| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py |
| ; RUN: llc -mtriple=riscv64 -mattr=+f,+d,+v -verify-machineinstrs < %s \ |
| ; RUN: -target-abi=lp64d -riscv-v-vector-bits-min=0 | FileCheck --check-prefix=CHECK-NOV %s |
| ; RUN: llc -mtriple=riscv64 -mattr=+f,+d,+v -verify-machineinstrs < %s \ |
| ; RUN: -target-abi=lp64d -riscv-v-vector-bits-min=-1 | FileCheck --check-prefix=CHECK-V %s |
| |
| ; i32 saturate |
| |
| define <2 x i32> @stest_f64i32(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: stest_f64i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: lui a2, 524288 |
| ; CHECK-NOV-NEXT: addiw a3, a2, -1 |
| ; CHECK-NOV-NEXT: fcvt.l.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a1, a3, .LBB0_5 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: bge a0, a3, .LBB0_6 |
| ; CHECK-NOV-NEXT: .LBB0_2: # %entry |
| ; CHECK-NOV-NEXT: bge a2, a0, .LBB0_7 |
| ; CHECK-NOV-NEXT: .LBB0_3: # %entry |
| ; CHECK-NOV-NEXT: bge a2, a1, .LBB0_8 |
| ; CHECK-NOV-NEXT: .LBB0_4: # %entry |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB0_5: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: blt a0, a3, .LBB0_2 |
| ; CHECK-NOV-NEXT: .LBB0_6: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a3 |
| ; CHECK-NOV-NEXT: blt a2, a3, .LBB0_3 |
| ; CHECK-NOV-NEXT: .LBB0_7: # %entry |
| ; CHECK-NOV-NEXT: lui a0, 524288 |
| ; CHECK-NOV-NEXT: blt a2, a1, .LBB0_4 |
| ; CHECK-NOV-NEXT: .LBB0_8: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 524288 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f64i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.x.f.v v8, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i64> |
| %0 = icmp slt <2 x i64> %conv, <i64 2147483647, i64 2147483647> |
| %spec.store.select = select <2 x i1> %0, <2 x i64> %conv, <2 x i64> <i64 2147483647, i64 2147483647> |
| %1 = icmp sgt <2 x i64> %spec.store.select, <i64 -2147483648, i64 -2147483648> |
| %spec.store.select7 = select <2 x i1> %1, <2 x i64> %spec.store.select, <2 x i64> <i64 -2147483648, i64 -2147483648> |
| %conv6 = trunc <2 x i64> %spec.store.select7 to <2 x i32> |
| ret <2 x i32> %conv6 |
| } |
| |
| define <2 x i32> @utest_f64i32(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: utest_f64i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: li a2, -1 |
| ; CHECK-NOV-NEXT: srli a2, a2, 32 |
| ; CHECK-NOV-NEXT: fcvt.lu.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: bgeu a0, a2, .LBB1_3 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: bgeu a1, a2, .LBB1_4 |
| ; CHECK-NOV-NEXT: .LBB1_2: # %entry |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB1_3: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: bltu a1, a2, .LBB1_2 |
| ; CHECK-NOV-NEXT: .LBB1_4: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utest_f64i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.xu.f.v v8, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x double> %x to <2 x i64> |
| %0 = icmp ult <2 x i64> %conv, <i64 4294967295, i64 4294967295> |
| %spec.store.select = select <2 x i1> %0, <2 x i64> %conv, <2 x i64> <i64 4294967295, i64 4294967295> |
| %conv6 = trunc <2 x i64> %spec.store.select to <2 x i32> |
| ret <2 x i32> %conv6 |
| } |
| |
| define <2 x i32> @ustest_f64i32(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: ustest_f64i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: li a2, -1 |
| ; CHECK-NOV-NEXT: srli a2, a2, 32 |
| ; CHECK-NOV-NEXT: fcvt.l.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a1, a2, .LBB2_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: .LBB2_2: # %entry |
| ; CHECK-NOV-NEXT: blt a0, a2, .LBB2_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: .LBB2_4: # %entry |
| ; CHECK-NOV-NEXT: sgtz a2, a1 |
| ; CHECK-NOV-NEXT: sgtz a3, a0 |
| ; CHECK-NOV-NEXT: neg a3, a3 |
| ; CHECK-NOV-NEXT: and a0, a3, a0 |
| ; CHECK-NOV-NEXT: neg a2, a2 |
| ; CHECK-NOV-NEXT: and a1, a2, a1 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f64i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.x.f.v v8, v8 |
| ; CHECK-V-NEXT: li a0, -1 |
| ; CHECK-V-NEXT: srli a0, a0, 32 |
| ; CHECK-V-NEXT: vmin.vx v8, v8, a0 |
| ; CHECK-V-NEXT: vmax.vx v8, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i64> |
| %0 = icmp slt <2 x i64> %conv, <i64 4294967295, i64 4294967295> |
| %spec.store.select = select <2 x i1> %0, <2 x i64> %conv, <2 x i64> <i64 4294967295, i64 4294967295> |
| %1 = icmp sgt <2 x i64> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <2 x i1> %1, <2 x i64> %spec.store.select, <2 x i64> zeroinitializer |
| %conv6 = trunc <2 x i64> %spec.store.select7 to <2 x i32> |
| ret <2 x i32> %conv6 |
| } |
| |
| define <4 x i32> @stest_f32i32(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: stest_f32i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fa3, rtz |
| ; CHECK-NOV-NEXT: lui a3, 524288 |
| ; CHECK-NOV-NEXT: addiw a6, a3, -1 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fa2, rtz |
| ; CHECK-NOV-NEXT: bge a1, a6, .LBB3_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fa1, rtz |
| ; CHECK-NOV-NEXT: bge a2, a6, .LBB3_11 |
| ; CHECK-NOV-NEXT: .LBB3_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a4, a6, .LBB3_12 |
| ; CHECK-NOV-NEXT: .LBB3_3: # %entry |
| ; CHECK-NOV-NEXT: bge a5, a6, .LBB3_13 |
| ; CHECK-NOV-NEXT: .LBB3_4: # %entry |
| ; CHECK-NOV-NEXT: bge a3, a5, .LBB3_14 |
| ; CHECK-NOV-NEXT: .LBB3_5: # %entry |
| ; CHECK-NOV-NEXT: bge a3, a4, .LBB3_15 |
| ; CHECK-NOV-NEXT: .LBB3_6: # %entry |
| ; CHECK-NOV-NEXT: bge a3, a2, .LBB3_16 |
| ; CHECK-NOV-NEXT: .LBB3_7: # %entry |
| ; CHECK-NOV-NEXT: blt a3, a1, .LBB3_9 |
| ; CHECK-NOV-NEXT: .LBB3_8: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 524288 |
| ; CHECK-NOV-NEXT: .LBB3_9: # %entry |
| ; CHECK-NOV-NEXT: sw a1, 12(a0) |
| ; CHECK-NOV-NEXT: sw a2, 8(a0) |
| ; CHECK-NOV-NEXT: sw a4, 4(a0) |
| ; CHECK-NOV-NEXT: sw a5, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB3_10: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a6 |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fa1, rtz |
| ; CHECK-NOV-NEXT: blt a2, a6, .LBB3_2 |
| ; CHECK-NOV-NEXT: .LBB3_11: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a6 |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a4, a6, .LBB3_3 |
| ; CHECK-NOV-NEXT: .LBB3_12: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a6 |
| ; CHECK-NOV-NEXT: blt a5, a6, .LBB3_4 |
| ; CHECK-NOV-NEXT: .LBB3_13: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a6 |
| ; CHECK-NOV-NEXT: blt a3, a6, .LBB3_5 |
| ; CHECK-NOV-NEXT: .LBB3_14: # %entry |
| ; CHECK-NOV-NEXT: lui a5, 524288 |
| ; CHECK-NOV-NEXT: blt a3, a4, .LBB3_6 |
| ; CHECK-NOV-NEXT: .LBB3_15: # %entry |
| ; CHECK-NOV-NEXT: lui a4, 524288 |
| ; CHECK-NOV-NEXT: blt a3, a2, .LBB3_7 |
| ; CHECK-NOV-NEXT: .LBB3_16: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 524288 |
| ; CHECK-NOV-NEXT: bge a3, a1, .LBB3_8 |
| ; CHECK-NOV-NEXT: j .LBB3_9 |
| ; |
| ; CHECK-V-LABEL: stest_f32i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfwcvt.rtz.x.f.v v10, v8 |
| ; CHECK-V-NEXT: vnclip.wi v8, v10, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x float> %x to <4 x i64> |
| %0 = icmp slt <4 x i64> %conv, <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647> |
| %spec.store.select = select <4 x i1> %0, <4 x i64> %conv, <4 x i64> <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647> |
| %1 = icmp sgt <4 x i64> %spec.store.select, <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648> |
| %spec.store.select7 = select <4 x i1> %1, <4 x i64> %spec.store.select, <4 x i64> <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648> |
| %conv6 = trunc <4 x i64> %spec.store.select7 to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @utest_f32i32(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: utest_f32i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a1, fa0, rtz |
| ; CHECK-NOV-NEXT: li a3, -1 |
| ; CHECK-NOV-NEXT: srli a3, a3, 32 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fa1, rtz |
| ; CHECK-NOV-NEXT: bgeu a1, a3, .LBB4_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a4, fa2, rtz |
| ; CHECK-NOV-NEXT: bgeu a2, a3, .LBB4_7 |
| ; CHECK-NOV-NEXT: .LBB4_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a5, fa3, rtz |
| ; CHECK-NOV-NEXT: bgeu a4, a3, .LBB4_8 |
| ; CHECK-NOV-NEXT: .LBB4_3: # %entry |
| ; CHECK-NOV-NEXT: bltu a5, a3, .LBB4_5 |
| ; CHECK-NOV-NEXT: .LBB4_4: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a3 |
| ; CHECK-NOV-NEXT: .LBB4_5: # %entry |
| ; CHECK-NOV-NEXT: sw a5, 12(a0) |
| ; CHECK-NOV-NEXT: sw a4, 8(a0) |
| ; CHECK-NOV-NEXT: sw a2, 4(a0) |
| ; CHECK-NOV-NEXT: sw a1, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB4_6: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a4, fa2, rtz |
| ; CHECK-NOV-NEXT: bltu a2, a3, .LBB4_2 |
| ; CHECK-NOV-NEXT: .LBB4_7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a5, fa3, rtz |
| ; CHECK-NOV-NEXT: bltu a4, a3, .LBB4_3 |
| ; CHECK-NOV-NEXT: .LBB4_8: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a3 |
| ; CHECK-NOV-NEXT: bgeu a5, a3, .LBB4_4 |
| ; CHECK-NOV-NEXT: j .LBB4_5 |
| ; |
| ; CHECK-V-LABEL: utest_f32i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfwcvt.rtz.xu.f.v v10, v8 |
| ; CHECK-V-NEXT: vnclipu.wi v8, v10, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <4 x float> %x to <4 x i64> |
| %0 = icmp ult <4 x i64> %conv, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295> |
| %spec.store.select = select <4 x i1> %0, <4 x i64> %conv, <4 x i64> <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295> |
| %conv6 = trunc <4 x i64> %spec.store.select to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @ustest_f32i32(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: ustest_f32i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fa3, rtz |
| ; CHECK-NOV-NEXT: li a4, -1 |
| ; CHECK-NOV-NEXT: srli a4, a4, 32 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fa2, rtz |
| ; CHECK-NOV-NEXT: bge a1, a4, .LBB5_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fa1, rtz |
| ; CHECK-NOV-NEXT: bge a2, a4, .LBB5_7 |
| ; CHECK-NOV-NEXT: .LBB5_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a3, a4, .LBB5_8 |
| ; CHECK-NOV-NEXT: .LBB5_3: # %entry |
| ; CHECK-NOV-NEXT: blt a5, a4, .LBB5_5 |
| ; CHECK-NOV-NEXT: .LBB5_4: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a4 |
| ; CHECK-NOV-NEXT: .LBB5_5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a4, a1 |
| ; CHECK-NOV-NEXT: sgtz a6, a2 |
| ; CHECK-NOV-NEXT: sgtz a7, a3 |
| ; CHECK-NOV-NEXT: sgtz t0, a5 |
| ; CHECK-NOV-NEXT: negw t0, t0 |
| ; CHECK-NOV-NEXT: and a5, t0, a5 |
| ; CHECK-NOV-NEXT: negw a7, a7 |
| ; CHECK-NOV-NEXT: and a3, a7, a3 |
| ; CHECK-NOV-NEXT: negw a6, a6 |
| ; CHECK-NOV-NEXT: and a2, a6, a2 |
| ; CHECK-NOV-NEXT: negw a4, a4 |
| ; CHECK-NOV-NEXT: and a1, a4, a1 |
| ; CHECK-NOV-NEXT: sw a1, 12(a0) |
| ; CHECK-NOV-NEXT: sw a2, 8(a0) |
| ; CHECK-NOV-NEXT: sw a3, 4(a0) |
| ; CHECK-NOV-NEXT: sw a5, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB5_6: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a4 |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fa1, rtz |
| ; CHECK-NOV-NEXT: blt a2, a4, .LBB5_2 |
| ; CHECK-NOV-NEXT: .LBB5_7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a4 |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a3, a4, .LBB5_3 |
| ; CHECK-NOV-NEXT: .LBB5_8: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a4 |
| ; CHECK-NOV-NEXT: bge a5, a4, .LBB5_4 |
| ; CHECK-NOV-NEXT: j .LBB5_5 |
| ; |
| ; CHECK-V-LABEL: ustest_f32i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfwcvt.rtz.x.f.v v10, v8 |
| ; CHECK-V-NEXT: li a0, -1 |
| ; CHECK-V-NEXT: srli a0, a0, 32 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e64, m2, ta, ma |
| ; CHECK-V-NEXT: vmin.vx v8, v10, a0 |
| ; CHECK-V-NEXT: vmax.vx v10, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v10, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x float> %x to <4 x i64> |
| %0 = icmp slt <4 x i64> %conv, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295> |
| %spec.store.select = select <4 x i1> %0, <4 x i64> %conv, <4 x i64> <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295> |
| %1 = icmp sgt <4 x i64> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <4 x i1> %1, <4 x i64> %spec.store.select, <4 x i64> zeroinitializer |
| %conv6 = trunc <4 x i64> %spec.store.select7 to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @stest_f16i32(<4 x half> %x) { |
| ; CHECK-NOV-LABEL: stest_f16i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -64 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-NOV-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -64 |
| ; CHECK-NOV-NEXT: lhu s1, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 8(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 16(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.l.s s1, fs2, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a1, 524288 |
| ; CHECK-NOV-NEXT: addiw a4, a1, -1 |
| ; CHECK-NOV-NEXT: bge a0, a4, .LBB6_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs1, rtz |
| ; CHECK-NOV-NEXT: bge s1, a4, .LBB6_11 |
| ; CHECK-NOV-NEXT: .LBB6_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: bge a2, a4, .LBB6_12 |
| ; CHECK-NOV-NEXT: .LBB6_3: # %entry |
| ; CHECK-NOV-NEXT: bge a3, a4, .LBB6_13 |
| ; CHECK-NOV-NEXT: .LBB6_4: # %entry |
| ; CHECK-NOV-NEXT: bge a1, a3, .LBB6_14 |
| ; CHECK-NOV-NEXT: .LBB6_5: # %entry |
| ; CHECK-NOV-NEXT: bge a1, a2, .LBB6_15 |
| ; CHECK-NOV-NEXT: .LBB6_6: # %entry |
| ; CHECK-NOV-NEXT: bge a1, s1, .LBB6_16 |
| ; CHECK-NOV-NEXT: .LBB6_7: # %entry |
| ; CHECK-NOV-NEXT: blt a1, a0, .LBB6_9 |
| ; CHECK-NOV-NEXT: .LBB6_8: # %entry |
| ; CHECK-NOV-NEXT: lui a0, 524288 |
| ; CHECK-NOV-NEXT: .LBB6_9: # %entry |
| ; CHECK-NOV-NEXT: sw a0, 12(s0) |
| ; CHECK-NOV-NEXT: sw s1, 8(s0) |
| ; CHECK-NOV-NEXT: sw a2, 4(s0) |
| ; CHECK-NOV-NEXT: sw a3, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 64 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB6_10: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a4 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs1, rtz |
| ; CHECK-NOV-NEXT: blt s1, a4, .LBB6_2 |
| ; CHECK-NOV-NEXT: .LBB6_11: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a4 |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: blt a2, a4, .LBB6_3 |
| ; CHECK-NOV-NEXT: .LBB6_12: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a4 |
| ; CHECK-NOV-NEXT: blt a3, a4, .LBB6_4 |
| ; CHECK-NOV-NEXT: .LBB6_13: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a4 |
| ; CHECK-NOV-NEXT: blt a1, a4, .LBB6_5 |
| ; CHECK-NOV-NEXT: .LBB6_14: # %entry |
| ; CHECK-NOV-NEXT: lui a3, 524288 |
| ; CHECK-NOV-NEXT: blt a1, a2, .LBB6_6 |
| ; CHECK-NOV-NEXT: .LBB6_15: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 524288 |
| ; CHECK-NOV-NEXT: blt a1, s1, .LBB6_7 |
| ; CHECK-NOV-NEXT: .LBB6_16: # %entry |
| ; CHECK-NOV-NEXT: lui s1, 524288 |
| ; CHECK-NOV-NEXT: bge a1, a0, .LBB6_8 |
| ; CHECK-NOV-NEXT: j .LBB6_9 |
| ; |
| ; CHECK-V-LABEL: stest_f16i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -48 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 48 |
| ; CHECK-V-NEXT: sd ra, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu a0, 24(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v10, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e64, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 2 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 48 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x half> %x to <4 x i64> |
| %0 = icmp slt <4 x i64> %conv, <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647> |
| %spec.store.select = select <4 x i1> %0, <4 x i64> %conv, <4 x i64> <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647> |
| %1 = icmp sgt <4 x i64> %spec.store.select, <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648> |
| %spec.store.select7 = select <4 x i1> %1, <4 x i64> %spec.store.select, <4 x i64> <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648> |
| %conv6 = trunc <4 x i64> %spec.store.select7 to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @utesth_f16i32(<4 x half> %x) { |
| ; CHECK-NOV-LABEL: utesth_f16i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -64 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-NOV-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -64 |
| ; CHECK-NOV-NEXT: lhu s1, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 16(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 8(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s s1, fs2, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: li a1, -1 |
| ; CHECK-NOV-NEXT: srli a1, a1, 32 |
| ; CHECK-NOV-NEXT: bgeu a0, a1, .LBB7_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fs1, rtz |
| ; CHECK-NOV-NEXT: bgeu s1, a1, .LBB7_7 |
| ; CHECK-NOV-NEXT: .LBB7_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: bgeu a2, a1, .LBB7_8 |
| ; CHECK-NOV-NEXT: .LBB7_3: # %entry |
| ; CHECK-NOV-NEXT: bltu a3, a1, .LBB7_5 |
| ; CHECK-NOV-NEXT: .LBB7_4: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a1 |
| ; CHECK-NOV-NEXT: .LBB7_5: # %entry |
| ; CHECK-NOV-NEXT: sw a3, 12(s0) |
| ; CHECK-NOV-NEXT: sw a2, 8(s0) |
| ; CHECK-NOV-NEXT: sw s1, 4(s0) |
| ; CHECK-NOV-NEXT: sw a0, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 64 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB7_6: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fs1, rtz |
| ; CHECK-NOV-NEXT: bltu s1, a1, .LBB7_2 |
| ; CHECK-NOV-NEXT: .LBB7_7: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: bltu a2, a1, .LBB7_3 |
| ; CHECK-NOV-NEXT: .LBB7_8: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: bgeu a3, a1, .LBB7_4 |
| ; CHECK-NOV-NEXT: j .LBB7_5 |
| ; |
| ; CHECK-V-LABEL: utesth_f16i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -48 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 48 |
| ; CHECK-V-NEXT: sd ra, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu a0, 24(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v10, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e64, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 2 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 48 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <4 x half> %x to <4 x i64> |
| %0 = icmp ult <4 x i64> %conv, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295> |
| %spec.store.select = select <4 x i1> %0, <4 x i64> %conv, <4 x i64> <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295> |
| %conv6 = trunc <4 x i64> %spec.store.select to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @ustest_f16i32(<4 x half> %x) { |
| ; CHECK-NOV-LABEL: ustest_f16i32: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -64 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-NOV-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -64 |
| ; CHECK-NOV-NEXT: lhu s1, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 8(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 16(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.l.s s1, fs2, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: li a2, -1 |
| ; CHECK-NOV-NEXT: srli a2, a2, 32 |
| ; CHECK-NOV-NEXT: bge a0, a2, .LBB8_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs1, rtz |
| ; CHECK-NOV-NEXT: bge s1, a2, .LBB8_7 |
| ; CHECK-NOV-NEXT: .LBB8_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: bge a1, a2, .LBB8_8 |
| ; CHECK-NOV-NEXT: .LBB8_3: # %entry |
| ; CHECK-NOV-NEXT: blt a3, a2, .LBB8_5 |
| ; CHECK-NOV-NEXT: .LBB8_4: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a2 |
| ; CHECK-NOV-NEXT: .LBB8_5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a2, a0 |
| ; CHECK-NOV-NEXT: sgtz a4, s1 |
| ; CHECK-NOV-NEXT: sgtz a5, a1 |
| ; CHECK-NOV-NEXT: sgtz a6, a3 |
| ; CHECK-NOV-NEXT: negw a6, a6 |
| ; CHECK-NOV-NEXT: and a3, a6, a3 |
| ; CHECK-NOV-NEXT: negw a5, a5 |
| ; CHECK-NOV-NEXT: and a1, a5, a1 |
| ; CHECK-NOV-NEXT: negw a4, a4 |
| ; CHECK-NOV-NEXT: and a4, a4, s1 |
| ; CHECK-NOV-NEXT: negw a2, a2 |
| ; CHECK-NOV-NEXT: and a0, a2, a0 |
| ; CHECK-NOV-NEXT: sw a0, 12(s0) |
| ; CHECK-NOV-NEXT: sw a4, 8(s0) |
| ; CHECK-NOV-NEXT: sw a1, 4(s0) |
| ; CHECK-NOV-NEXT: sw a3, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 64 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB8_6: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs1, rtz |
| ; CHECK-NOV-NEXT: blt s1, a2, .LBB8_2 |
| ; CHECK-NOV-NEXT: .LBB8_7: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: blt a1, a2, .LBB8_3 |
| ; CHECK-NOV-NEXT: .LBB8_8: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: bge a3, a2, .LBB8_4 |
| ; CHECK-NOV-NEXT: j .LBB8_5 |
| ; |
| ; CHECK-V-LABEL: ustest_f16i32: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -48 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 48 |
| ; CHECK-V-NEXT: sd ra, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu a0, 24(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e64, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v10, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v10, 2 |
| ; CHECK-V-NEXT: li a0, -1 |
| ; CHECK-V-NEXT: srli a0, a0, 32 |
| ; CHECK-V-NEXT: vmin.vx v8, v8, a0 |
| ; CHECK-V-NEXT: vmax.vx v10, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 48 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x half> %x to <4 x i64> |
| %0 = icmp slt <4 x i64> %conv, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295> |
| %spec.store.select = select <4 x i1> %0, <4 x i64> %conv, <4 x i64> <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295> |
| %1 = icmp sgt <4 x i64> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <4 x i1> %1, <4 x i64> %spec.store.select, <4 x i64> zeroinitializer |
| %conv6 = trunc <4 x i64> %spec.store.select7 to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| ; i16 saturate |
| |
| define <2 x i16> @stest_f64i16(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: stest_f64i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: lui a2, 8 |
| ; CHECK-NOV-NEXT: addiw a2, a2, -1 |
| ; CHECK-NOV-NEXT: fcvt.w.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a1, a2, .LBB9_5 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: bge a0, a2, .LBB9_6 |
| ; CHECK-NOV-NEXT: .LBB9_2: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 1048568 |
| ; CHECK-NOV-NEXT: bge a2, a0, .LBB9_7 |
| ; CHECK-NOV-NEXT: .LBB9_3: # %entry |
| ; CHECK-NOV-NEXT: bge a2, a1, .LBB9_8 |
| ; CHECK-NOV-NEXT: .LBB9_4: # %entry |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB9_5: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: blt a0, a2, .LBB9_2 |
| ; CHECK-NOV-NEXT: .LBB9_6: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: lui a2, 1048568 |
| ; CHECK-NOV-NEXT: blt a2, a0, .LBB9_3 |
| ; CHECK-NOV-NEXT: .LBB9_7: # %entry |
| ; CHECK-NOV-NEXT: lui a0, 1048568 |
| ; CHECK-NOV-NEXT: blt a2, a1, .LBB9_4 |
| ; CHECK-NOV-NEXT: .LBB9_8: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 1048568 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f64i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vfncvt.rtz.x.f.w v9, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf4, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v9, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i32> |
| %0 = icmp slt <2 x i32> %conv, <i32 32767, i32 32767> |
| %spec.store.select = select <2 x i1> %0, <2 x i32> %conv, <2 x i32> <i32 32767, i32 32767> |
| %1 = icmp sgt <2 x i32> %spec.store.select, <i32 -32768, i32 -32768> |
| %spec.store.select7 = select <2 x i1> %1, <2 x i32> %spec.store.select, <2 x i32> <i32 -32768, i32 -32768> |
| %conv6 = trunc <2 x i32> %spec.store.select7 to <2 x i16> |
| ret <2 x i16> %conv6 |
| } |
| |
| define <2 x i16> @utest_f64i16(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: utest_f64i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.wu.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a2, 16 |
| ; CHECK-NOV-NEXT: addiw a2, a2, -1 |
| ; CHECK-NOV-NEXT: fcvt.wu.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: bgeu a0, a2, .LBB10_3 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: bgeu a1, a2, .LBB10_4 |
| ; CHECK-NOV-NEXT: .LBB10_2: # %entry |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB10_3: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: bltu a1, a2, .LBB10_2 |
| ; CHECK-NOV-NEXT: .LBB10_4: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utest_f64i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vfncvt.rtz.xu.f.w v9, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf4, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v9, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x double> %x to <2 x i32> |
| %0 = icmp ult <2 x i32> %conv, <i32 65535, i32 65535> |
| %spec.store.select = select <2 x i1> %0, <2 x i32> %conv, <2 x i32> <i32 65535, i32 65535> |
| %conv6 = trunc <2 x i32> %spec.store.select to <2 x i16> |
| ret <2 x i16> %conv6 |
| } |
| |
| define <2 x i16> @ustest_f64i16(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: ustest_f64i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: lui a2, 16 |
| ; CHECK-NOV-NEXT: addiw a2, a2, -1 |
| ; CHECK-NOV-NEXT: fcvt.w.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a1, a2, .LBB11_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: .LBB11_2: # %entry |
| ; CHECK-NOV-NEXT: blt a0, a2, .LBB11_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: .LBB11_4: # %entry |
| ; CHECK-NOV-NEXT: sgtz a2, a1 |
| ; CHECK-NOV-NEXT: sgtz a3, a0 |
| ; CHECK-NOV-NEXT: neg a3, a3 |
| ; CHECK-NOV-NEXT: and a0, a3, a0 |
| ; CHECK-NOV-NEXT: neg a2, a2 |
| ; CHECK-NOV-NEXT: and a1, a2, a1 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f64i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vfncvt.rtz.x.f.w v9, v8 |
| ; CHECK-V-NEXT: lui a0, 16 |
| ; CHECK-V-NEXT: addi a0, a0, -1 |
| ; CHECK-V-NEXT: vmin.vx v8, v9, a0 |
| ; CHECK-V-NEXT: vmax.vx v8, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf4, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i32> |
| %0 = icmp slt <2 x i32> %conv, <i32 65535, i32 65535> |
| %spec.store.select = select <2 x i1> %0, <2 x i32> %conv, <2 x i32> <i32 65535, i32 65535> |
| %1 = icmp sgt <2 x i32> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <2 x i1> %1, <2 x i32> %spec.store.select, <2 x i32> zeroinitializer |
| %conv6 = trunc <2 x i32> %spec.store.select7 to <2 x i16> |
| ret <2 x i16> %conv6 |
| } |
| |
| define <4 x i16> @stest_f32i16(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: stest_f32i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a1, fa3, rtz |
| ; CHECK-NOV-NEXT: lui a5, 8 |
| ; CHECK-NOV-NEXT: addiw a5, a5, -1 |
| ; CHECK-NOV-NEXT: fcvt.w.s a2, fa2, rtz |
| ; CHECK-NOV-NEXT: bge a1, a5, .LBB12_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a3, fa1, rtz |
| ; CHECK-NOV-NEXT: bge a2, a5, .LBB12_11 |
| ; CHECK-NOV-NEXT: .LBB12_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a4, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a3, a5, .LBB12_12 |
| ; CHECK-NOV-NEXT: .LBB12_3: # %entry |
| ; CHECK-NOV-NEXT: bge a4, a5, .LBB12_13 |
| ; CHECK-NOV-NEXT: .LBB12_4: # %entry |
| ; CHECK-NOV-NEXT: lui a5, 1048568 |
| ; CHECK-NOV-NEXT: bge a5, a4, .LBB12_14 |
| ; CHECK-NOV-NEXT: .LBB12_5: # %entry |
| ; CHECK-NOV-NEXT: bge a5, a3, .LBB12_15 |
| ; CHECK-NOV-NEXT: .LBB12_6: # %entry |
| ; CHECK-NOV-NEXT: bge a5, a2, .LBB12_16 |
| ; CHECK-NOV-NEXT: .LBB12_7: # %entry |
| ; CHECK-NOV-NEXT: blt a5, a1, .LBB12_9 |
| ; CHECK-NOV-NEXT: .LBB12_8: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 1048568 |
| ; CHECK-NOV-NEXT: .LBB12_9: # %entry |
| ; CHECK-NOV-NEXT: sh a1, 6(a0) |
| ; CHECK-NOV-NEXT: sh a2, 4(a0) |
| ; CHECK-NOV-NEXT: sh a3, 2(a0) |
| ; CHECK-NOV-NEXT: sh a4, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB12_10: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a5 |
| ; CHECK-NOV-NEXT: fcvt.w.s a3, fa1, rtz |
| ; CHECK-NOV-NEXT: blt a2, a5, .LBB12_2 |
| ; CHECK-NOV-NEXT: .LBB12_11: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a5 |
| ; CHECK-NOV-NEXT: fcvt.w.s a4, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a3, a5, .LBB12_3 |
| ; CHECK-NOV-NEXT: .LBB12_12: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a5 |
| ; CHECK-NOV-NEXT: blt a4, a5, .LBB12_4 |
| ; CHECK-NOV-NEXT: .LBB12_13: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a5 |
| ; CHECK-NOV-NEXT: lui a5, 1048568 |
| ; CHECK-NOV-NEXT: blt a5, a4, .LBB12_5 |
| ; CHECK-NOV-NEXT: .LBB12_14: # %entry |
| ; CHECK-NOV-NEXT: lui a4, 1048568 |
| ; CHECK-NOV-NEXT: blt a5, a3, .LBB12_6 |
| ; CHECK-NOV-NEXT: .LBB12_15: # %entry |
| ; CHECK-NOV-NEXT: lui a3, 1048568 |
| ; CHECK-NOV-NEXT: blt a5, a2, .LBB12_7 |
| ; CHECK-NOV-NEXT: .LBB12_16: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 1048568 |
| ; CHECK-NOV-NEXT: bge a5, a1, .LBB12_8 |
| ; CHECK-NOV-NEXT: j .LBB12_9 |
| ; |
| ; CHECK-V-LABEL: stest_f32i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.x.f.v v8, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf2, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x float> %x to <4 x i32> |
| %0 = icmp slt <4 x i32> %conv, <i32 32767, i32 32767, i32 32767, i32 32767> |
| %spec.store.select = select <4 x i1> %0, <4 x i32> %conv, <4 x i32> <i32 32767, i32 32767, i32 32767, i32 32767> |
| %1 = icmp sgt <4 x i32> %spec.store.select, <i32 -32768, i32 -32768, i32 -32768, i32 -32768> |
| %spec.store.select7 = select <4 x i1> %1, <4 x i32> %spec.store.select, <4 x i32> <i32 -32768, i32 -32768, i32 -32768, i32 -32768> |
| %conv6 = trunc <4 x i32> %spec.store.select7 to <4 x i16> |
| ret <4 x i16> %conv6 |
| } |
| |
| define <4 x i16> @utest_f32i16(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: utest_f32i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.wu.s a1, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a3, 16 |
| ; CHECK-NOV-NEXT: addiw a3, a3, -1 |
| ; CHECK-NOV-NEXT: fcvt.wu.s a2, fa1, rtz |
| ; CHECK-NOV-NEXT: bgeu a1, a3, .LBB13_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.wu.s a4, fa2, rtz |
| ; CHECK-NOV-NEXT: bgeu a2, a3, .LBB13_7 |
| ; CHECK-NOV-NEXT: .LBB13_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.wu.s a5, fa3, rtz |
| ; CHECK-NOV-NEXT: bgeu a4, a3, .LBB13_8 |
| ; CHECK-NOV-NEXT: .LBB13_3: # %entry |
| ; CHECK-NOV-NEXT: bltu a5, a3, .LBB13_5 |
| ; CHECK-NOV-NEXT: .LBB13_4: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a3 |
| ; CHECK-NOV-NEXT: .LBB13_5: # %entry |
| ; CHECK-NOV-NEXT: sh a5, 6(a0) |
| ; CHECK-NOV-NEXT: sh a4, 4(a0) |
| ; CHECK-NOV-NEXT: sh a2, 2(a0) |
| ; CHECK-NOV-NEXT: sh a1, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB13_6: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: fcvt.wu.s a4, fa2, rtz |
| ; CHECK-NOV-NEXT: bltu a2, a3, .LBB13_2 |
| ; CHECK-NOV-NEXT: .LBB13_7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: fcvt.wu.s a5, fa3, rtz |
| ; CHECK-NOV-NEXT: bltu a4, a3, .LBB13_3 |
| ; CHECK-NOV-NEXT: .LBB13_8: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a3 |
| ; CHECK-NOV-NEXT: bgeu a5, a3, .LBB13_4 |
| ; CHECK-NOV-NEXT: j .LBB13_5 |
| ; |
| ; CHECK-V-LABEL: utest_f32i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.xu.f.v v8, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf2, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <4 x float> %x to <4 x i32> |
| %0 = icmp ult <4 x i32> %conv, <i32 65535, i32 65535, i32 65535, i32 65535> |
| %spec.store.select = select <4 x i1> %0, <4 x i32> %conv, <4 x i32> <i32 65535, i32 65535, i32 65535, i32 65535> |
| %conv6 = trunc <4 x i32> %spec.store.select to <4 x i16> |
| ret <4 x i16> %conv6 |
| } |
| |
| define <4 x i16> @ustest_f32i16(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: ustest_f32i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a1, fa3, rtz |
| ; CHECK-NOV-NEXT: lui a4, 16 |
| ; CHECK-NOV-NEXT: addiw a4, a4, -1 |
| ; CHECK-NOV-NEXT: fcvt.w.s a2, fa2, rtz |
| ; CHECK-NOV-NEXT: bge a1, a4, .LBB14_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a3, fa1, rtz |
| ; CHECK-NOV-NEXT: bge a2, a4, .LBB14_7 |
| ; CHECK-NOV-NEXT: .LBB14_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a3, a4, .LBB14_8 |
| ; CHECK-NOV-NEXT: .LBB14_3: # %entry |
| ; CHECK-NOV-NEXT: blt a5, a4, .LBB14_5 |
| ; CHECK-NOV-NEXT: .LBB14_4: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a4 |
| ; CHECK-NOV-NEXT: .LBB14_5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a4, a1 |
| ; CHECK-NOV-NEXT: sgtz a6, a2 |
| ; CHECK-NOV-NEXT: sgtz a7, a3 |
| ; CHECK-NOV-NEXT: sgtz t0, a5 |
| ; CHECK-NOV-NEXT: negw t0, t0 |
| ; CHECK-NOV-NEXT: and a5, t0, a5 |
| ; CHECK-NOV-NEXT: negw a7, a7 |
| ; CHECK-NOV-NEXT: and a3, a7, a3 |
| ; CHECK-NOV-NEXT: negw a6, a6 |
| ; CHECK-NOV-NEXT: and a2, a6, a2 |
| ; CHECK-NOV-NEXT: negw a4, a4 |
| ; CHECK-NOV-NEXT: and a1, a4, a1 |
| ; CHECK-NOV-NEXT: sh a1, 6(a0) |
| ; CHECK-NOV-NEXT: sh a2, 4(a0) |
| ; CHECK-NOV-NEXT: sh a3, 2(a0) |
| ; CHECK-NOV-NEXT: sh a5, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB14_6: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a4 |
| ; CHECK-NOV-NEXT: fcvt.w.s a3, fa1, rtz |
| ; CHECK-NOV-NEXT: blt a2, a4, .LBB14_2 |
| ; CHECK-NOV-NEXT: .LBB14_7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a4 |
| ; CHECK-NOV-NEXT: fcvt.w.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a3, a4, .LBB14_3 |
| ; CHECK-NOV-NEXT: .LBB14_8: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a4 |
| ; CHECK-NOV-NEXT: bge a5, a4, .LBB14_4 |
| ; CHECK-NOV-NEXT: j .LBB14_5 |
| ; |
| ; CHECK-V-LABEL: ustest_f32i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.x.f.v v8, v8 |
| ; CHECK-V-NEXT: lui a0, 16 |
| ; CHECK-V-NEXT: addi a0, a0, -1 |
| ; CHECK-V-NEXT: vmin.vx v8, v8, a0 |
| ; CHECK-V-NEXT: vmax.vx v8, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf2, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x float> %x to <4 x i32> |
| %0 = icmp slt <4 x i32> %conv, <i32 65535, i32 65535, i32 65535, i32 65535> |
| %spec.store.select = select <4 x i1> %0, <4 x i32> %conv, <4 x i32> <i32 65535, i32 65535, i32 65535, i32 65535> |
| %1 = icmp sgt <4 x i32> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <4 x i1> %1, <4 x i32> %spec.store.select, <4 x i32> zeroinitializer |
| %conv6 = trunc <4 x i32> %spec.store.select7 to <4 x i16> |
| ret <4 x i16> %conv6 |
| } |
| |
| define <8 x i16> @stest_f16i16(<8 x half> %x) { |
| ; CHECK-NOV-LABEL: stest_f16i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -128 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 128 |
| ; CHECK-NOV-NEXT: sd ra, 120(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 112(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 104(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 96(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 88(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s4, 80(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s5, 72(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s6, 64(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s7, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs4, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs5, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs6, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset s4, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset s5, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset s6, -64 |
| ; CHECK-NOV-NEXT: .cfi_offset s7, -72 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -80 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -88 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -96 |
| ; CHECK-NOV-NEXT: .cfi_offset fs3, -104 |
| ; CHECK-NOV-NEXT: .cfi_offset fs4, -112 |
| ; CHECK-NOV-NEXT: .cfi_offset fs5, -120 |
| ; CHECK-NOV-NEXT: .cfi_offset fs6, -128 |
| ; CHECK-NOV-NEXT: lhu s1, 56(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 8(a1) |
| ; CHECK-NOV-NEXT: lhu s4, 16(a1) |
| ; CHECK-NOV-NEXT: lhu s5, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s6, 32(a1) |
| ; CHECK-NOV-NEXT: lhu s7, 40(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 48(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs6, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s7 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs5, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs4, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs3, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.l.s s1, fs6, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a7, 8 |
| ; CHECK-NOV-NEXT: addiw a7, a7, -1 |
| ; CHECK-NOV-NEXT: bge a0, a7, .LBB15_18 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs5, rtz |
| ; CHECK-NOV-NEXT: bge s1, a7, .LBB15_19 |
| ; CHECK-NOV-NEXT: .LBB15_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs4, rtz |
| ; CHECK-NOV-NEXT: bge a1, a7, .LBB15_20 |
| ; CHECK-NOV-NEXT: .LBB15_3: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs3, rtz |
| ; CHECK-NOV-NEXT: bge a2, a7, .LBB15_21 |
| ; CHECK-NOV-NEXT: .LBB15_4: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fs2, rtz |
| ; CHECK-NOV-NEXT: bge a3, a7, .LBB15_22 |
| ; CHECK-NOV-NEXT: .LBB15_5: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fs1, rtz |
| ; CHECK-NOV-NEXT: bge a4, a7, .LBB15_23 |
| ; CHECK-NOV-NEXT: .LBB15_6: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a6, fs0, rtz |
| ; CHECK-NOV-NEXT: bge a5, a7, .LBB15_24 |
| ; CHECK-NOV-NEXT: .LBB15_7: # %entry |
| ; CHECK-NOV-NEXT: bge a6, a7, .LBB15_25 |
| ; CHECK-NOV-NEXT: .LBB15_8: # %entry |
| ; CHECK-NOV-NEXT: lui a7, 1048568 |
| ; CHECK-NOV-NEXT: bge a7, a6, .LBB15_26 |
| ; CHECK-NOV-NEXT: .LBB15_9: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a5, .LBB15_27 |
| ; CHECK-NOV-NEXT: .LBB15_10: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a4, .LBB15_28 |
| ; CHECK-NOV-NEXT: .LBB15_11: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a3, .LBB15_29 |
| ; CHECK-NOV-NEXT: .LBB15_12: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a2, .LBB15_30 |
| ; CHECK-NOV-NEXT: .LBB15_13: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a1, .LBB15_31 |
| ; CHECK-NOV-NEXT: .LBB15_14: # %entry |
| ; CHECK-NOV-NEXT: bge a7, s1, .LBB15_32 |
| ; CHECK-NOV-NEXT: .LBB15_15: # %entry |
| ; CHECK-NOV-NEXT: blt a7, a0, .LBB15_17 |
| ; CHECK-NOV-NEXT: .LBB15_16: # %entry |
| ; CHECK-NOV-NEXT: lui a0, 1048568 |
| ; CHECK-NOV-NEXT: .LBB15_17: # %entry |
| ; CHECK-NOV-NEXT: sh a0, 14(s0) |
| ; CHECK-NOV-NEXT: sh s1, 12(s0) |
| ; CHECK-NOV-NEXT: sh a1, 10(s0) |
| ; CHECK-NOV-NEXT: sh a2, 8(s0) |
| ; CHECK-NOV-NEXT: sh a3, 6(s0) |
| ; CHECK-NOV-NEXT: sh a4, 4(s0) |
| ; CHECK-NOV-NEXT: sh a5, 2(s0) |
| ; CHECK-NOV-NEXT: sh a6, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 120(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 112(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 104(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 96(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 88(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s4, 80(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s5, 72(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s6, 64(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s7, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs4, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs5, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs6, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 128 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB15_18: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs5, rtz |
| ; CHECK-NOV-NEXT: blt s1, a7, .LBB15_2 |
| ; CHECK-NOV-NEXT: .LBB15_19: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs4, rtz |
| ; CHECK-NOV-NEXT: blt a1, a7, .LBB15_3 |
| ; CHECK-NOV-NEXT: .LBB15_20: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs3, rtz |
| ; CHECK-NOV-NEXT: blt a2, a7, .LBB15_4 |
| ; CHECK-NOV-NEXT: .LBB15_21: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fs2, rtz |
| ; CHECK-NOV-NEXT: blt a3, a7, .LBB15_5 |
| ; CHECK-NOV-NEXT: .LBB15_22: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fs1, rtz |
| ; CHECK-NOV-NEXT: blt a4, a7, .LBB15_6 |
| ; CHECK-NOV-NEXT: .LBB15_23: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a6, fs0, rtz |
| ; CHECK-NOV-NEXT: blt a5, a7, .LBB15_7 |
| ; CHECK-NOV-NEXT: .LBB15_24: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a7 |
| ; CHECK-NOV-NEXT: blt a6, a7, .LBB15_8 |
| ; CHECK-NOV-NEXT: .LBB15_25: # %entry |
| ; CHECK-NOV-NEXT: mv a6, a7 |
| ; CHECK-NOV-NEXT: lui a7, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a6, .LBB15_9 |
| ; CHECK-NOV-NEXT: .LBB15_26: # %entry |
| ; CHECK-NOV-NEXT: lui a6, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a5, .LBB15_10 |
| ; CHECK-NOV-NEXT: .LBB15_27: # %entry |
| ; CHECK-NOV-NEXT: lui a5, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a4, .LBB15_11 |
| ; CHECK-NOV-NEXT: .LBB15_28: # %entry |
| ; CHECK-NOV-NEXT: lui a4, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a3, .LBB15_12 |
| ; CHECK-NOV-NEXT: .LBB15_29: # %entry |
| ; CHECK-NOV-NEXT: lui a3, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a2, .LBB15_13 |
| ; CHECK-NOV-NEXT: .LBB15_30: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a1, .LBB15_14 |
| ; CHECK-NOV-NEXT: .LBB15_31: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, s1, .LBB15_15 |
| ; CHECK-NOV-NEXT: .LBB15_32: # %entry |
| ; CHECK-NOV-NEXT: lui s1, 1048568 |
| ; CHECK-NOV-NEXT: bge a7, a0, .LBB15_16 |
| ; CHECK-NOV-NEXT: j .LBB15_17 |
| ; |
| ; CHECK-V-LABEL: stest_f16i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -80 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 80 |
| ; CHECK-V-NEXT: sd ra, 72(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 64(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s3, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s4, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s5, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s6, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: .cfi_offset s3, -40 |
| ; CHECK-V-NEXT: .cfi_offset s4, -48 |
| ; CHECK-V-NEXT: .cfi_offset s5, -56 |
| ; CHECK-V-NEXT: .cfi_offset s6, -64 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xd0, 0x00, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 80 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu s3, 24(a0) |
| ; CHECK-V-NEXT: lhu s4, 32(a0) |
| ; CHECK-V-NEXT: lhu s5, 40(a0) |
| ; CHECK-V-NEXT: lhu s6, 48(a0) |
| ; CHECK-V-NEXT: lhu a0, 56(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 2 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v10, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 2 |
| ; CHECK-V-NEXT: vsetivli zero, 8, e32, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 4 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, m1, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 72(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 64(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s3, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s4, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s5, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s6, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 80 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <8 x half> %x to <8 x i32> |
| %0 = icmp slt <8 x i32> %conv, <i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767> |
| %spec.store.select = select <8 x i1> %0, <8 x i32> %conv, <8 x i32> <i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767> |
| %1 = icmp sgt <8 x i32> %spec.store.select, <i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768> |
| %spec.store.select7 = select <8 x i1> %1, <8 x i32> %spec.store.select, <8 x i32> <i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768> |
| %conv6 = trunc <8 x i32> %spec.store.select7 to <8 x i16> |
| ret <8 x i16> %conv6 |
| } |
| |
| define <8 x i16> @utesth_f16i16(<8 x half> %x) { |
| ; CHECK-NOV-LABEL: utesth_f16i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -128 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 128 |
| ; CHECK-NOV-NEXT: sd ra, 120(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 112(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 104(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 96(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 88(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s4, 80(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s5, 72(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s6, 64(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s7, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs4, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs5, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs6, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset s4, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset s5, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset s6, -64 |
| ; CHECK-NOV-NEXT: .cfi_offset s7, -72 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -80 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -88 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -96 |
| ; CHECK-NOV-NEXT: .cfi_offset fs3, -104 |
| ; CHECK-NOV-NEXT: .cfi_offset fs4, -112 |
| ; CHECK-NOV-NEXT: .cfi_offset fs5, -120 |
| ; CHECK-NOV-NEXT: .cfi_offset fs6, -128 |
| ; CHECK-NOV-NEXT: lhu s1, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 56(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 48(a1) |
| ; CHECK-NOV-NEXT: lhu s4, 40(a1) |
| ; CHECK-NOV-NEXT: lhu s5, 32(a1) |
| ; CHECK-NOV-NEXT: lhu s6, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s7, 16(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 8(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs6, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s7 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs5, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs4, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs3, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s s1, fs6, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a1, 16 |
| ; CHECK-NOV-NEXT: addiw a1, a1, -1 |
| ; CHECK-NOV-NEXT: bgeu a0, a1, .LBB16_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fs5, rtz |
| ; CHECK-NOV-NEXT: bgeu s1, a1, .LBB16_11 |
| ; CHECK-NOV-NEXT: .LBB16_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a3, fs4, rtz |
| ; CHECK-NOV-NEXT: bgeu a2, a1, .LBB16_12 |
| ; CHECK-NOV-NEXT: .LBB16_3: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a4, fs3, rtz |
| ; CHECK-NOV-NEXT: bgeu a3, a1, .LBB16_13 |
| ; CHECK-NOV-NEXT: .LBB16_4: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a5, fs2, rtz |
| ; CHECK-NOV-NEXT: bgeu a4, a1, .LBB16_14 |
| ; CHECK-NOV-NEXT: .LBB16_5: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a6, fs1, rtz |
| ; CHECK-NOV-NEXT: bgeu a5, a1, .LBB16_15 |
| ; CHECK-NOV-NEXT: .LBB16_6: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a7, fs0, rtz |
| ; CHECK-NOV-NEXT: bgeu a6, a1, .LBB16_16 |
| ; CHECK-NOV-NEXT: .LBB16_7: # %entry |
| ; CHECK-NOV-NEXT: bltu a7, a1, .LBB16_9 |
| ; CHECK-NOV-NEXT: .LBB16_8: # %entry |
| ; CHECK-NOV-NEXT: mv a7, a1 |
| ; CHECK-NOV-NEXT: .LBB16_9: # %entry |
| ; CHECK-NOV-NEXT: sh a7, 14(s0) |
| ; CHECK-NOV-NEXT: sh a6, 12(s0) |
| ; CHECK-NOV-NEXT: sh a5, 10(s0) |
| ; CHECK-NOV-NEXT: sh a4, 8(s0) |
| ; CHECK-NOV-NEXT: sh a3, 6(s0) |
| ; CHECK-NOV-NEXT: sh a2, 4(s0) |
| ; CHECK-NOV-NEXT: sh s1, 2(s0) |
| ; CHECK-NOV-NEXT: sh a0, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 120(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 112(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 104(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 96(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 88(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s4, 80(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s5, 72(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s6, 64(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s7, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs4, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs5, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs6, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 128 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB16_10: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fs5, rtz |
| ; CHECK-NOV-NEXT: bltu s1, a1, .LBB16_2 |
| ; CHECK-NOV-NEXT: .LBB16_11: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a3, fs4, rtz |
| ; CHECK-NOV-NEXT: bltu a2, a1, .LBB16_3 |
| ; CHECK-NOV-NEXT: .LBB16_12: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a4, fs3, rtz |
| ; CHECK-NOV-NEXT: bltu a3, a1, .LBB16_4 |
| ; CHECK-NOV-NEXT: .LBB16_13: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a5, fs2, rtz |
| ; CHECK-NOV-NEXT: bltu a4, a1, .LBB16_5 |
| ; CHECK-NOV-NEXT: .LBB16_14: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a6, fs1, rtz |
| ; CHECK-NOV-NEXT: bltu a5, a1, .LBB16_6 |
| ; CHECK-NOV-NEXT: .LBB16_15: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a7, fs0, rtz |
| ; CHECK-NOV-NEXT: bltu a6, a1, .LBB16_7 |
| ; CHECK-NOV-NEXT: .LBB16_16: # %entry |
| ; CHECK-NOV-NEXT: mv a6, a1 |
| ; CHECK-NOV-NEXT: bgeu a7, a1, .LBB16_8 |
| ; CHECK-NOV-NEXT: j .LBB16_9 |
| ; |
| ; CHECK-V-LABEL: utesth_f16i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -80 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 80 |
| ; CHECK-V-NEXT: sd ra, 72(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 64(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s3, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s4, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s5, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s6, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: .cfi_offset s3, -40 |
| ; CHECK-V-NEXT: .cfi_offset s4, -48 |
| ; CHECK-V-NEXT: .cfi_offset s5, -56 |
| ; CHECK-V-NEXT: .cfi_offset s6, -64 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xd0, 0x00, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 80 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu s3, 24(a0) |
| ; CHECK-V-NEXT: lhu s4, 32(a0) |
| ; CHECK-V-NEXT: lhu s5, 40(a0) |
| ; CHECK-V-NEXT: lhu s6, 48(a0) |
| ; CHECK-V-NEXT: lhu a0, 56(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 2 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v10, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 2 |
| ; CHECK-V-NEXT: vsetivli zero, 8, e32, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 4 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, m1, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 72(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 64(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s3, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s4, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s5, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s6, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 80 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <8 x half> %x to <8 x i32> |
| %0 = icmp ult <8 x i32> %conv, <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535> |
| %spec.store.select = select <8 x i1> %0, <8 x i32> %conv, <8 x i32> <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535> |
| %conv6 = trunc <8 x i32> %spec.store.select to <8 x i16> |
| ret <8 x i16> %conv6 |
| } |
| |
| define <8 x i16> @ustest_f16i16(<8 x half> %x) { |
| ; CHECK-NOV-LABEL: ustest_f16i16: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -128 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 128 |
| ; CHECK-NOV-NEXT: sd ra, 120(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 112(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 104(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 96(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 88(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s4, 80(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s5, 72(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s6, 64(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s7, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs4, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs5, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs6, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset s4, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset s5, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset s6, -64 |
| ; CHECK-NOV-NEXT: .cfi_offset s7, -72 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -80 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -88 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -96 |
| ; CHECK-NOV-NEXT: .cfi_offset fs3, -104 |
| ; CHECK-NOV-NEXT: .cfi_offset fs4, -112 |
| ; CHECK-NOV-NEXT: .cfi_offset fs5, -120 |
| ; CHECK-NOV-NEXT: .cfi_offset fs6, -128 |
| ; CHECK-NOV-NEXT: lhu s1, 56(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 8(a1) |
| ; CHECK-NOV-NEXT: lhu s4, 16(a1) |
| ; CHECK-NOV-NEXT: lhu s5, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s6, 32(a1) |
| ; CHECK-NOV-NEXT: lhu s7, 40(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 48(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs6, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s7 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs5, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs4, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs3, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.l.s s1, fs6, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a3, 16 |
| ; CHECK-NOV-NEXT: addiw a3, a3, -1 |
| ; CHECK-NOV-NEXT: bge a0, a3, .LBB17_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs5, rtz |
| ; CHECK-NOV-NEXT: bge s1, a3, .LBB17_11 |
| ; CHECK-NOV-NEXT: .LBB17_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs4, rtz |
| ; CHECK-NOV-NEXT: bge a1, a3, .LBB17_12 |
| ; CHECK-NOV-NEXT: .LBB17_3: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fs3, rtz |
| ; CHECK-NOV-NEXT: bge a2, a3, .LBB17_13 |
| ; CHECK-NOV-NEXT: .LBB17_4: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fs2, rtz |
| ; CHECK-NOV-NEXT: bge a4, a3, .LBB17_14 |
| ; CHECK-NOV-NEXT: .LBB17_5: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a6, fs1, rtz |
| ; CHECK-NOV-NEXT: bge a5, a3, .LBB17_15 |
| ; CHECK-NOV-NEXT: .LBB17_6: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a7, fs0, rtz |
| ; CHECK-NOV-NEXT: bge a6, a3, .LBB17_16 |
| ; CHECK-NOV-NEXT: .LBB17_7: # %entry |
| ; CHECK-NOV-NEXT: blt a7, a3, .LBB17_9 |
| ; CHECK-NOV-NEXT: .LBB17_8: # %entry |
| ; CHECK-NOV-NEXT: mv a7, a3 |
| ; CHECK-NOV-NEXT: .LBB17_9: # %entry |
| ; CHECK-NOV-NEXT: sgtz a3, a0 |
| ; CHECK-NOV-NEXT: sgtz t0, s1 |
| ; CHECK-NOV-NEXT: sgtz t1, a1 |
| ; CHECK-NOV-NEXT: sgtz t2, a2 |
| ; CHECK-NOV-NEXT: sgtz t3, a4 |
| ; CHECK-NOV-NEXT: sgtz t4, a5 |
| ; CHECK-NOV-NEXT: sgtz t5, a6 |
| ; CHECK-NOV-NEXT: sgtz t6, a7 |
| ; CHECK-NOV-NEXT: negw t6, t6 |
| ; CHECK-NOV-NEXT: and a7, t6, a7 |
| ; CHECK-NOV-NEXT: negw t5, t5 |
| ; CHECK-NOV-NEXT: and a6, t5, a6 |
| ; CHECK-NOV-NEXT: negw t4, t4 |
| ; CHECK-NOV-NEXT: and a5, t4, a5 |
| ; CHECK-NOV-NEXT: negw t3, t3 |
| ; CHECK-NOV-NEXT: and a4, t3, a4 |
| ; CHECK-NOV-NEXT: negw t2, t2 |
| ; CHECK-NOV-NEXT: and a2, t2, a2 |
| ; CHECK-NOV-NEXT: negw t1, t1 |
| ; CHECK-NOV-NEXT: and a1, t1, a1 |
| ; CHECK-NOV-NEXT: negw t0, t0 |
| ; CHECK-NOV-NEXT: and t0, t0, s1 |
| ; CHECK-NOV-NEXT: negw a3, a3 |
| ; CHECK-NOV-NEXT: and a0, a3, a0 |
| ; CHECK-NOV-NEXT: sh a0, 14(s0) |
| ; CHECK-NOV-NEXT: sh t0, 12(s0) |
| ; CHECK-NOV-NEXT: sh a1, 10(s0) |
| ; CHECK-NOV-NEXT: sh a2, 8(s0) |
| ; CHECK-NOV-NEXT: sh a4, 6(s0) |
| ; CHECK-NOV-NEXT: sh a5, 4(s0) |
| ; CHECK-NOV-NEXT: sh a6, 2(s0) |
| ; CHECK-NOV-NEXT: sh a7, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 120(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 112(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 104(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 96(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 88(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s4, 80(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s5, 72(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s6, 64(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s7, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs4, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs5, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs6, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 128 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB17_10: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs5, rtz |
| ; CHECK-NOV-NEXT: blt s1, a3, .LBB17_2 |
| ; CHECK-NOV-NEXT: .LBB17_11: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs4, rtz |
| ; CHECK-NOV-NEXT: blt a1, a3, .LBB17_3 |
| ; CHECK-NOV-NEXT: .LBB17_12: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fs3, rtz |
| ; CHECK-NOV-NEXT: blt a2, a3, .LBB17_4 |
| ; CHECK-NOV-NEXT: .LBB17_13: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fs2, rtz |
| ; CHECK-NOV-NEXT: blt a4, a3, .LBB17_5 |
| ; CHECK-NOV-NEXT: .LBB17_14: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a6, fs1, rtz |
| ; CHECK-NOV-NEXT: blt a5, a3, .LBB17_6 |
| ; CHECK-NOV-NEXT: .LBB17_15: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a7, fs0, rtz |
| ; CHECK-NOV-NEXT: blt a6, a3, .LBB17_7 |
| ; CHECK-NOV-NEXT: .LBB17_16: # %entry |
| ; CHECK-NOV-NEXT: mv a6, a3 |
| ; CHECK-NOV-NEXT: bge a7, a3, .LBB17_8 |
| ; CHECK-NOV-NEXT: j .LBB17_9 |
| ; |
| ; CHECK-V-LABEL: ustest_f16i16: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -80 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 80 |
| ; CHECK-V-NEXT: sd ra, 72(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 64(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s3, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s4, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s5, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s6, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: .cfi_offset s3, -40 |
| ; CHECK-V-NEXT: .cfi_offset s4, -48 |
| ; CHECK-V-NEXT: .cfi_offset s5, -56 |
| ; CHECK-V-NEXT: .cfi_offset s6, -64 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xd0, 0x00, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 80 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu s3, 24(a0) |
| ; CHECK-V-NEXT: lhu s4, 32(a0) |
| ; CHECK-V-NEXT: lhu s5, 40(a0) |
| ; CHECK-V-NEXT: lhu s6, 48(a0) |
| ; CHECK-V-NEXT: lhu a0, 56(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 2 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 2 |
| ; CHECK-V-NEXT: vsetivli zero, 8, e32, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v10, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v10, 4 |
| ; CHECK-V-NEXT: lui a0, 16 |
| ; CHECK-V-NEXT: addi a0, a0, -1 |
| ; CHECK-V-NEXT: vmin.vx v8, v8, a0 |
| ; CHECK-V-NEXT: vmax.vx v10, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, m1, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 72(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 64(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s3, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s4, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s5, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s6, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 80 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <8 x half> %x to <8 x i32> |
| %0 = icmp slt <8 x i32> %conv, <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535> |
| %spec.store.select = select <8 x i1> %0, <8 x i32> %conv, <8 x i32> <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535> |
| %1 = icmp sgt <8 x i32> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <8 x i1> %1, <8 x i32> %spec.store.select, <8 x i32> zeroinitializer |
| %conv6 = trunc <8 x i32> %spec.store.select7 to <8 x i16> |
| ret <8 x i16> %conv6 |
| } |
| |
| ; i64 saturate |
| |
| define <2 x i64> @stest_f64i64(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: stest_f64i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.d fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixdfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.d fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixdfti |
| ; CHECK-NOV-NEXT: mv a2, a0 |
| ; CHECK-NOV-NEXT: li a0, -1 |
| ; CHECK-NOV-NEXT: srli a3, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB18_3 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: slti a4, a1, 0 |
| ; CHECK-NOV-NEXT: bnez s1, .LBB18_4 |
| ; CHECK-NOV-NEXT: .LBB18_2: |
| ; CHECK-NOV-NEXT: sltu a5, s0, a3 |
| ; CHECK-NOV-NEXT: beqz a5, .LBB18_5 |
| ; CHECK-NOV-NEXT: j .LBB18_6 |
| ; CHECK-NOV-NEXT: .LBB18_3: |
| ; CHECK-NOV-NEXT: sltu a4, a2, a3 |
| ; CHECK-NOV-NEXT: beqz s1, .LBB18_2 |
| ; CHECK-NOV-NEXT: .LBB18_4: # %entry |
| ; CHECK-NOV-NEXT: slti a5, s1, 0 |
| ; CHECK-NOV-NEXT: bnez a5, .LBB18_6 |
| ; CHECK-NOV-NEXT: .LBB18_5: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB18_6: # %entry |
| ; CHECK-NOV-NEXT: neg a6, a5 |
| ; CHECK-NOV-NEXT: neg a5, a4 |
| ; CHECK-NOV-NEXT: and a5, a5, a1 |
| ; CHECK-NOV-NEXT: bnez a4, .LBB18_8 |
| ; CHECK-NOV-NEXT: # %bb.7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB18_8: # %entry |
| ; CHECK-NOV-NEXT: and a4, a6, s1 |
| ; CHECK-NOV-NEXT: slli a1, a0, 63 |
| ; CHECK-NOV-NEXT: beq a5, a0, .LBB18_11 |
| ; CHECK-NOV-NEXT: # %bb.9: # %entry |
| ; CHECK-NOV-NEXT: slti a3, a5, 0 |
| ; CHECK-NOV-NEXT: xori a3, a3, 1 |
| ; CHECK-NOV-NEXT: bne a4, a0, .LBB18_12 |
| ; CHECK-NOV-NEXT: .LBB18_10: |
| ; CHECK-NOV-NEXT: sltu a0, a1, s0 |
| ; CHECK-NOV-NEXT: beqz a0, .LBB18_13 |
| ; CHECK-NOV-NEXT: j .LBB18_14 |
| ; CHECK-NOV-NEXT: .LBB18_11: |
| ; CHECK-NOV-NEXT: sltu a3, a1, a2 |
| ; CHECK-NOV-NEXT: beq a4, a0, .LBB18_10 |
| ; CHECK-NOV-NEXT: .LBB18_12: # %entry |
| ; CHECK-NOV-NEXT: slti a0, a4, 0 |
| ; CHECK-NOV-NEXT: xori a0, a0, 1 |
| ; CHECK-NOV-NEXT: bnez a0, .LBB18_14 |
| ; CHECK-NOV-NEXT: .LBB18_13: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a1 |
| ; CHECK-NOV-NEXT: .LBB18_14: # %entry |
| ; CHECK-NOV-NEXT: bnez a3, .LBB18_16 |
| ; CHECK-NOV-NEXT: # %bb.15: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: .LBB18_16: # %entry |
| ; CHECK-NOV-NEXT: mv a0, s0 |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f64i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixdfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixdfti |
| ; CHECK-V-NEXT: li a2, -1 |
| ; CHECK-V-NEXT: srli a3, a2, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB18_3 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 0 |
| ; CHECK-V-NEXT: bnez s1, .LBB18_4 |
| ; CHECK-V-NEXT: .LBB18_2: |
| ; CHECK-V-NEXT: sltu a5, s0, a3 |
| ; CHECK-V-NEXT: beqz a5, .LBB18_5 |
| ; CHECK-V-NEXT: j .LBB18_6 |
| ; CHECK-V-NEXT: .LBB18_3: |
| ; CHECK-V-NEXT: sltu a4, a0, a3 |
| ; CHECK-V-NEXT: beqz s1, .LBB18_2 |
| ; CHECK-V-NEXT: .LBB18_4: # %entry |
| ; CHECK-V-NEXT: slti a5, s1, 0 |
| ; CHECK-V-NEXT: bnez a5, .LBB18_6 |
| ; CHECK-V-NEXT: .LBB18_5: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB18_6: # %entry |
| ; CHECK-V-NEXT: neg a6, a5 |
| ; CHECK-V-NEXT: neg a5, a4 |
| ; CHECK-V-NEXT: and a5, a5, a1 |
| ; CHECK-V-NEXT: bnez a4, .LBB18_8 |
| ; CHECK-V-NEXT: # %bb.7: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB18_8: # %entry |
| ; CHECK-V-NEXT: and a4, a6, s1 |
| ; CHECK-V-NEXT: slli a1, a2, 63 |
| ; CHECK-V-NEXT: beq a5, a2, .LBB18_11 |
| ; CHECK-V-NEXT: # %bb.9: # %entry |
| ; CHECK-V-NEXT: slti a3, a5, 0 |
| ; CHECK-V-NEXT: xori a3, a3, 1 |
| ; CHECK-V-NEXT: bne a4, a2, .LBB18_12 |
| ; CHECK-V-NEXT: .LBB18_10: |
| ; CHECK-V-NEXT: sltu a2, a1, s0 |
| ; CHECK-V-NEXT: beqz a2, .LBB18_13 |
| ; CHECK-V-NEXT: j .LBB18_14 |
| ; CHECK-V-NEXT: .LBB18_11: |
| ; CHECK-V-NEXT: sltu a3, a1, a0 |
| ; CHECK-V-NEXT: beq a4, a2, .LBB18_10 |
| ; CHECK-V-NEXT: .LBB18_12: # %entry |
| ; CHECK-V-NEXT: slti a2, a4, 0 |
| ; CHECK-V-NEXT: xori a2, a2, 1 |
| ; CHECK-V-NEXT: bnez a2, .LBB18_14 |
| ; CHECK-V-NEXT: .LBB18_13: # %entry |
| ; CHECK-V-NEXT: mv s0, a1 |
| ; CHECK-V-NEXT: .LBB18_14: # %entry |
| ; CHECK-V-NEXT: bnez a3, .LBB18_16 |
| ; CHECK-V-NEXT: # %bb.15: # %entry |
| ; CHECK-V-NEXT: mv a0, a1 |
| ; CHECK-V-NEXT: .LBB18_16: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, s0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i128> |
| %0 = icmp slt <2 x i128> %conv, <i128 9223372036854775807, i128 9223372036854775807> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 9223372036854775807, i128 9223372036854775807> |
| %1 = icmp sgt <2 x i128> %spec.store.select, <i128 -9223372036854775808, i128 -9223372036854775808> |
| %spec.store.select7 = select <2 x i1> %1, <2 x i128> %spec.store.select, <2 x i128> <i128 -9223372036854775808, i128 -9223372036854775808> |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @utest_f64i64(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: utest_f64i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.d fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixunsdfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.d fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixunsdfti |
| ; CHECK-NOV-NEXT: snez a1, a1 |
| ; CHECK-NOV-NEXT: snez a2, s1 |
| ; CHECK-NOV-NEXT: addi a2, a2, -1 |
| ; CHECK-NOV-NEXT: and a2, a2, s0 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a1, a1, a0 |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utest_f64i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixunsdfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixunsdfti |
| ; CHECK-V-NEXT: snez a1, a1 |
| ; CHECK-V-NEXT: snez a2, s1 |
| ; CHECK-V-NEXT: addi a2, a2, -1 |
| ; CHECK-V-NEXT: and a2, a2, s0 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, a2 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x double> %x to <2 x i128> |
| %0 = icmp ult <2 x i128> %conv, <i128 18446744073709551616, i128 18446744073709551616> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616> |
| %conv6 = trunc <2 x i128> %spec.store.select to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @ustest_f64i64(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: ustest_f64i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.d fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixdfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.d fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixdfti |
| ; CHECK-NOV-NEXT: mv a2, s1 |
| ; CHECK-NOV-NEXT: blez s1, .LBB20_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: li a2, 1 |
| ; CHECK-NOV-NEXT: .LBB20_2: # %entry |
| ; CHECK-NOV-NEXT: slti a3, a1, 1 |
| ; CHECK-NOV-NEXT: slti a4, s1, 1 |
| ; CHECK-NOV-NEXT: blez a1, .LBB20_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: li a1, 1 |
| ; CHECK-NOV-NEXT: .LBB20_4: # %entry |
| ; CHECK-NOV-NEXT: neg a4, a4 |
| ; CHECK-NOV-NEXT: neg a3, a3 |
| ; CHECK-NOV-NEXT: and a3, a3, a0 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB20_7 |
| ; CHECK-NOV-NEXT: # %bb.5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a1, a1 |
| ; CHECK-NOV-NEXT: and a4, a4, s0 |
| ; CHECK-NOV-NEXT: bnez a2, .LBB20_8 |
| ; CHECK-NOV-NEXT: .LBB20_6: |
| ; CHECK-NOV-NEXT: snez a0, a4 |
| ; CHECK-NOV-NEXT: j .LBB20_9 |
| ; CHECK-NOV-NEXT: .LBB20_7: |
| ; CHECK-NOV-NEXT: snez a1, a3 |
| ; CHECK-NOV-NEXT: and a4, a4, s0 |
| ; CHECK-NOV-NEXT: beqz a2, .LBB20_6 |
| ; CHECK-NOV-NEXT: .LBB20_8: # %entry |
| ; CHECK-NOV-NEXT: sgtz a0, a2 |
| ; CHECK-NOV-NEXT: .LBB20_9: # %entry |
| ; CHECK-NOV-NEXT: neg a0, a0 |
| ; CHECK-NOV-NEXT: and a0, a0, a4 |
| ; CHECK-NOV-NEXT: neg a1, a1 |
| ; CHECK-NOV-NEXT: and a1, a1, a3 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f64i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixdfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixdfti |
| ; CHECK-V-NEXT: mv a2, s1 |
| ; CHECK-V-NEXT: blez s1, .LBB20_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: li a2, 1 |
| ; CHECK-V-NEXT: .LBB20_2: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 1 |
| ; CHECK-V-NEXT: slti a3, s1, 1 |
| ; CHECK-V-NEXT: blez a1, .LBB20_4 |
| ; CHECK-V-NEXT: # %bb.3: # %entry |
| ; CHECK-V-NEXT: li a1, 1 |
| ; CHECK-V-NEXT: .LBB20_4: # %entry |
| ; CHECK-V-NEXT: neg a3, a3 |
| ; CHECK-V-NEXT: neg a4, a4 |
| ; CHECK-V-NEXT: and a0, a4, a0 |
| ; CHECK-V-NEXT: beqz a1, .LBB20_7 |
| ; CHECK-V-NEXT: # %bb.5: # %entry |
| ; CHECK-V-NEXT: sgtz a1, a1 |
| ; CHECK-V-NEXT: and a3, a3, s0 |
| ; CHECK-V-NEXT: bnez a2, .LBB20_8 |
| ; CHECK-V-NEXT: .LBB20_6: |
| ; CHECK-V-NEXT: snez a2, a3 |
| ; CHECK-V-NEXT: j .LBB20_9 |
| ; CHECK-V-NEXT: .LBB20_7: |
| ; CHECK-V-NEXT: snez a1, a0 |
| ; CHECK-V-NEXT: and a3, a3, s0 |
| ; CHECK-V-NEXT: beqz a2, .LBB20_6 |
| ; CHECK-V-NEXT: .LBB20_8: # %entry |
| ; CHECK-V-NEXT: sgtz a2, a2 |
| ; CHECK-V-NEXT: .LBB20_9: # %entry |
| ; CHECK-V-NEXT: neg a2, a2 |
| ; CHECK-V-NEXT: and a2, a2, a3 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, a2 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i128> |
| %0 = icmp slt <2 x i128> %conv, <i128 18446744073709551616, i128 18446744073709551616> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616> |
| %1 = icmp sgt <2 x i128> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <2 x i1> %1, <2 x i128> %spec.store.select, <2 x i128> zeroinitializer |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @stest_f32i64(<2 x float> %x) { |
| ; CHECK-NOV-LABEL: stest_f32i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.s fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv a2, a0 |
| ; CHECK-NOV-NEXT: li a0, -1 |
| ; CHECK-NOV-NEXT: srli a3, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB21_3 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: slti a4, a1, 0 |
| ; CHECK-NOV-NEXT: bnez s1, .LBB21_4 |
| ; CHECK-NOV-NEXT: .LBB21_2: |
| ; CHECK-NOV-NEXT: sltu a5, s0, a3 |
| ; CHECK-NOV-NEXT: beqz a5, .LBB21_5 |
| ; CHECK-NOV-NEXT: j .LBB21_6 |
| ; CHECK-NOV-NEXT: .LBB21_3: |
| ; CHECK-NOV-NEXT: sltu a4, a2, a3 |
| ; CHECK-NOV-NEXT: beqz s1, .LBB21_2 |
| ; CHECK-NOV-NEXT: .LBB21_4: # %entry |
| ; CHECK-NOV-NEXT: slti a5, s1, 0 |
| ; CHECK-NOV-NEXT: bnez a5, .LBB21_6 |
| ; CHECK-NOV-NEXT: .LBB21_5: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB21_6: # %entry |
| ; CHECK-NOV-NEXT: neg a6, a5 |
| ; CHECK-NOV-NEXT: neg a5, a4 |
| ; CHECK-NOV-NEXT: and a5, a5, a1 |
| ; CHECK-NOV-NEXT: bnez a4, .LBB21_8 |
| ; CHECK-NOV-NEXT: # %bb.7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB21_8: # %entry |
| ; CHECK-NOV-NEXT: and a4, a6, s1 |
| ; CHECK-NOV-NEXT: slli a1, a0, 63 |
| ; CHECK-NOV-NEXT: beq a5, a0, .LBB21_11 |
| ; CHECK-NOV-NEXT: # %bb.9: # %entry |
| ; CHECK-NOV-NEXT: slti a3, a5, 0 |
| ; CHECK-NOV-NEXT: xori a3, a3, 1 |
| ; CHECK-NOV-NEXT: bne a4, a0, .LBB21_12 |
| ; CHECK-NOV-NEXT: .LBB21_10: |
| ; CHECK-NOV-NEXT: sltu a0, a1, s0 |
| ; CHECK-NOV-NEXT: beqz a0, .LBB21_13 |
| ; CHECK-NOV-NEXT: j .LBB21_14 |
| ; CHECK-NOV-NEXT: .LBB21_11: |
| ; CHECK-NOV-NEXT: sltu a3, a1, a2 |
| ; CHECK-NOV-NEXT: beq a4, a0, .LBB21_10 |
| ; CHECK-NOV-NEXT: .LBB21_12: # %entry |
| ; CHECK-NOV-NEXT: slti a0, a4, 0 |
| ; CHECK-NOV-NEXT: xori a0, a0, 1 |
| ; CHECK-NOV-NEXT: bnez a0, .LBB21_14 |
| ; CHECK-NOV-NEXT: .LBB21_13: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a1 |
| ; CHECK-NOV-NEXT: .LBB21_14: # %entry |
| ; CHECK-NOV-NEXT: bnez a3, .LBB21_16 |
| ; CHECK-NOV-NEXT: # %bb.15: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: .LBB21_16: # %entry |
| ; CHECK-NOV-NEXT: mv a0, s0 |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f32i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: li a2, -1 |
| ; CHECK-V-NEXT: srli a3, a2, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB21_3 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 0 |
| ; CHECK-V-NEXT: bnez s1, .LBB21_4 |
| ; CHECK-V-NEXT: .LBB21_2: |
| ; CHECK-V-NEXT: sltu a5, s0, a3 |
| ; CHECK-V-NEXT: beqz a5, .LBB21_5 |
| ; CHECK-V-NEXT: j .LBB21_6 |
| ; CHECK-V-NEXT: .LBB21_3: |
| ; CHECK-V-NEXT: sltu a4, a0, a3 |
| ; CHECK-V-NEXT: beqz s1, .LBB21_2 |
| ; CHECK-V-NEXT: .LBB21_4: # %entry |
| ; CHECK-V-NEXT: slti a5, s1, 0 |
| ; CHECK-V-NEXT: bnez a5, .LBB21_6 |
| ; CHECK-V-NEXT: .LBB21_5: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB21_6: # %entry |
| ; CHECK-V-NEXT: neg a6, a5 |
| ; CHECK-V-NEXT: neg a5, a4 |
| ; CHECK-V-NEXT: and a5, a5, a1 |
| ; CHECK-V-NEXT: bnez a4, .LBB21_8 |
| ; CHECK-V-NEXT: # %bb.7: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB21_8: # %entry |
| ; CHECK-V-NEXT: and a4, a6, s1 |
| ; CHECK-V-NEXT: slli a1, a2, 63 |
| ; CHECK-V-NEXT: beq a5, a2, .LBB21_11 |
| ; CHECK-V-NEXT: # %bb.9: # %entry |
| ; CHECK-V-NEXT: slti a3, a5, 0 |
| ; CHECK-V-NEXT: xori a3, a3, 1 |
| ; CHECK-V-NEXT: bne a4, a2, .LBB21_12 |
| ; CHECK-V-NEXT: .LBB21_10: |
| ; CHECK-V-NEXT: sltu a2, a1, s0 |
| ; CHECK-V-NEXT: beqz a2, .LBB21_13 |
| ; CHECK-V-NEXT: j .LBB21_14 |
| ; CHECK-V-NEXT: .LBB21_11: |
| ; CHECK-V-NEXT: sltu a3, a1, a0 |
| ; CHECK-V-NEXT: beq a4, a2, .LBB21_10 |
| ; CHECK-V-NEXT: .LBB21_12: # %entry |
| ; CHECK-V-NEXT: slti a2, a4, 0 |
| ; CHECK-V-NEXT: xori a2, a2, 1 |
| ; CHECK-V-NEXT: bnez a2, .LBB21_14 |
| ; CHECK-V-NEXT: .LBB21_13: # %entry |
| ; CHECK-V-NEXT: mv s0, a1 |
| ; CHECK-V-NEXT: .LBB21_14: # %entry |
| ; CHECK-V-NEXT: bnez a3, .LBB21_16 |
| ; CHECK-V-NEXT: # %bb.15: # %entry |
| ; CHECK-V-NEXT: mv a0, a1 |
| ; CHECK-V-NEXT: .LBB21_16: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, s0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x float> %x to <2 x i128> |
| %0 = icmp slt <2 x i128> %conv, <i128 9223372036854775807, i128 9223372036854775807> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 9223372036854775807, i128 9223372036854775807> |
| %1 = icmp sgt <2 x i128> %spec.store.select, <i128 -9223372036854775808, i128 -9223372036854775808> |
| %spec.store.select7 = select <2 x i1> %1, <2 x i128> %spec.store.select, <2 x i128> <i128 -9223372036854775808, i128 -9223372036854775808> |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @utest_f32i64(<2 x float> %x) { |
| ; CHECK-NOV-LABEL: utest_f32i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixunssfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.s fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixunssfti |
| ; CHECK-NOV-NEXT: snez a1, a1 |
| ; CHECK-NOV-NEXT: snez a2, s1 |
| ; CHECK-NOV-NEXT: addi a2, a2, -1 |
| ; CHECK-NOV-NEXT: and a2, a2, s0 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a1, a1, a0 |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utest_f32i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixunssfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixunssfti |
| ; CHECK-V-NEXT: snez a1, a1 |
| ; CHECK-V-NEXT: snez a2, s1 |
| ; CHECK-V-NEXT: addi a2, a2, -1 |
| ; CHECK-V-NEXT: and a2, a2, s0 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, a2 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x float> %x to <2 x i128> |
| %0 = icmp ult <2 x i128> %conv, <i128 18446744073709551616, i128 18446744073709551616> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616> |
| %conv6 = trunc <2 x i128> %spec.store.select to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @ustest_f32i64(<2 x float> %x) { |
| ; CHECK-NOV-LABEL: ustest_f32i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.s fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv a2, s1 |
| ; CHECK-NOV-NEXT: blez s1, .LBB23_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: li a2, 1 |
| ; CHECK-NOV-NEXT: .LBB23_2: # %entry |
| ; CHECK-NOV-NEXT: slti a3, a1, 1 |
| ; CHECK-NOV-NEXT: slti a4, s1, 1 |
| ; CHECK-NOV-NEXT: blez a1, .LBB23_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: li a1, 1 |
| ; CHECK-NOV-NEXT: .LBB23_4: # %entry |
| ; CHECK-NOV-NEXT: neg a4, a4 |
| ; CHECK-NOV-NEXT: neg a3, a3 |
| ; CHECK-NOV-NEXT: and a3, a3, a0 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB23_7 |
| ; CHECK-NOV-NEXT: # %bb.5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a1, a1 |
| ; CHECK-NOV-NEXT: and a4, a4, s0 |
| ; CHECK-NOV-NEXT: bnez a2, .LBB23_8 |
| ; CHECK-NOV-NEXT: .LBB23_6: |
| ; CHECK-NOV-NEXT: snez a0, a4 |
| ; CHECK-NOV-NEXT: j .LBB23_9 |
| ; CHECK-NOV-NEXT: .LBB23_7: |
| ; CHECK-NOV-NEXT: snez a1, a3 |
| ; CHECK-NOV-NEXT: and a4, a4, s0 |
| ; CHECK-NOV-NEXT: beqz a2, .LBB23_6 |
| ; CHECK-NOV-NEXT: .LBB23_8: # %entry |
| ; CHECK-NOV-NEXT: sgtz a0, a2 |
| ; CHECK-NOV-NEXT: .LBB23_9: # %entry |
| ; CHECK-NOV-NEXT: neg a0, a0 |
| ; CHECK-NOV-NEXT: and a0, a0, a4 |
| ; CHECK-NOV-NEXT: neg a1, a1 |
| ; CHECK-NOV-NEXT: and a1, a1, a3 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f32i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv a2, s1 |
| ; CHECK-V-NEXT: blez s1, .LBB23_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: li a2, 1 |
| ; CHECK-V-NEXT: .LBB23_2: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 1 |
| ; CHECK-V-NEXT: slti a3, s1, 1 |
| ; CHECK-V-NEXT: blez a1, .LBB23_4 |
| ; CHECK-V-NEXT: # %bb.3: # %entry |
| ; CHECK-V-NEXT: li a1, 1 |
| ; CHECK-V-NEXT: .LBB23_4: # %entry |
| ; CHECK-V-NEXT: neg a3, a3 |
| ; CHECK-V-NEXT: neg a4, a4 |
| ; CHECK-V-NEXT: and a0, a4, a0 |
| ; CHECK-V-NEXT: beqz a1, .LBB23_7 |
| ; CHECK-V-NEXT: # %bb.5: # %entry |
| ; CHECK-V-NEXT: sgtz a1, a1 |
| ; CHECK-V-NEXT: and a3, a3, s0 |
| ; CHECK-V-NEXT: bnez a2, .LBB23_8 |
| ; CHECK-V-NEXT: .LBB23_6: |
| ; CHECK-V-NEXT: snez a2, a3 |
| ; CHECK-V-NEXT: j .LBB23_9 |
| ; CHECK-V-NEXT: .LBB23_7: |
| ; CHECK-V-NEXT: snez a1, a0 |
| ; CHECK-V-NEXT: and a3, a3, s0 |
| ; CHECK-V-NEXT: beqz a2, .LBB23_6 |
| ; CHECK-V-NEXT: .LBB23_8: # %entry |
| ; CHECK-V-NEXT: sgtz a2, a2 |
| ; CHECK-V-NEXT: .LBB23_9: # %entry |
| ; CHECK-V-NEXT: neg a2, a2 |
| ; CHECK-V-NEXT: and a2, a2, a3 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, a2 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x float> %x to <2 x i128> |
| %0 = icmp slt <2 x i128> %conv, <i128 18446744073709551616, i128 18446744073709551616> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616> |
| %1 = icmp sgt <2 x i128> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <2 x i1> %1, <2 x i128> %spec.store.select, <2 x i128> zeroinitializer |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @stest_f16i64(<2 x half> %x) { |
| ; CHECK-NOV-LABEL: stest_f16i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: mv s2, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv a2, a0 |
| ; CHECK-NOV-NEXT: li a0, -1 |
| ; CHECK-NOV-NEXT: srli a3, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB24_3 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: slti a4, a1, 0 |
| ; CHECK-NOV-NEXT: bnez s1, .LBB24_4 |
| ; CHECK-NOV-NEXT: .LBB24_2: |
| ; CHECK-NOV-NEXT: sltu a5, s0, a3 |
| ; CHECK-NOV-NEXT: beqz a5, .LBB24_5 |
| ; CHECK-NOV-NEXT: j .LBB24_6 |
| ; CHECK-NOV-NEXT: .LBB24_3: |
| ; CHECK-NOV-NEXT: sltu a4, a2, a3 |
| ; CHECK-NOV-NEXT: beqz s1, .LBB24_2 |
| ; CHECK-NOV-NEXT: .LBB24_4: # %entry |
| ; CHECK-NOV-NEXT: slti a5, s1, 0 |
| ; CHECK-NOV-NEXT: bnez a5, .LBB24_6 |
| ; CHECK-NOV-NEXT: .LBB24_5: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB24_6: # %entry |
| ; CHECK-NOV-NEXT: neg a6, a5 |
| ; CHECK-NOV-NEXT: neg a5, a4 |
| ; CHECK-NOV-NEXT: and a5, a5, a1 |
| ; CHECK-NOV-NEXT: bnez a4, .LBB24_8 |
| ; CHECK-NOV-NEXT: # %bb.7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB24_8: # %entry |
| ; CHECK-NOV-NEXT: and a4, a6, s1 |
| ; CHECK-NOV-NEXT: slli a1, a0, 63 |
| ; CHECK-NOV-NEXT: beq a5, a0, .LBB24_11 |
| ; CHECK-NOV-NEXT: # %bb.9: # %entry |
| ; CHECK-NOV-NEXT: slti a3, a5, 0 |
| ; CHECK-NOV-NEXT: xori a3, a3, 1 |
| ; CHECK-NOV-NEXT: bne a4, a0, .LBB24_12 |
| ; CHECK-NOV-NEXT: .LBB24_10: |
| ; CHECK-NOV-NEXT: sltu a0, a1, s0 |
| ; CHECK-NOV-NEXT: beqz a0, .LBB24_13 |
| ; CHECK-NOV-NEXT: j .LBB24_14 |
| ; CHECK-NOV-NEXT: .LBB24_11: |
| ; CHECK-NOV-NEXT: sltu a3, a1, a2 |
| ; CHECK-NOV-NEXT: beq a4, a0, .LBB24_10 |
| ; CHECK-NOV-NEXT: .LBB24_12: # %entry |
| ; CHECK-NOV-NEXT: slti a0, a4, 0 |
| ; CHECK-NOV-NEXT: xori a0, a0, 1 |
| ; CHECK-NOV-NEXT: bnez a0, .LBB24_14 |
| ; CHECK-NOV-NEXT: .LBB24_13: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a1 |
| ; CHECK-NOV-NEXT: .LBB24_14: # %entry |
| ; CHECK-NOV-NEXT: bnez a3, .LBB24_16 |
| ; CHECK-NOV-NEXT: # %bb.15: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: .LBB24_16: # %entry |
| ; CHECK-NOV-NEXT: mv a0, s0 |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f16i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -32 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-V-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: mv s2, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: li a2, -1 |
| ; CHECK-V-NEXT: srli a3, a2, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB24_3 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 0 |
| ; CHECK-V-NEXT: bnez s1, .LBB24_4 |
| ; CHECK-V-NEXT: .LBB24_2: |
| ; CHECK-V-NEXT: sltu a5, s0, a3 |
| ; CHECK-V-NEXT: beqz a5, .LBB24_5 |
| ; CHECK-V-NEXT: j .LBB24_6 |
| ; CHECK-V-NEXT: .LBB24_3: |
| ; CHECK-V-NEXT: sltu a4, a0, a3 |
| ; CHECK-V-NEXT: beqz s1, .LBB24_2 |
| ; CHECK-V-NEXT: .LBB24_4: # %entry |
| ; CHECK-V-NEXT: slti a5, s1, 0 |
| ; CHECK-V-NEXT: bnez a5, .LBB24_6 |
| ; CHECK-V-NEXT: .LBB24_5: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB24_6: # %entry |
| ; CHECK-V-NEXT: neg a6, a5 |
| ; CHECK-V-NEXT: neg a5, a4 |
| ; CHECK-V-NEXT: and a5, a5, a1 |
| ; CHECK-V-NEXT: bnez a4, .LBB24_8 |
| ; CHECK-V-NEXT: # %bb.7: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB24_8: # %entry |
| ; CHECK-V-NEXT: and a4, a6, s1 |
| ; CHECK-V-NEXT: slli a1, a2, 63 |
| ; CHECK-V-NEXT: beq a5, a2, .LBB24_11 |
| ; CHECK-V-NEXT: # %bb.9: # %entry |
| ; CHECK-V-NEXT: slti a3, a5, 0 |
| ; CHECK-V-NEXT: xori a3, a3, 1 |
| ; CHECK-V-NEXT: bne a4, a2, .LBB24_12 |
| ; CHECK-V-NEXT: .LBB24_10: |
| ; CHECK-V-NEXT: sltu a2, a1, s0 |
| ; CHECK-V-NEXT: beqz a2, .LBB24_13 |
| ; CHECK-V-NEXT: j .LBB24_14 |
| ; CHECK-V-NEXT: .LBB24_11: |
| ; CHECK-V-NEXT: sltu a3, a1, a0 |
| ; CHECK-V-NEXT: beq a4, a2, .LBB24_10 |
| ; CHECK-V-NEXT: .LBB24_12: # %entry |
| ; CHECK-V-NEXT: slti a2, a4, 0 |
| ; CHECK-V-NEXT: xori a2, a2, 1 |
| ; CHECK-V-NEXT: bnez a2, .LBB24_14 |
| ; CHECK-V-NEXT: .LBB24_13: # %entry |
| ; CHECK-V-NEXT: mv s0, a1 |
| ; CHECK-V-NEXT: .LBB24_14: # %entry |
| ; CHECK-V-NEXT: bnez a3, .LBB24_16 |
| ; CHECK-V-NEXT: # %bb.15: # %entry |
| ; CHECK-V-NEXT: mv a0, a1 |
| ; CHECK-V-NEXT: .LBB24_16: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v9, a0 |
| ; CHECK-V-NEXT: vmv.s.x v8, s0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 32 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x half> %x to <2 x i128> |
| %0 = icmp slt <2 x i128> %conv, <i128 9223372036854775807, i128 9223372036854775807> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 9223372036854775807, i128 9223372036854775807> |
| %1 = icmp sgt <2 x i128> %spec.store.select, <i128 -9223372036854775808, i128 -9223372036854775808> |
| %spec.store.select7 = select <2 x i1> %1, <2 x i128> %spec.store.select, <2 x i128> <i128 -9223372036854775808, i128 -9223372036854775808> |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @utesth_f16i64(<2 x half> %x) { |
| ; CHECK-NOV-LABEL: utesth_f16i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: mv s0, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixunssfti |
| ; CHECK-NOV-NEXT: mv s1, a0 |
| ; CHECK-NOV-NEXT: mv s2, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixunssfti |
| ; CHECK-NOV-NEXT: snez a1, a1 |
| ; CHECK-NOV-NEXT: snez a2, s2 |
| ; CHECK-NOV-NEXT: addi a2, a2, -1 |
| ; CHECK-NOV-NEXT: and a2, a2, s1 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a1, a1, a0 |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utesth_f16i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -32 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-V-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: mv s0, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixunssfti |
| ; CHECK-V-NEXT: mv s1, a0 |
| ; CHECK-V-NEXT: mv s2, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixunssfti |
| ; CHECK-V-NEXT: snez a1, a1 |
| ; CHECK-V-NEXT: snez a2, s2 |
| ; CHECK-V-NEXT: addi a2, a2, -1 |
| ; CHECK-V-NEXT: and a2, a2, s1 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v9, a0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a2 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 32 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x half> %x to <2 x i128> |
| %0 = icmp ult <2 x i128> %conv, <i128 18446744073709551616, i128 18446744073709551616> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616> |
| %conv6 = trunc <2 x i128> %spec.store.select to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @ustest_f16i64(<2 x half> %x) { |
| ; CHECK-NOV-LABEL: ustest_f16i64: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: mv s2, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv a2, s1 |
| ; CHECK-NOV-NEXT: blez s1, .LBB26_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: li a2, 1 |
| ; CHECK-NOV-NEXT: .LBB26_2: # %entry |
| ; CHECK-NOV-NEXT: slti a3, a1, 1 |
| ; CHECK-NOV-NEXT: slti a4, s1, 1 |
| ; CHECK-NOV-NEXT: blez a1, .LBB26_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: li a1, 1 |
| ; CHECK-NOV-NEXT: .LBB26_4: # %entry |
| ; CHECK-NOV-NEXT: neg a4, a4 |
| ; CHECK-NOV-NEXT: neg a3, a3 |
| ; CHECK-NOV-NEXT: and a3, a3, a0 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB26_7 |
| ; CHECK-NOV-NEXT: # %bb.5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a1, a1 |
| ; CHECK-NOV-NEXT: and a4, a4, s0 |
| ; CHECK-NOV-NEXT: bnez a2, .LBB26_8 |
| ; CHECK-NOV-NEXT: .LBB26_6: |
| ; CHECK-NOV-NEXT: snez a0, a4 |
| ; CHECK-NOV-NEXT: j .LBB26_9 |
| ; CHECK-NOV-NEXT: .LBB26_7: |
| ; CHECK-NOV-NEXT: snez a1, a3 |
| ; CHECK-NOV-NEXT: and a4, a4, s0 |
| ; CHECK-NOV-NEXT: beqz a2, .LBB26_6 |
| ; CHECK-NOV-NEXT: .LBB26_8: # %entry |
| ; CHECK-NOV-NEXT: sgtz a0, a2 |
| ; CHECK-NOV-NEXT: .LBB26_9: # %entry |
| ; CHECK-NOV-NEXT: neg a0, a0 |
| ; CHECK-NOV-NEXT: and a0, a0, a4 |
| ; CHECK-NOV-NEXT: neg a1, a1 |
| ; CHECK-NOV-NEXT: and a1, a1, a3 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f16i64: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -32 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-V-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: mv s2, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv a2, s1 |
| ; CHECK-V-NEXT: blez s1, .LBB26_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: li a2, 1 |
| ; CHECK-V-NEXT: .LBB26_2: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 1 |
| ; CHECK-V-NEXT: slti a3, s1, 1 |
| ; CHECK-V-NEXT: blez a1, .LBB26_4 |
| ; CHECK-V-NEXT: # %bb.3: # %entry |
| ; CHECK-V-NEXT: li a1, 1 |
| ; CHECK-V-NEXT: .LBB26_4: # %entry |
| ; CHECK-V-NEXT: neg a3, a3 |
| ; CHECK-V-NEXT: neg a4, a4 |
| ; CHECK-V-NEXT: and a0, a4, a0 |
| ; CHECK-V-NEXT: beqz a1, .LBB26_7 |
| ; CHECK-V-NEXT: # %bb.5: # %entry |
| ; CHECK-V-NEXT: sgtz a1, a1 |
| ; CHECK-V-NEXT: and a3, a3, s0 |
| ; CHECK-V-NEXT: bnez a2, .LBB26_8 |
| ; CHECK-V-NEXT: .LBB26_6: |
| ; CHECK-V-NEXT: snez a2, a3 |
| ; CHECK-V-NEXT: j .LBB26_9 |
| ; CHECK-V-NEXT: .LBB26_7: |
| ; CHECK-V-NEXT: snez a1, a0 |
| ; CHECK-V-NEXT: and a3, a3, s0 |
| ; CHECK-V-NEXT: beqz a2, .LBB26_6 |
| ; CHECK-V-NEXT: .LBB26_8: # %entry |
| ; CHECK-V-NEXT: sgtz a2, a2 |
| ; CHECK-V-NEXT: .LBB26_9: # %entry |
| ; CHECK-V-NEXT: neg a2, a2 |
| ; CHECK-V-NEXT: and a2, a2, a3 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v9, a0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a2 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 32 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x half> %x to <2 x i128> |
| %0 = icmp slt <2 x i128> %conv, <i128 18446744073709551616, i128 18446744073709551616> |
| %spec.store.select = select <2 x i1> %0, <2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616> |
| %1 = icmp sgt <2 x i128> %spec.store.select, zeroinitializer |
| %spec.store.select7 = select <2 x i1> %1, <2 x i128> %spec.store.select, <2 x i128> zeroinitializer |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| |
| |
| ; i32 saturate |
| |
| define <2 x i32> @stest_f64i32_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: stest_f64i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: lui a2, 524288 |
| ; CHECK-NOV-NEXT: addiw a3, a2, -1 |
| ; CHECK-NOV-NEXT: fcvt.l.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a1, a3, .LBB27_5 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: bge a0, a3, .LBB27_6 |
| ; CHECK-NOV-NEXT: .LBB27_2: # %entry |
| ; CHECK-NOV-NEXT: bge a2, a0, .LBB27_7 |
| ; CHECK-NOV-NEXT: .LBB27_3: # %entry |
| ; CHECK-NOV-NEXT: bge a2, a1, .LBB27_8 |
| ; CHECK-NOV-NEXT: .LBB27_4: # %entry |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB27_5: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: blt a0, a3, .LBB27_2 |
| ; CHECK-NOV-NEXT: .LBB27_6: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a3 |
| ; CHECK-NOV-NEXT: blt a2, a3, .LBB27_3 |
| ; CHECK-NOV-NEXT: .LBB27_7: # %entry |
| ; CHECK-NOV-NEXT: lui a0, 524288 |
| ; CHECK-NOV-NEXT: blt a2, a1, .LBB27_4 |
| ; CHECK-NOV-NEXT: .LBB27_8: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 524288 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f64i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.x.f.v v8, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i64> |
| %spec.store.select = call <2 x i64> @llvm.smin.v2i64(<2 x i64> %conv, <2 x i64> <i64 2147483647, i64 2147483647>) |
| %spec.store.select7 = call <2 x i64> @llvm.smax.v2i64(<2 x i64> %spec.store.select, <2 x i64> <i64 -2147483648, i64 -2147483648>) |
| %conv6 = trunc <2 x i64> %spec.store.select7 to <2 x i32> |
| ret <2 x i32> %conv6 |
| } |
| |
| define <2 x i32> @utest_f64i32_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: utest_f64i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: li a2, -1 |
| ; CHECK-NOV-NEXT: srli a2, a2, 32 |
| ; CHECK-NOV-NEXT: fcvt.lu.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: bgeu a0, a2, .LBB28_3 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: bgeu a1, a2, .LBB28_4 |
| ; CHECK-NOV-NEXT: .LBB28_2: # %entry |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB28_3: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: bltu a1, a2, .LBB28_2 |
| ; CHECK-NOV-NEXT: .LBB28_4: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utest_f64i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.xu.f.v v8, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x double> %x to <2 x i64> |
| %spec.store.select = call <2 x i64> @llvm.umin.v2i64(<2 x i64> %conv, <2 x i64> <i64 4294967295, i64 4294967295>) |
| %conv6 = trunc <2 x i64> %spec.store.select to <2 x i32> |
| ret <2 x i32> %conv6 |
| } |
| |
| define <2 x i32> @ustest_f64i32_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: ustest_f64i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: li a2, -1 |
| ; CHECK-NOV-NEXT: srli a2, a2, 32 |
| ; CHECK-NOV-NEXT: fcvt.l.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a1, a2, .LBB29_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: .LBB29_2: # %entry |
| ; CHECK-NOV-NEXT: blt a0, a2, .LBB29_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: .LBB29_4: # %entry |
| ; CHECK-NOV-NEXT: sgtz a2, a0 |
| ; CHECK-NOV-NEXT: neg a2, a2 |
| ; CHECK-NOV-NEXT: and a0, a2, a0 |
| ; CHECK-NOV-NEXT: sgtz a2, a1 |
| ; CHECK-NOV-NEXT: neg a2, a2 |
| ; CHECK-NOV-NEXT: and a1, a2, a1 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f64i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.x.f.v v8, v8 |
| ; CHECK-V-NEXT: li a0, -1 |
| ; CHECK-V-NEXT: srli a0, a0, 32 |
| ; CHECK-V-NEXT: vmin.vx v8, v8, a0 |
| ; CHECK-V-NEXT: vmax.vx v8, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i64> |
| %spec.store.select = call <2 x i64> @llvm.smin.v2i64(<2 x i64> %conv, <2 x i64> <i64 4294967295, i64 4294967295>) |
| %spec.store.select7 = call <2 x i64> @llvm.smax.v2i64(<2 x i64> %spec.store.select, <2 x i64> zeroinitializer) |
| %conv6 = trunc <2 x i64> %spec.store.select7 to <2 x i32> |
| ret <2 x i32> %conv6 |
| } |
| |
| define <4 x i32> @stest_f32i32_mm(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: stest_f32i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fa3, rtz |
| ; CHECK-NOV-NEXT: lui a3, 524288 |
| ; CHECK-NOV-NEXT: addiw a6, a3, -1 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fa2, rtz |
| ; CHECK-NOV-NEXT: bge a1, a6, .LBB30_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fa1, rtz |
| ; CHECK-NOV-NEXT: bge a2, a6, .LBB30_11 |
| ; CHECK-NOV-NEXT: .LBB30_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a4, a6, .LBB30_12 |
| ; CHECK-NOV-NEXT: .LBB30_3: # %entry |
| ; CHECK-NOV-NEXT: bge a5, a6, .LBB30_13 |
| ; CHECK-NOV-NEXT: .LBB30_4: # %entry |
| ; CHECK-NOV-NEXT: bge a3, a5, .LBB30_14 |
| ; CHECK-NOV-NEXT: .LBB30_5: # %entry |
| ; CHECK-NOV-NEXT: bge a3, a4, .LBB30_15 |
| ; CHECK-NOV-NEXT: .LBB30_6: # %entry |
| ; CHECK-NOV-NEXT: bge a3, a2, .LBB30_16 |
| ; CHECK-NOV-NEXT: .LBB30_7: # %entry |
| ; CHECK-NOV-NEXT: blt a3, a1, .LBB30_9 |
| ; CHECK-NOV-NEXT: .LBB30_8: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 524288 |
| ; CHECK-NOV-NEXT: .LBB30_9: # %entry |
| ; CHECK-NOV-NEXT: sw a1, 12(a0) |
| ; CHECK-NOV-NEXT: sw a2, 8(a0) |
| ; CHECK-NOV-NEXT: sw a4, 4(a0) |
| ; CHECK-NOV-NEXT: sw a5, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB30_10: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a6 |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fa1, rtz |
| ; CHECK-NOV-NEXT: blt a2, a6, .LBB30_2 |
| ; CHECK-NOV-NEXT: .LBB30_11: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a6 |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a4, a6, .LBB30_3 |
| ; CHECK-NOV-NEXT: .LBB30_12: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a6 |
| ; CHECK-NOV-NEXT: blt a5, a6, .LBB30_4 |
| ; CHECK-NOV-NEXT: .LBB30_13: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a6 |
| ; CHECK-NOV-NEXT: blt a3, a6, .LBB30_5 |
| ; CHECK-NOV-NEXT: .LBB30_14: # %entry |
| ; CHECK-NOV-NEXT: lui a5, 524288 |
| ; CHECK-NOV-NEXT: blt a3, a4, .LBB30_6 |
| ; CHECK-NOV-NEXT: .LBB30_15: # %entry |
| ; CHECK-NOV-NEXT: lui a4, 524288 |
| ; CHECK-NOV-NEXT: blt a3, a2, .LBB30_7 |
| ; CHECK-NOV-NEXT: .LBB30_16: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 524288 |
| ; CHECK-NOV-NEXT: bge a3, a1, .LBB30_8 |
| ; CHECK-NOV-NEXT: j .LBB30_9 |
| ; |
| ; CHECK-V-LABEL: stest_f32i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfwcvt.rtz.x.f.v v10, v8 |
| ; CHECK-V-NEXT: vnclip.wi v8, v10, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x float> %x to <4 x i64> |
| %spec.store.select = call <4 x i64> @llvm.smin.v4i64(<4 x i64> %conv, <4 x i64> <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647>) |
| %spec.store.select7 = call <4 x i64> @llvm.smax.v4i64(<4 x i64> %spec.store.select, <4 x i64> <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648>) |
| %conv6 = trunc <4 x i64> %spec.store.select7 to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @utest_f32i32_mm(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: utest_f32i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a1, fa0, rtz |
| ; CHECK-NOV-NEXT: li a3, -1 |
| ; CHECK-NOV-NEXT: srli a3, a3, 32 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fa1, rtz |
| ; CHECK-NOV-NEXT: bgeu a1, a3, .LBB31_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a4, fa2, rtz |
| ; CHECK-NOV-NEXT: bgeu a2, a3, .LBB31_7 |
| ; CHECK-NOV-NEXT: .LBB31_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a5, fa3, rtz |
| ; CHECK-NOV-NEXT: bgeu a4, a3, .LBB31_8 |
| ; CHECK-NOV-NEXT: .LBB31_3: # %entry |
| ; CHECK-NOV-NEXT: bltu a5, a3, .LBB31_5 |
| ; CHECK-NOV-NEXT: .LBB31_4: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a3 |
| ; CHECK-NOV-NEXT: .LBB31_5: # %entry |
| ; CHECK-NOV-NEXT: sw a5, 12(a0) |
| ; CHECK-NOV-NEXT: sw a4, 8(a0) |
| ; CHECK-NOV-NEXT: sw a2, 4(a0) |
| ; CHECK-NOV-NEXT: sw a1, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB31_6: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a4, fa2, rtz |
| ; CHECK-NOV-NEXT: bltu a2, a3, .LBB31_2 |
| ; CHECK-NOV-NEXT: .LBB31_7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a5, fa3, rtz |
| ; CHECK-NOV-NEXT: bltu a4, a3, .LBB31_3 |
| ; CHECK-NOV-NEXT: .LBB31_8: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a3 |
| ; CHECK-NOV-NEXT: bgeu a5, a3, .LBB31_4 |
| ; CHECK-NOV-NEXT: j .LBB31_5 |
| ; |
| ; CHECK-V-LABEL: utest_f32i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfwcvt.rtz.xu.f.v v10, v8 |
| ; CHECK-V-NEXT: vnclipu.wi v8, v10, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <4 x float> %x to <4 x i64> |
| %spec.store.select = call <4 x i64> @llvm.umin.v4i64(<4 x i64> %conv, <4 x i64> <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>) |
| %conv6 = trunc <4 x i64> %spec.store.select to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @ustest_f32i32_mm(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: ustest_f32i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fa3, rtz |
| ; CHECK-NOV-NEXT: li a3, -1 |
| ; CHECK-NOV-NEXT: srli a3, a3, 32 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fa2, rtz |
| ; CHECK-NOV-NEXT: bge a1, a3, .LBB32_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fa1, rtz |
| ; CHECK-NOV-NEXT: bge a2, a3, .LBB32_7 |
| ; CHECK-NOV-NEXT: .LBB32_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a4, a3, .LBB32_8 |
| ; CHECK-NOV-NEXT: .LBB32_3: # %entry |
| ; CHECK-NOV-NEXT: blt a5, a3, .LBB32_5 |
| ; CHECK-NOV-NEXT: .LBB32_4: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a3 |
| ; CHECK-NOV-NEXT: .LBB32_5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a3, a5 |
| ; CHECK-NOV-NEXT: negw a3, a3 |
| ; CHECK-NOV-NEXT: and a3, a3, a5 |
| ; CHECK-NOV-NEXT: sgtz a5, a4 |
| ; CHECK-NOV-NEXT: negw a5, a5 |
| ; CHECK-NOV-NEXT: and a4, a5, a4 |
| ; CHECK-NOV-NEXT: sgtz a5, a2 |
| ; CHECK-NOV-NEXT: negw a5, a5 |
| ; CHECK-NOV-NEXT: and a2, a5, a2 |
| ; CHECK-NOV-NEXT: sgtz a5, a1 |
| ; CHECK-NOV-NEXT: negw a5, a5 |
| ; CHECK-NOV-NEXT: and a1, a5, a1 |
| ; CHECK-NOV-NEXT: sw a1, 12(a0) |
| ; CHECK-NOV-NEXT: sw a2, 8(a0) |
| ; CHECK-NOV-NEXT: sw a4, 4(a0) |
| ; CHECK-NOV-NEXT: sw a3, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB32_6: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fa1, rtz |
| ; CHECK-NOV-NEXT: blt a2, a3, .LBB32_2 |
| ; CHECK-NOV-NEXT: .LBB32_7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a4, a3, .LBB32_3 |
| ; CHECK-NOV-NEXT: .LBB32_8: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a3 |
| ; CHECK-NOV-NEXT: bge a5, a3, .LBB32_4 |
| ; CHECK-NOV-NEXT: j .LBB32_5 |
| ; |
| ; CHECK-V-LABEL: ustest_f32i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfwcvt.rtz.x.f.v v10, v8 |
| ; CHECK-V-NEXT: li a0, -1 |
| ; CHECK-V-NEXT: srli a0, a0, 32 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e64, m2, ta, ma |
| ; CHECK-V-NEXT: vmin.vx v8, v10, a0 |
| ; CHECK-V-NEXT: vmax.vx v10, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v10, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x float> %x to <4 x i64> |
| %spec.store.select = call <4 x i64> @llvm.smin.v4i64(<4 x i64> %conv, <4 x i64> <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>) |
| %spec.store.select7 = call <4 x i64> @llvm.smax.v4i64(<4 x i64> %spec.store.select, <4 x i64> zeroinitializer) |
| %conv6 = trunc <4 x i64> %spec.store.select7 to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @stest_f16i32_mm(<4 x half> %x) { |
| ; CHECK-NOV-LABEL: stest_f16i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -64 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-NOV-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -64 |
| ; CHECK-NOV-NEXT: lhu s1, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 8(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 16(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.l.s s1, fs2, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a1, 524288 |
| ; CHECK-NOV-NEXT: addiw a4, a1, -1 |
| ; CHECK-NOV-NEXT: bge a0, a4, .LBB33_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs1, rtz |
| ; CHECK-NOV-NEXT: bge s1, a4, .LBB33_11 |
| ; CHECK-NOV-NEXT: .LBB33_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: bge a2, a4, .LBB33_12 |
| ; CHECK-NOV-NEXT: .LBB33_3: # %entry |
| ; CHECK-NOV-NEXT: bge a3, a4, .LBB33_13 |
| ; CHECK-NOV-NEXT: .LBB33_4: # %entry |
| ; CHECK-NOV-NEXT: bge a1, a3, .LBB33_14 |
| ; CHECK-NOV-NEXT: .LBB33_5: # %entry |
| ; CHECK-NOV-NEXT: bge a1, a2, .LBB33_15 |
| ; CHECK-NOV-NEXT: .LBB33_6: # %entry |
| ; CHECK-NOV-NEXT: bge a1, s1, .LBB33_16 |
| ; CHECK-NOV-NEXT: .LBB33_7: # %entry |
| ; CHECK-NOV-NEXT: blt a1, a0, .LBB33_9 |
| ; CHECK-NOV-NEXT: .LBB33_8: # %entry |
| ; CHECK-NOV-NEXT: lui a0, 524288 |
| ; CHECK-NOV-NEXT: .LBB33_9: # %entry |
| ; CHECK-NOV-NEXT: sw a0, 12(s0) |
| ; CHECK-NOV-NEXT: sw s1, 8(s0) |
| ; CHECK-NOV-NEXT: sw a2, 4(s0) |
| ; CHECK-NOV-NEXT: sw a3, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 64 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB33_10: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a4 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs1, rtz |
| ; CHECK-NOV-NEXT: blt s1, a4, .LBB33_2 |
| ; CHECK-NOV-NEXT: .LBB33_11: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a4 |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: blt a2, a4, .LBB33_3 |
| ; CHECK-NOV-NEXT: .LBB33_12: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a4 |
| ; CHECK-NOV-NEXT: blt a3, a4, .LBB33_4 |
| ; CHECK-NOV-NEXT: .LBB33_13: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a4 |
| ; CHECK-NOV-NEXT: blt a1, a4, .LBB33_5 |
| ; CHECK-NOV-NEXT: .LBB33_14: # %entry |
| ; CHECK-NOV-NEXT: lui a3, 524288 |
| ; CHECK-NOV-NEXT: blt a1, a2, .LBB33_6 |
| ; CHECK-NOV-NEXT: .LBB33_15: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 524288 |
| ; CHECK-NOV-NEXT: blt a1, s1, .LBB33_7 |
| ; CHECK-NOV-NEXT: .LBB33_16: # %entry |
| ; CHECK-NOV-NEXT: lui s1, 524288 |
| ; CHECK-NOV-NEXT: bge a1, a0, .LBB33_8 |
| ; CHECK-NOV-NEXT: j .LBB33_9 |
| ; |
| ; CHECK-V-LABEL: stest_f16i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -48 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 48 |
| ; CHECK-V-NEXT: sd ra, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu a0, 24(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v10, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e64, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 2 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 48 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x half> %x to <4 x i64> |
| %spec.store.select = call <4 x i64> @llvm.smin.v4i64(<4 x i64> %conv, <4 x i64> <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647>) |
| %spec.store.select7 = call <4 x i64> @llvm.smax.v4i64(<4 x i64> %spec.store.select, <4 x i64> <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648>) |
| %conv6 = trunc <4 x i64> %spec.store.select7 to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @utesth_f16i32_mm(<4 x half> %x) { |
| ; CHECK-NOV-LABEL: utesth_f16i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -64 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-NOV-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -64 |
| ; CHECK-NOV-NEXT: lhu s1, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 16(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 8(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s s1, fs2, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: li a1, -1 |
| ; CHECK-NOV-NEXT: srli a1, a1, 32 |
| ; CHECK-NOV-NEXT: bgeu a0, a1, .LBB34_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fs1, rtz |
| ; CHECK-NOV-NEXT: bgeu s1, a1, .LBB34_7 |
| ; CHECK-NOV-NEXT: .LBB34_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: bgeu a2, a1, .LBB34_8 |
| ; CHECK-NOV-NEXT: .LBB34_3: # %entry |
| ; CHECK-NOV-NEXT: bltu a3, a1, .LBB34_5 |
| ; CHECK-NOV-NEXT: .LBB34_4: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a1 |
| ; CHECK-NOV-NEXT: .LBB34_5: # %entry |
| ; CHECK-NOV-NEXT: sw a3, 12(s0) |
| ; CHECK-NOV-NEXT: sw a2, 8(s0) |
| ; CHECK-NOV-NEXT: sw s1, 4(s0) |
| ; CHECK-NOV-NEXT: sw a0, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 64 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB34_6: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fs1, rtz |
| ; CHECK-NOV-NEXT: bltu s1, a1, .LBB34_2 |
| ; CHECK-NOV-NEXT: .LBB34_7: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: bltu a2, a1, .LBB34_3 |
| ; CHECK-NOV-NEXT: .LBB34_8: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: bgeu a3, a1, .LBB34_4 |
| ; CHECK-NOV-NEXT: j .LBB34_5 |
| ; |
| ; CHECK-V-LABEL: utesth_f16i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -48 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 48 |
| ; CHECK-V-NEXT: sd ra, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu a0, 24(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v10, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e64, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 2 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 48 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <4 x half> %x to <4 x i64> |
| %spec.store.select = call <4 x i64> @llvm.umin.v4i64(<4 x i64> %conv, <4 x i64> <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>) |
| %conv6 = trunc <4 x i64> %spec.store.select to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| define <4 x i32> @ustest_f16i32_mm(<4 x half> %x) { |
| ; CHECK-NOV-LABEL: ustest_f16i32_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -64 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-NOV-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -64 |
| ; CHECK-NOV-NEXT: lhu s1, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 8(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 16(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.l.s s1, fs2, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: li a2, -1 |
| ; CHECK-NOV-NEXT: srli a2, a2, 32 |
| ; CHECK-NOV-NEXT: bge a0, a2, .LBB35_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs1, rtz |
| ; CHECK-NOV-NEXT: bge s1, a2, .LBB35_7 |
| ; CHECK-NOV-NEXT: .LBB35_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: bge a1, a2, .LBB35_8 |
| ; CHECK-NOV-NEXT: .LBB35_3: # %entry |
| ; CHECK-NOV-NEXT: blt a3, a2, .LBB35_5 |
| ; CHECK-NOV-NEXT: .LBB35_4: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a2 |
| ; CHECK-NOV-NEXT: .LBB35_5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a2, a3 |
| ; CHECK-NOV-NEXT: negw a2, a2 |
| ; CHECK-NOV-NEXT: and a2, a2, a3 |
| ; CHECK-NOV-NEXT: sgtz a3, a1 |
| ; CHECK-NOV-NEXT: negw a3, a3 |
| ; CHECK-NOV-NEXT: and a1, a3, a1 |
| ; CHECK-NOV-NEXT: sgtz a3, s1 |
| ; CHECK-NOV-NEXT: negw a3, a3 |
| ; CHECK-NOV-NEXT: and a3, a3, s1 |
| ; CHECK-NOV-NEXT: sgtz a4, a0 |
| ; CHECK-NOV-NEXT: negw a4, a4 |
| ; CHECK-NOV-NEXT: and a0, a4, a0 |
| ; CHECK-NOV-NEXT: sw a0, 12(s0) |
| ; CHECK-NOV-NEXT: sw a3, 8(s0) |
| ; CHECK-NOV-NEXT: sw a1, 4(s0) |
| ; CHECK-NOV-NEXT: sw a2, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 64 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB35_6: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs1, rtz |
| ; CHECK-NOV-NEXT: blt s1, a2, .LBB35_2 |
| ; CHECK-NOV-NEXT: .LBB35_7: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs0, rtz |
| ; CHECK-NOV-NEXT: blt a1, a2, .LBB35_3 |
| ; CHECK-NOV-NEXT: .LBB35_8: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: bge a3, a2, .LBB35_4 |
| ; CHECK-NOV-NEXT: j .LBB35_5 |
| ; |
| ; CHECK-V-LABEL: ustest_f16i32_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -48 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 48 |
| ; CHECK-V-NEXT: sd ra, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0d, 0x72, 0x00, 0x11, 0x30, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 48 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu a0, 24(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e64, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v10, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v10, 2 |
| ; CHECK-V-NEXT: li a0, -1 |
| ; CHECK-V-NEXT: srli a0, a0, 32 |
| ; CHECK-V-NEXT: vmin.vx v8, v8, a0 |
| ; CHECK-V-NEXT: vmax.vx v10, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 48 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x half> %x to <4 x i64> |
| %spec.store.select = call <4 x i64> @llvm.smin.v4i64(<4 x i64> %conv, <4 x i64> <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>) |
| %spec.store.select7 = call <4 x i64> @llvm.smax.v4i64(<4 x i64> %spec.store.select, <4 x i64> zeroinitializer) |
| %conv6 = trunc <4 x i64> %spec.store.select7 to <4 x i32> |
| ret <4 x i32> %conv6 |
| } |
| |
| ; i16 saturate |
| |
| define <2 x i16> @stest_f64i16_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: stest_f64i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: lui a2, 8 |
| ; CHECK-NOV-NEXT: addiw a2, a2, -1 |
| ; CHECK-NOV-NEXT: fcvt.w.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a1, a2, .LBB36_5 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: bge a0, a2, .LBB36_6 |
| ; CHECK-NOV-NEXT: .LBB36_2: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 1048568 |
| ; CHECK-NOV-NEXT: bge a2, a0, .LBB36_7 |
| ; CHECK-NOV-NEXT: .LBB36_3: # %entry |
| ; CHECK-NOV-NEXT: bge a2, a1, .LBB36_8 |
| ; CHECK-NOV-NEXT: .LBB36_4: # %entry |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB36_5: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: blt a0, a2, .LBB36_2 |
| ; CHECK-NOV-NEXT: .LBB36_6: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: lui a2, 1048568 |
| ; CHECK-NOV-NEXT: blt a2, a0, .LBB36_3 |
| ; CHECK-NOV-NEXT: .LBB36_7: # %entry |
| ; CHECK-NOV-NEXT: lui a0, 1048568 |
| ; CHECK-NOV-NEXT: blt a2, a1, .LBB36_4 |
| ; CHECK-NOV-NEXT: .LBB36_8: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 1048568 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f64i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vfncvt.rtz.x.f.w v9, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf4, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v9, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i32> |
| %spec.store.select = call <2 x i32> @llvm.smin.v2i32(<2 x i32> %conv, <2 x i32> <i32 32767, i32 32767>) |
| %spec.store.select7 = call <2 x i32> @llvm.smax.v2i32(<2 x i32> %spec.store.select, <2 x i32> <i32 -32768, i32 -32768>) |
| %conv6 = trunc <2 x i32> %spec.store.select7 to <2 x i16> |
| ret <2 x i16> %conv6 |
| } |
| |
| define <2 x i16> @utest_f64i16_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: utest_f64i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.wu.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a2, 16 |
| ; CHECK-NOV-NEXT: addiw a2, a2, -1 |
| ; CHECK-NOV-NEXT: fcvt.wu.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: bgeu a0, a2, .LBB37_3 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: bgeu a1, a2, .LBB37_4 |
| ; CHECK-NOV-NEXT: .LBB37_2: # %entry |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB37_3: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: bltu a1, a2, .LBB37_2 |
| ; CHECK-NOV-NEXT: .LBB37_4: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utest_f64i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vfncvt.rtz.xu.f.w v9, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf4, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v9, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x double> %x to <2 x i32> |
| %spec.store.select = call <2 x i32> @llvm.umin.v2i32(<2 x i32> %conv, <2 x i32> <i32 65535, i32 65535>) |
| %conv6 = trunc <2 x i32> %spec.store.select to <2 x i16> |
| ret <2 x i16> %conv6 |
| } |
| |
| define <2 x i16> @ustest_f64i16_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: ustest_f64i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.d a1, fa1, rtz |
| ; CHECK-NOV-NEXT: lui a2, 16 |
| ; CHECK-NOV-NEXT: addiw a2, a2, -1 |
| ; CHECK-NOV-NEXT: fcvt.w.d a0, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a1, a2, .LBB38_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: .LBB38_2: # %entry |
| ; CHECK-NOV-NEXT: blt a0, a2, .LBB38_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a2 |
| ; CHECK-NOV-NEXT: .LBB38_4: # %entry |
| ; CHECK-NOV-NEXT: sgtz a2, a0 |
| ; CHECK-NOV-NEXT: neg a2, a2 |
| ; CHECK-NOV-NEXT: and a0, a2, a0 |
| ; CHECK-NOV-NEXT: sgtz a2, a1 |
| ; CHECK-NOV-NEXT: neg a2, a2 |
| ; CHECK-NOV-NEXT: and a1, a2, a1 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f64i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vfncvt.rtz.x.f.w v9, v8 |
| ; CHECK-V-NEXT: lui a0, 16 |
| ; CHECK-V-NEXT: addi a0, a0, -1 |
| ; CHECK-V-NEXT: vmin.vx v8, v9, a0 |
| ; CHECK-V-NEXT: vmax.vx v8, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf4, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i32> |
| %spec.store.select = call <2 x i32> @llvm.smin.v2i32(<2 x i32> %conv, <2 x i32> <i32 65535, i32 65535>) |
| %spec.store.select7 = call <2 x i32> @llvm.smax.v2i32(<2 x i32> %spec.store.select, <2 x i32> zeroinitializer) |
| %conv6 = trunc <2 x i32> %spec.store.select7 to <2 x i16> |
| ret <2 x i16> %conv6 |
| } |
| |
| define <4 x i16> @stest_f32i16_mm(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: stest_f32i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a1, fa3, rtz |
| ; CHECK-NOV-NEXT: lui a5, 8 |
| ; CHECK-NOV-NEXT: addiw a5, a5, -1 |
| ; CHECK-NOV-NEXT: fcvt.w.s a2, fa2, rtz |
| ; CHECK-NOV-NEXT: bge a1, a5, .LBB39_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a3, fa1, rtz |
| ; CHECK-NOV-NEXT: bge a2, a5, .LBB39_11 |
| ; CHECK-NOV-NEXT: .LBB39_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a4, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a3, a5, .LBB39_12 |
| ; CHECK-NOV-NEXT: .LBB39_3: # %entry |
| ; CHECK-NOV-NEXT: bge a4, a5, .LBB39_13 |
| ; CHECK-NOV-NEXT: .LBB39_4: # %entry |
| ; CHECK-NOV-NEXT: lui a5, 1048568 |
| ; CHECK-NOV-NEXT: bge a5, a4, .LBB39_14 |
| ; CHECK-NOV-NEXT: .LBB39_5: # %entry |
| ; CHECK-NOV-NEXT: bge a5, a3, .LBB39_15 |
| ; CHECK-NOV-NEXT: .LBB39_6: # %entry |
| ; CHECK-NOV-NEXT: bge a5, a2, .LBB39_16 |
| ; CHECK-NOV-NEXT: .LBB39_7: # %entry |
| ; CHECK-NOV-NEXT: blt a5, a1, .LBB39_9 |
| ; CHECK-NOV-NEXT: .LBB39_8: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 1048568 |
| ; CHECK-NOV-NEXT: .LBB39_9: # %entry |
| ; CHECK-NOV-NEXT: sh a1, 6(a0) |
| ; CHECK-NOV-NEXT: sh a2, 4(a0) |
| ; CHECK-NOV-NEXT: sh a3, 2(a0) |
| ; CHECK-NOV-NEXT: sh a4, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB39_10: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a5 |
| ; CHECK-NOV-NEXT: fcvt.w.s a3, fa1, rtz |
| ; CHECK-NOV-NEXT: blt a2, a5, .LBB39_2 |
| ; CHECK-NOV-NEXT: .LBB39_11: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a5 |
| ; CHECK-NOV-NEXT: fcvt.w.s a4, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a3, a5, .LBB39_3 |
| ; CHECK-NOV-NEXT: .LBB39_12: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a5 |
| ; CHECK-NOV-NEXT: blt a4, a5, .LBB39_4 |
| ; CHECK-NOV-NEXT: .LBB39_13: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a5 |
| ; CHECK-NOV-NEXT: lui a5, 1048568 |
| ; CHECK-NOV-NEXT: blt a5, a4, .LBB39_5 |
| ; CHECK-NOV-NEXT: .LBB39_14: # %entry |
| ; CHECK-NOV-NEXT: lui a4, 1048568 |
| ; CHECK-NOV-NEXT: blt a5, a3, .LBB39_6 |
| ; CHECK-NOV-NEXT: .LBB39_15: # %entry |
| ; CHECK-NOV-NEXT: lui a3, 1048568 |
| ; CHECK-NOV-NEXT: blt a5, a2, .LBB39_7 |
| ; CHECK-NOV-NEXT: .LBB39_16: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 1048568 |
| ; CHECK-NOV-NEXT: bge a5, a1, .LBB39_8 |
| ; CHECK-NOV-NEXT: j .LBB39_9 |
| ; |
| ; CHECK-V-LABEL: stest_f32i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.x.f.v v8, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf2, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x float> %x to <4 x i32> |
| %spec.store.select = call <4 x i32> @llvm.smin.v4i32(<4 x i32> %conv, <4 x i32> <i32 32767, i32 32767, i32 32767, i32 32767>) |
| %spec.store.select7 = call <4 x i32> @llvm.smax.v4i32(<4 x i32> %spec.store.select, <4 x i32> <i32 -32768, i32 -32768, i32 -32768, i32 -32768>) |
| %conv6 = trunc <4 x i32> %spec.store.select7 to <4 x i16> |
| ret <4 x i16> %conv6 |
| } |
| |
| define <4 x i16> @utest_f32i16_mm(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: utest_f32i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.wu.s a1, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a3, 16 |
| ; CHECK-NOV-NEXT: addiw a3, a3, -1 |
| ; CHECK-NOV-NEXT: fcvt.wu.s a2, fa1, rtz |
| ; CHECK-NOV-NEXT: bgeu a1, a3, .LBB40_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.wu.s a4, fa2, rtz |
| ; CHECK-NOV-NEXT: bgeu a2, a3, .LBB40_7 |
| ; CHECK-NOV-NEXT: .LBB40_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.wu.s a5, fa3, rtz |
| ; CHECK-NOV-NEXT: bgeu a4, a3, .LBB40_8 |
| ; CHECK-NOV-NEXT: .LBB40_3: # %entry |
| ; CHECK-NOV-NEXT: bltu a5, a3, .LBB40_5 |
| ; CHECK-NOV-NEXT: .LBB40_4: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a3 |
| ; CHECK-NOV-NEXT: .LBB40_5: # %entry |
| ; CHECK-NOV-NEXT: sh a5, 6(a0) |
| ; CHECK-NOV-NEXT: sh a4, 4(a0) |
| ; CHECK-NOV-NEXT: sh a2, 2(a0) |
| ; CHECK-NOV-NEXT: sh a1, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB40_6: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: fcvt.wu.s a4, fa2, rtz |
| ; CHECK-NOV-NEXT: bltu a2, a3, .LBB40_2 |
| ; CHECK-NOV-NEXT: .LBB40_7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: fcvt.wu.s a5, fa3, rtz |
| ; CHECK-NOV-NEXT: bltu a4, a3, .LBB40_3 |
| ; CHECK-NOV-NEXT: .LBB40_8: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a3 |
| ; CHECK-NOV-NEXT: bgeu a5, a3, .LBB40_4 |
| ; CHECK-NOV-NEXT: j .LBB40_5 |
| ; |
| ; CHECK-V-LABEL: utest_f32i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.xu.f.v v8, v8 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf2, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <4 x float> %x to <4 x i32> |
| %spec.store.select = call <4 x i32> @llvm.umin.v4i32(<4 x i32> %conv, <4 x i32> <i32 65535, i32 65535, i32 65535, i32 65535>) |
| %conv6 = trunc <4 x i32> %spec.store.select to <4 x i16> |
| ret <4 x i16> %conv6 |
| } |
| |
| define <4 x i16> @ustest_f32i16_mm(<4 x float> %x) { |
| ; CHECK-NOV-LABEL: ustest_f32i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a1, fa3, rtz |
| ; CHECK-NOV-NEXT: lui a3, 16 |
| ; CHECK-NOV-NEXT: addiw a3, a3, -1 |
| ; CHECK-NOV-NEXT: fcvt.w.s a2, fa2, rtz |
| ; CHECK-NOV-NEXT: bge a1, a3, .LBB41_6 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a4, fa1, rtz |
| ; CHECK-NOV-NEXT: bge a2, a3, .LBB41_7 |
| ; CHECK-NOV-NEXT: .LBB41_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.w.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: bge a4, a3, .LBB41_8 |
| ; CHECK-NOV-NEXT: .LBB41_3: # %entry |
| ; CHECK-NOV-NEXT: blt a5, a3, .LBB41_5 |
| ; CHECK-NOV-NEXT: .LBB41_4: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a3 |
| ; CHECK-NOV-NEXT: .LBB41_5: # %entry |
| ; CHECK-NOV-NEXT: sgtz a3, a5 |
| ; CHECK-NOV-NEXT: negw a3, a3 |
| ; CHECK-NOV-NEXT: and a3, a3, a5 |
| ; CHECK-NOV-NEXT: sgtz a5, a4 |
| ; CHECK-NOV-NEXT: negw a5, a5 |
| ; CHECK-NOV-NEXT: and a4, a5, a4 |
| ; CHECK-NOV-NEXT: sgtz a5, a2 |
| ; CHECK-NOV-NEXT: negw a5, a5 |
| ; CHECK-NOV-NEXT: and a2, a5, a2 |
| ; CHECK-NOV-NEXT: sgtz a5, a1 |
| ; CHECK-NOV-NEXT: negw a5, a5 |
| ; CHECK-NOV-NEXT: and a1, a5, a1 |
| ; CHECK-NOV-NEXT: sh a1, 6(a0) |
| ; CHECK-NOV-NEXT: sh a2, 4(a0) |
| ; CHECK-NOV-NEXT: sh a4, 2(a0) |
| ; CHECK-NOV-NEXT: sh a3, 0(a0) |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB41_6: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: fcvt.w.s a4, fa1, rtz |
| ; CHECK-NOV-NEXT: blt a2, a3, .LBB41_2 |
| ; CHECK-NOV-NEXT: .LBB41_7: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: fcvt.w.s a5, fa0, rtz |
| ; CHECK-NOV-NEXT: blt a4, a3, .LBB41_3 |
| ; CHECK-NOV-NEXT: .LBB41_8: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a3 |
| ; CHECK-NOV-NEXT: bge a5, a3, .LBB41_4 |
| ; CHECK-NOV-NEXT: j .LBB41_5 |
| ; |
| ; CHECK-V-LABEL: ustest_f32i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: vfcvt.rtz.x.f.v v8, v8 |
| ; CHECK-V-NEXT: lui a0, 16 |
| ; CHECK-V-NEXT: addi a0, a0, -1 |
| ; CHECK-V-NEXT: vmin.vx v8, v8, a0 |
| ; CHECK-V-NEXT: vmax.vx v8, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, mf2, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v8, 0 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <4 x float> %x to <4 x i32> |
| %spec.store.select = call <4 x i32> @llvm.smin.v4i32(<4 x i32> %conv, <4 x i32> <i32 65535, i32 65535, i32 65535, i32 65535>) |
| %spec.store.select7 = call <4 x i32> @llvm.smax.v4i32(<4 x i32> %spec.store.select, <4 x i32> zeroinitializer) |
| %conv6 = trunc <4 x i32> %spec.store.select7 to <4 x i16> |
| ret <4 x i16> %conv6 |
| } |
| |
| define <8 x i16> @stest_f16i16_mm(<8 x half> %x) { |
| ; CHECK-NOV-LABEL: stest_f16i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -128 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 128 |
| ; CHECK-NOV-NEXT: sd ra, 120(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 112(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 104(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 96(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 88(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s4, 80(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s5, 72(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s6, 64(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s7, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs4, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs5, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs6, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset s4, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset s5, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset s6, -64 |
| ; CHECK-NOV-NEXT: .cfi_offset s7, -72 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -80 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -88 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -96 |
| ; CHECK-NOV-NEXT: .cfi_offset fs3, -104 |
| ; CHECK-NOV-NEXT: .cfi_offset fs4, -112 |
| ; CHECK-NOV-NEXT: .cfi_offset fs5, -120 |
| ; CHECK-NOV-NEXT: .cfi_offset fs6, -128 |
| ; CHECK-NOV-NEXT: lhu s1, 56(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 8(a1) |
| ; CHECK-NOV-NEXT: lhu s4, 16(a1) |
| ; CHECK-NOV-NEXT: lhu s5, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s6, 32(a1) |
| ; CHECK-NOV-NEXT: lhu s7, 40(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 48(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs6, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s7 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs5, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs4, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs3, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.l.s s1, fs6, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a7, 8 |
| ; CHECK-NOV-NEXT: addiw a7, a7, -1 |
| ; CHECK-NOV-NEXT: bge a0, a7, .LBB42_18 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs5, rtz |
| ; CHECK-NOV-NEXT: bge s1, a7, .LBB42_19 |
| ; CHECK-NOV-NEXT: .LBB42_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs4, rtz |
| ; CHECK-NOV-NEXT: bge a1, a7, .LBB42_20 |
| ; CHECK-NOV-NEXT: .LBB42_3: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs3, rtz |
| ; CHECK-NOV-NEXT: bge a2, a7, .LBB42_21 |
| ; CHECK-NOV-NEXT: .LBB42_4: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fs2, rtz |
| ; CHECK-NOV-NEXT: bge a3, a7, .LBB42_22 |
| ; CHECK-NOV-NEXT: .LBB42_5: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fs1, rtz |
| ; CHECK-NOV-NEXT: bge a4, a7, .LBB42_23 |
| ; CHECK-NOV-NEXT: .LBB42_6: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a6, fs0, rtz |
| ; CHECK-NOV-NEXT: bge a5, a7, .LBB42_24 |
| ; CHECK-NOV-NEXT: .LBB42_7: # %entry |
| ; CHECK-NOV-NEXT: bge a6, a7, .LBB42_25 |
| ; CHECK-NOV-NEXT: .LBB42_8: # %entry |
| ; CHECK-NOV-NEXT: lui a7, 1048568 |
| ; CHECK-NOV-NEXT: bge a7, a6, .LBB42_26 |
| ; CHECK-NOV-NEXT: .LBB42_9: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a5, .LBB42_27 |
| ; CHECK-NOV-NEXT: .LBB42_10: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a4, .LBB42_28 |
| ; CHECK-NOV-NEXT: .LBB42_11: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a3, .LBB42_29 |
| ; CHECK-NOV-NEXT: .LBB42_12: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a2, .LBB42_30 |
| ; CHECK-NOV-NEXT: .LBB42_13: # %entry |
| ; CHECK-NOV-NEXT: bge a7, a1, .LBB42_31 |
| ; CHECK-NOV-NEXT: .LBB42_14: # %entry |
| ; CHECK-NOV-NEXT: bge a7, s1, .LBB42_32 |
| ; CHECK-NOV-NEXT: .LBB42_15: # %entry |
| ; CHECK-NOV-NEXT: blt a7, a0, .LBB42_17 |
| ; CHECK-NOV-NEXT: .LBB42_16: # %entry |
| ; CHECK-NOV-NEXT: lui a0, 1048568 |
| ; CHECK-NOV-NEXT: .LBB42_17: # %entry |
| ; CHECK-NOV-NEXT: sh a0, 14(s0) |
| ; CHECK-NOV-NEXT: sh s1, 12(s0) |
| ; CHECK-NOV-NEXT: sh a1, 10(s0) |
| ; CHECK-NOV-NEXT: sh a2, 8(s0) |
| ; CHECK-NOV-NEXT: sh a3, 6(s0) |
| ; CHECK-NOV-NEXT: sh a4, 4(s0) |
| ; CHECK-NOV-NEXT: sh a5, 2(s0) |
| ; CHECK-NOV-NEXT: sh a6, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 120(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 112(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 104(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 96(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 88(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s4, 80(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s5, 72(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s6, 64(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s7, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs4, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs5, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs6, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 128 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB42_18: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs5, rtz |
| ; CHECK-NOV-NEXT: blt s1, a7, .LBB42_2 |
| ; CHECK-NOV-NEXT: .LBB42_19: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs4, rtz |
| ; CHECK-NOV-NEXT: blt a1, a7, .LBB42_3 |
| ; CHECK-NOV-NEXT: .LBB42_20: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a3, fs3, rtz |
| ; CHECK-NOV-NEXT: blt a2, a7, .LBB42_4 |
| ; CHECK-NOV-NEXT: .LBB42_21: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fs2, rtz |
| ; CHECK-NOV-NEXT: blt a3, a7, .LBB42_5 |
| ; CHECK-NOV-NEXT: .LBB42_22: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fs1, rtz |
| ; CHECK-NOV-NEXT: blt a4, a7, .LBB42_6 |
| ; CHECK-NOV-NEXT: .LBB42_23: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a7 |
| ; CHECK-NOV-NEXT: fcvt.l.s a6, fs0, rtz |
| ; CHECK-NOV-NEXT: blt a5, a7, .LBB42_7 |
| ; CHECK-NOV-NEXT: .LBB42_24: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a7 |
| ; CHECK-NOV-NEXT: blt a6, a7, .LBB42_8 |
| ; CHECK-NOV-NEXT: .LBB42_25: # %entry |
| ; CHECK-NOV-NEXT: mv a6, a7 |
| ; CHECK-NOV-NEXT: lui a7, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a6, .LBB42_9 |
| ; CHECK-NOV-NEXT: .LBB42_26: # %entry |
| ; CHECK-NOV-NEXT: lui a6, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a5, .LBB42_10 |
| ; CHECK-NOV-NEXT: .LBB42_27: # %entry |
| ; CHECK-NOV-NEXT: lui a5, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a4, .LBB42_11 |
| ; CHECK-NOV-NEXT: .LBB42_28: # %entry |
| ; CHECK-NOV-NEXT: lui a4, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a3, .LBB42_12 |
| ; CHECK-NOV-NEXT: .LBB42_29: # %entry |
| ; CHECK-NOV-NEXT: lui a3, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a2, .LBB42_13 |
| ; CHECK-NOV-NEXT: .LBB42_30: # %entry |
| ; CHECK-NOV-NEXT: lui a2, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, a1, .LBB42_14 |
| ; CHECK-NOV-NEXT: .LBB42_31: # %entry |
| ; CHECK-NOV-NEXT: lui a1, 1048568 |
| ; CHECK-NOV-NEXT: blt a7, s1, .LBB42_15 |
| ; CHECK-NOV-NEXT: .LBB42_32: # %entry |
| ; CHECK-NOV-NEXT: lui s1, 1048568 |
| ; CHECK-NOV-NEXT: bge a7, a0, .LBB42_16 |
| ; CHECK-NOV-NEXT: j .LBB42_17 |
| ; |
| ; CHECK-V-LABEL: stest_f16i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -80 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 80 |
| ; CHECK-V-NEXT: sd ra, 72(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 64(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s3, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s4, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s5, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s6, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: .cfi_offset s3, -40 |
| ; CHECK-V-NEXT: .cfi_offset s4, -48 |
| ; CHECK-V-NEXT: .cfi_offset s5, -56 |
| ; CHECK-V-NEXT: .cfi_offset s6, -64 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xd0, 0x00, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 80 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu s3, 24(a0) |
| ; CHECK-V-NEXT: lhu s4, 32(a0) |
| ; CHECK-V-NEXT: lhu s5, 40(a0) |
| ; CHECK-V-NEXT: lhu s6, 48(a0) |
| ; CHECK-V-NEXT: lhu a0, 56(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 2 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v10, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 2 |
| ; CHECK-V-NEXT: vsetivli zero, 8, e32, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 4 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, m1, ta, ma |
| ; CHECK-V-NEXT: vnclip.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 72(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 64(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s3, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s4, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s5, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s6, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 80 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <8 x half> %x to <8 x i32> |
| %spec.store.select = call <8 x i32> @llvm.smin.v8i32(<8 x i32> %conv, <8 x i32> <i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767>) |
| %spec.store.select7 = call <8 x i32> @llvm.smax.v8i32(<8 x i32> %spec.store.select, <8 x i32> <i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768>) |
| %conv6 = trunc <8 x i32> %spec.store.select7 to <8 x i16> |
| ret <8 x i16> %conv6 |
| } |
| |
| define <8 x i16> @utesth_f16i16_mm(<8 x half> %x) { |
| ; CHECK-NOV-LABEL: utesth_f16i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -128 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 128 |
| ; CHECK-NOV-NEXT: sd ra, 120(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 112(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 104(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 96(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 88(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s4, 80(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s5, 72(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s6, 64(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s7, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs4, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs5, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs6, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset s4, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset s5, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset s6, -64 |
| ; CHECK-NOV-NEXT: .cfi_offset s7, -72 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -80 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -88 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -96 |
| ; CHECK-NOV-NEXT: .cfi_offset fs3, -104 |
| ; CHECK-NOV-NEXT: .cfi_offset fs4, -112 |
| ; CHECK-NOV-NEXT: .cfi_offset fs5, -120 |
| ; CHECK-NOV-NEXT: .cfi_offset fs6, -128 |
| ; CHECK-NOV-NEXT: lhu s1, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 56(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 48(a1) |
| ; CHECK-NOV-NEXT: lhu s4, 40(a1) |
| ; CHECK-NOV-NEXT: lhu s5, 32(a1) |
| ; CHECK-NOV-NEXT: lhu s6, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s7, 16(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 8(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs6, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s7 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs5, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs4, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs3, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s s1, fs6, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a1, 16 |
| ; CHECK-NOV-NEXT: addiw a1, a1, -1 |
| ; CHECK-NOV-NEXT: bgeu a0, a1, .LBB43_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fs5, rtz |
| ; CHECK-NOV-NEXT: bgeu s1, a1, .LBB43_11 |
| ; CHECK-NOV-NEXT: .LBB43_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a3, fs4, rtz |
| ; CHECK-NOV-NEXT: bgeu a2, a1, .LBB43_12 |
| ; CHECK-NOV-NEXT: .LBB43_3: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a4, fs3, rtz |
| ; CHECK-NOV-NEXT: bgeu a3, a1, .LBB43_13 |
| ; CHECK-NOV-NEXT: .LBB43_4: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a5, fs2, rtz |
| ; CHECK-NOV-NEXT: bgeu a4, a1, .LBB43_14 |
| ; CHECK-NOV-NEXT: .LBB43_5: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a6, fs1, rtz |
| ; CHECK-NOV-NEXT: bgeu a5, a1, .LBB43_15 |
| ; CHECK-NOV-NEXT: .LBB43_6: # %entry |
| ; CHECK-NOV-NEXT: fcvt.lu.s a7, fs0, rtz |
| ; CHECK-NOV-NEXT: bgeu a6, a1, .LBB43_16 |
| ; CHECK-NOV-NEXT: .LBB43_7: # %entry |
| ; CHECK-NOV-NEXT: bltu a7, a1, .LBB43_9 |
| ; CHECK-NOV-NEXT: .LBB43_8: # %entry |
| ; CHECK-NOV-NEXT: mv a7, a1 |
| ; CHECK-NOV-NEXT: .LBB43_9: # %entry |
| ; CHECK-NOV-NEXT: sh a7, 14(s0) |
| ; CHECK-NOV-NEXT: sh a6, 12(s0) |
| ; CHECK-NOV-NEXT: sh a5, 10(s0) |
| ; CHECK-NOV-NEXT: sh a4, 8(s0) |
| ; CHECK-NOV-NEXT: sh a3, 6(s0) |
| ; CHECK-NOV-NEXT: sh a2, 4(s0) |
| ; CHECK-NOV-NEXT: sh s1, 2(s0) |
| ; CHECK-NOV-NEXT: sh a0, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 120(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 112(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 104(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 96(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 88(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s4, 80(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s5, 72(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s6, 64(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s7, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs4, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs5, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs6, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 128 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB43_10: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a2, fs5, rtz |
| ; CHECK-NOV-NEXT: bltu s1, a1, .LBB43_2 |
| ; CHECK-NOV-NEXT: .LBB43_11: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a3, fs4, rtz |
| ; CHECK-NOV-NEXT: bltu a2, a1, .LBB43_3 |
| ; CHECK-NOV-NEXT: .LBB43_12: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a4, fs3, rtz |
| ; CHECK-NOV-NEXT: bltu a3, a1, .LBB43_4 |
| ; CHECK-NOV-NEXT: .LBB43_13: # %entry |
| ; CHECK-NOV-NEXT: mv a3, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a5, fs2, rtz |
| ; CHECK-NOV-NEXT: bltu a4, a1, .LBB43_5 |
| ; CHECK-NOV-NEXT: .LBB43_14: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a6, fs1, rtz |
| ; CHECK-NOV-NEXT: bltu a5, a1, .LBB43_6 |
| ; CHECK-NOV-NEXT: .LBB43_15: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a1 |
| ; CHECK-NOV-NEXT: fcvt.lu.s a7, fs0, rtz |
| ; CHECK-NOV-NEXT: bltu a6, a1, .LBB43_7 |
| ; CHECK-NOV-NEXT: .LBB43_16: # %entry |
| ; CHECK-NOV-NEXT: mv a6, a1 |
| ; CHECK-NOV-NEXT: bgeu a7, a1, .LBB43_8 |
| ; CHECK-NOV-NEXT: j .LBB43_9 |
| ; |
| ; CHECK-V-LABEL: utesth_f16i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -80 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 80 |
| ; CHECK-V-NEXT: sd ra, 72(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 64(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s3, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s4, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s5, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s6, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: .cfi_offset s3, -40 |
| ; CHECK-V-NEXT: .cfi_offset s4, -48 |
| ; CHECK-V-NEXT: .cfi_offset s5, -56 |
| ; CHECK-V-NEXT: .cfi_offset s6, -64 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xd0, 0x00, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 80 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu s3, 24(a0) |
| ; CHECK-V-NEXT: lhu s4, 32(a0) |
| ; CHECK-V-NEXT: lhu s5, 40(a0) |
| ; CHECK-V-NEXT: lhu s6, 48(a0) |
| ; CHECK-V-NEXT: lhu a0, 56(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 2 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.lu.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v10, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 2 |
| ; CHECK-V-NEXT: vsetivli zero, 8, e32, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v10, v8, 4 |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, m1, ta, ma |
| ; CHECK-V-NEXT: vnclipu.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 72(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 64(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s3, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s4, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s5, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s6, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 80 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <8 x half> %x to <8 x i32> |
| %spec.store.select = call <8 x i32> @llvm.umin.v8i32(<8 x i32> %conv, <8 x i32> <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535>) |
| %conv6 = trunc <8 x i32> %spec.store.select to <8 x i16> |
| ret <8 x i16> %conv6 |
| } |
| |
| define <8 x i16> @ustest_f16i16_mm(<8 x half> %x) { |
| ; CHECK-NOV-LABEL: ustest_f16i16_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -128 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 128 |
| ; CHECK-NOV-NEXT: sd ra, 120(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 112(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 104(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 96(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s3, 88(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s4, 80(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s5, 72(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s6, 64(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s7, 56(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs2, 32(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs3, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs4, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs5, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs6, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: .cfi_offset s3, -40 |
| ; CHECK-NOV-NEXT: .cfi_offset s4, -48 |
| ; CHECK-NOV-NEXT: .cfi_offset s5, -56 |
| ; CHECK-NOV-NEXT: .cfi_offset s6, -64 |
| ; CHECK-NOV-NEXT: .cfi_offset s7, -72 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -80 |
| ; CHECK-NOV-NEXT: .cfi_offset fs1, -88 |
| ; CHECK-NOV-NEXT: .cfi_offset fs2, -96 |
| ; CHECK-NOV-NEXT: .cfi_offset fs3, -104 |
| ; CHECK-NOV-NEXT: .cfi_offset fs4, -112 |
| ; CHECK-NOV-NEXT: .cfi_offset fs5, -120 |
| ; CHECK-NOV-NEXT: .cfi_offset fs6, -128 |
| ; CHECK-NOV-NEXT: lhu s1, 56(a1) |
| ; CHECK-NOV-NEXT: lhu s2, 0(a1) |
| ; CHECK-NOV-NEXT: lhu s3, 8(a1) |
| ; CHECK-NOV-NEXT: lhu s4, 16(a1) |
| ; CHECK-NOV-NEXT: lhu s5, 24(a1) |
| ; CHECK-NOV-NEXT: lhu s6, 32(a1) |
| ; CHECK-NOV-NEXT: lhu s7, 40(a1) |
| ; CHECK-NOV-NEXT: lhu a1, 48(a1) |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs6, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s7 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs5, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs4, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs3, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs2, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs1, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-NOV-NEXT: fcvt.l.s s1, fs6, rtz |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-NOV-NEXT: lui a3, 16 |
| ; CHECK-NOV-NEXT: addiw a3, a3, -1 |
| ; CHECK-NOV-NEXT: bge a0, a3, .LBB44_10 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs5, rtz |
| ; CHECK-NOV-NEXT: bge s1, a3, .LBB44_11 |
| ; CHECK-NOV-NEXT: .LBB44_2: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs4, rtz |
| ; CHECK-NOV-NEXT: bge a1, a3, .LBB44_12 |
| ; CHECK-NOV-NEXT: .LBB44_3: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fs3, rtz |
| ; CHECK-NOV-NEXT: bge a2, a3, .LBB44_13 |
| ; CHECK-NOV-NEXT: .LBB44_4: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fs2, rtz |
| ; CHECK-NOV-NEXT: bge a4, a3, .LBB44_14 |
| ; CHECK-NOV-NEXT: .LBB44_5: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a6, fs1, rtz |
| ; CHECK-NOV-NEXT: bge a5, a3, .LBB44_15 |
| ; CHECK-NOV-NEXT: .LBB44_6: # %entry |
| ; CHECK-NOV-NEXT: fcvt.l.s a7, fs0, rtz |
| ; CHECK-NOV-NEXT: bge a6, a3, .LBB44_16 |
| ; CHECK-NOV-NEXT: .LBB44_7: # %entry |
| ; CHECK-NOV-NEXT: blt a7, a3, .LBB44_9 |
| ; CHECK-NOV-NEXT: .LBB44_8: # %entry |
| ; CHECK-NOV-NEXT: mv a7, a3 |
| ; CHECK-NOV-NEXT: .LBB44_9: # %entry |
| ; CHECK-NOV-NEXT: sgtz a3, a7 |
| ; CHECK-NOV-NEXT: negw a3, a3 |
| ; CHECK-NOV-NEXT: and a3, a3, a7 |
| ; CHECK-NOV-NEXT: sgtz a7, a6 |
| ; CHECK-NOV-NEXT: negw a7, a7 |
| ; CHECK-NOV-NEXT: and a6, a7, a6 |
| ; CHECK-NOV-NEXT: sgtz a7, a5 |
| ; CHECK-NOV-NEXT: negw a7, a7 |
| ; CHECK-NOV-NEXT: and a5, a7, a5 |
| ; CHECK-NOV-NEXT: sgtz a7, a4 |
| ; CHECK-NOV-NEXT: negw a7, a7 |
| ; CHECK-NOV-NEXT: and a4, a7, a4 |
| ; CHECK-NOV-NEXT: sgtz a7, a2 |
| ; CHECK-NOV-NEXT: negw a7, a7 |
| ; CHECK-NOV-NEXT: and a2, a7, a2 |
| ; CHECK-NOV-NEXT: sgtz a7, a1 |
| ; CHECK-NOV-NEXT: negw a7, a7 |
| ; CHECK-NOV-NEXT: and a1, a7, a1 |
| ; CHECK-NOV-NEXT: sgtz a7, s1 |
| ; CHECK-NOV-NEXT: negw a7, a7 |
| ; CHECK-NOV-NEXT: and a7, a7, s1 |
| ; CHECK-NOV-NEXT: sgtz t0, a0 |
| ; CHECK-NOV-NEXT: negw t0, t0 |
| ; CHECK-NOV-NEXT: and a0, t0, a0 |
| ; CHECK-NOV-NEXT: sh a0, 14(s0) |
| ; CHECK-NOV-NEXT: sh a7, 12(s0) |
| ; CHECK-NOV-NEXT: sh a1, 10(s0) |
| ; CHECK-NOV-NEXT: sh a2, 8(s0) |
| ; CHECK-NOV-NEXT: sh a4, 6(s0) |
| ; CHECK-NOV-NEXT: sh a5, 4(s0) |
| ; CHECK-NOV-NEXT: sh a6, 2(s0) |
| ; CHECK-NOV-NEXT: sh a3, 0(s0) |
| ; CHECK-NOV-NEXT: ld ra, 120(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 112(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 104(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 96(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s3, 88(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s4, 80(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s5, 72(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s6, 64(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s7, 56(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs2, 32(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs3, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs4, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs5, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs6, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 128 |
| ; CHECK-NOV-NEXT: ret |
| ; CHECK-NOV-NEXT: .LBB44_10: # %entry |
| ; CHECK-NOV-NEXT: mv a0, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a1, fs5, rtz |
| ; CHECK-NOV-NEXT: blt s1, a3, .LBB44_2 |
| ; CHECK-NOV-NEXT: .LBB44_11: # %entry |
| ; CHECK-NOV-NEXT: mv s1, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a2, fs4, rtz |
| ; CHECK-NOV-NEXT: blt a1, a3, .LBB44_3 |
| ; CHECK-NOV-NEXT: .LBB44_12: # %entry |
| ; CHECK-NOV-NEXT: mv a1, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a4, fs3, rtz |
| ; CHECK-NOV-NEXT: blt a2, a3, .LBB44_4 |
| ; CHECK-NOV-NEXT: .LBB44_13: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a5, fs2, rtz |
| ; CHECK-NOV-NEXT: blt a4, a3, .LBB44_5 |
| ; CHECK-NOV-NEXT: .LBB44_14: # %entry |
| ; CHECK-NOV-NEXT: mv a4, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a6, fs1, rtz |
| ; CHECK-NOV-NEXT: blt a5, a3, .LBB44_6 |
| ; CHECK-NOV-NEXT: .LBB44_15: # %entry |
| ; CHECK-NOV-NEXT: mv a5, a3 |
| ; CHECK-NOV-NEXT: fcvt.l.s a7, fs0, rtz |
| ; CHECK-NOV-NEXT: blt a6, a3, .LBB44_7 |
| ; CHECK-NOV-NEXT: .LBB44_16: # %entry |
| ; CHECK-NOV-NEXT: mv a6, a3 |
| ; CHECK-NOV-NEXT: bge a7, a3, .LBB44_8 |
| ; CHECK-NOV-NEXT: j .LBB44_9 |
| ; |
| ; CHECK-V-LABEL: ustest_f16i16_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -80 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 80 |
| ; CHECK-V-NEXT: sd ra, 72(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 64(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s3, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s4, 32(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s5, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s6, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: .cfi_offset s3, -40 |
| ; CHECK-V-NEXT: .cfi_offset s4, -48 |
| ; CHECK-V-NEXT: .cfi_offset s5, -56 |
| ; CHECK-V-NEXT: .cfi_offset s6, -64 |
| ; CHECK-V-NEXT: csrr a1, vlenb |
| ; CHECK-V-NEXT: slli a1, a1, 2 |
| ; CHECK-V-NEXT: sub sp, sp, a1 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xd0, 0x00, 0x22, 0x11, 0x04, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 80 + 4 * vlenb |
| ; CHECK-V-NEXT: lhu s0, 0(a0) |
| ; CHECK-V-NEXT: lhu s1, 8(a0) |
| ; CHECK-V-NEXT: lhu s2, 16(a0) |
| ; CHECK-V-NEXT: lhu s3, 24(a0) |
| ; CHECK-V-NEXT: lhu s4, 32(a0) |
| ; CHECK-V-NEXT: lhu s5, 40(a0) |
| ; CHECK-V-NEXT: lhu s6, 48(a0) |
| ; CHECK-V-NEXT: lhu a0, 56(a0) |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s6 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s5 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s4 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 2 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs2r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s3 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: fmv.w.x fa0, s1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, m1, ta, ma |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: fcvt.l.s a0, fa0, rtz |
| ; CHECK-V-NEXT: vsetivli zero, 2, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: addi a0, sp, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: vsetivli zero, 4, e32, m1, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl1r.v v9, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 2 |
| ; CHECK-V-NEXT: vsetivli zero, 8, e32, m2, ta, ma |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add a0, sp, a0 |
| ; CHECK-V-NEXT: addi a0, a0, 16 |
| ; CHECK-V-NEXT: vl2r.v v10, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslideup.vi v8, v10, 4 |
| ; CHECK-V-NEXT: lui a0, 16 |
| ; CHECK-V-NEXT: addi a0, a0, -1 |
| ; CHECK-V-NEXT: vmin.vx v8, v8, a0 |
| ; CHECK-V-NEXT: vmax.vx v10, v8, zero |
| ; CHECK-V-NEXT: vsetvli zero, zero, e16, m1, ta, ma |
| ; CHECK-V-NEXT: vnsrl.wi v8, v10, 0 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 2 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 72(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 64(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s3, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s4, 32(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s5, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s6, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 80 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <8 x half> %x to <8 x i32> |
| %spec.store.select = call <8 x i32> @llvm.smin.v8i32(<8 x i32> %conv, <8 x i32> <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535>) |
| %spec.store.select7 = call <8 x i32> @llvm.smax.v8i32(<8 x i32> %spec.store.select, <8 x i32> zeroinitializer) |
| %conv6 = trunc <8 x i32> %spec.store.select7 to <8 x i16> |
| ret <8 x i16> %conv6 |
| } |
| |
| ; i64 saturate |
| |
| define <2 x i64> @stest_f64i64_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: stest_f64i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.d fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixdfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.d fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixdfti |
| ; CHECK-NOV-NEXT: mv a2, a0 |
| ; CHECK-NOV-NEXT: li a0, -1 |
| ; CHECK-NOV-NEXT: srli a3, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB45_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: slti a4, a1, 0 |
| ; CHECK-NOV-NEXT: beqz a4, .LBB45_3 |
| ; CHECK-NOV-NEXT: j .LBB45_4 |
| ; CHECK-NOV-NEXT: .LBB45_2: |
| ; CHECK-NOV-NEXT: sltu a4, a2, a3 |
| ; CHECK-NOV-NEXT: bnez a4, .LBB45_4 |
| ; CHECK-NOV-NEXT: .LBB45_3: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB45_4: # %entry |
| ; CHECK-NOV-NEXT: beqz s1, .LBB45_6 |
| ; CHECK-NOV-NEXT: # %bb.5: # %entry |
| ; CHECK-NOV-NEXT: slti a6, s1, 0 |
| ; CHECK-NOV-NEXT: j .LBB45_7 |
| ; CHECK-NOV-NEXT: .LBB45_6: |
| ; CHECK-NOV-NEXT: sltu a6, s0, a3 |
| ; CHECK-NOV-NEXT: .LBB45_7: # %entry |
| ; CHECK-NOV-NEXT: neg a5, a6 |
| ; CHECK-NOV-NEXT: and a5, a5, s1 |
| ; CHECK-NOV-NEXT: bnez a6, .LBB45_9 |
| ; CHECK-NOV-NEXT: # %bb.8: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB45_9: # %entry |
| ; CHECK-NOV-NEXT: neg a4, a4 |
| ; CHECK-NOV-NEXT: slli a3, a0, 63 |
| ; CHECK-NOV-NEXT: beq a5, a0, .LBB45_11 |
| ; CHECK-NOV-NEXT: # %bb.10: # %entry |
| ; CHECK-NOV-NEXT: slti a5, a5, 0 |
| ; CHECK-NOV-NEXT: xori a5, a5, 1 |
| ; CHECK-NOV-NEXT: and a1, a4, a1 |
| ; CHECK-NOV-NEXT: beqz a5, .LBB45_12 |
| ; CHECK-NOV-NEXT: j .LBB45_13 |
| ; CHECK-NOV-NEXT: .LBB45_11: |
| ; CHECK-NOV-NEXT: sltu a5, a3, s0 |
| ; CHECK-NOV-NEXT: and a1, a4, a1 |
| ; CHECK-NOV-NEXT: bnez a5, .LBB45_13 |
| ; CHECK-NOV-NEXT: .LBB45_12: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB45_13: # %entry |
| ; CHECK-NOV-NEXT: beq a1, a0, .LBB45_15 |
| ; CHECK-NOV-NEXT: # %bb.14: # %entry |
| ; CHECK-NOV-NEXT: slti a0, a1, 0 |
| ; CHECK-NOV-NEXT: xori a0, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a0, .LBB45_16 |
| ; CHECK-NOV-NEXT: j .LBB45_17 |
| ; CHECK-NOV-NEXT: .LBB45_15: |
| ; CHECK-NOV-NEXT: sltu a0, a3, a2 |
| ; CHECK-NOV-NEXT: bnez a0, .LBB45_17 |
| ; CHECK-NOV-NEXT: .LBB45_16: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB45_17: # %entry |
| ; CHECK-NOV-NEXT: mv a0, s0 |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f64i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixdfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixdfti |
| ; CHECK-V-NEXT: li a2, -1 |
| ; CHECK-V-NEXT: srli a3, a2, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB45_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 0 |
| ; CHECK-V-NEXT: beqz a4, .LBB45_3 |
| ; CHECK-V-NEXT: j .LBB45_4 |
| ; CHECK-V-NEXT: .LBB45_2: |
| ; CHECK-V-NEXT: sltu a4, a0, a3 |
| ; CHECK-V-NEXT: bnez a4, .LBB45_4 |
| ; CHECK-V-NEXT: .LBB45_3: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB45_4: # %entry |
| ; CHECK-V-NEXT: beqz s1, .LBB45_6 |
| ; CHECK-V-NEXT: # %bb.5: # %entry |
| ; CHECK-V-NEXT: slti a6, s1, 0 |
| ; CHECK-V-NEXT: j .LBB45_7 |
| ; CHECK-V-NEXT: .LBB45_6: |
| ; CHECK-V-NEXT: sltu a6, s0, a3 |
| ; CHECK-V-NEXT: .LBB45_7: # %entry |
| ; CHECK-V-NEXT: neg a5, a6 |
| ; CHECK-V-NEXT: and a5, a5, s1 |
| ; CHECK-V-NEXT: bnez a6, .LBB45_9 |
| ; CHECK-V-NEXT: # %bb.8: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB45_9: # %entry |
| ; CHECK-V-NEXT: neg a4, a4 |
| ; CHECK-V-NEXT: slli a3, a2, 63 |
| ; CHECK-V-NEXT: beq a5, a2, .LBB45_11 |
| ; CHECK-V-NEXT: # %bb.10: # %entry |
| ; CHECK-V-NEXT: slti a5, a5, 0 |
| ; CHECK-V-NEXT: xori a5, a5, 1 |
| ; CHECK-V-NEXT: and a1, a4, a1 |
| ; CHECK-V-NEXT: beqz a5, .LBB45_12 |
| ; CHECK-V-NEXT: j .LBB45_13 |
| ; CHECK-V-NEXT: .LBB45_11: |
| ; CHECK-V-NEXT: sltu a5, a3, s0 |
| ; CHECK-V-NEXT: and a1, a4, a1 |
| ; CHECK-V-NEXT: bnez a5, .LBB45_13 |
| ; CHECK-V-NEXT: .LBB45_12: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB45_13: # %entry |
| ; CHECK-V-NEXT: beq a1, a2, .LBB45_15 |
| ; CHECK-V-NEXT: # %bb.14: # %entry |
| ; CHECK-V-NEXT: slti a1, a1, 0 |
| ; CHECK-V-NEXT: xori a1, a1, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB45_16 |
| ; CHECK-V-NEXT: j .LBB45_17 |
| ; CHECK-V-NEXT: .LBB45_15: |
| ; CHECK-V-NEXT: sltu a1, a3, a0 |
| ; CHECK-V-NEXT: bnez a1, .LBB45_17 |
| ; CHECK-V-NEXT: .LBB45_16: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB45_17: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, s0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.smin.v2i128(<2 x i128> %conv, <2 x i128> <i128 9223372036854775807, i128 9223372036854775807>) |
| %spec.store.select7 = call <2 x i128> @llvm.smax.v2i128(<2 x i128> %spec.store.select, <2 x i128> <i128 -9223372036854775808, i128 -9223372036854775808>) |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @utest_f64i64_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: utest_f64i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.d fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.d fa0, fa1 |
| ; CHECK-NOV-NEXT: call __fixunsdfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.d fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixunsdfti |
| ; CHECK-NOV-NEXT: snez a1, a1 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a0, a1, a0 |
| ; CHECK-NOV-NEXT: snez a1, s1 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a1, a1, s0 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utest_f64i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixunsdfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslidedown.vi v8, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixunsdfti |
| ; CHECK-V-NEXT: snez a1, a1 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: snez a1, s1 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a1, a1, s0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a1 |
| ; CHECK-V-NEXT: vmv.s.x v9, a0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x double> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.umin.v2i128(<2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616>) |
| %conv6 = trunc <2 x i128> %spec.store.select to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @ustest_f64i64_mm(<2 x double> %x) { |
| ; CHECK-NOV-LABEL: ustest_f64i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.d fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixdfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.d fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixdfti |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: blez a1, .LBB47_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: li a2, 1 |
| ; CHECK-NOV-NEXT: .LBB47_2: # %entry |
| ; CHECK-NOV-NEXT: mv a3, s1 |
| ; CHECK-NOV-NEXT: blez s1, .LBB47_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: li a3, 1 |
| ; CHECK-NOV-NEXT: .LBB47_4: # %entry |
| ; CHECK-NOV-NEXT: slti a1, a1, 1 |
| ; CHECK-NOV-NEXT: neg a1, a1 |
| ; CHECK-NOV-NEXT: and a1, a1, a0 |
| ; CHECK-NOV-NEXT: slti a0, s1, 1 |
| ; CHECK-NOV-NEXT: neg a0, a0 |
| ; CHECK-NOV-NEXT: and a0, a0, s0 |
| ; CHECK-NOV-NEXT: slti a3, a3, 0 |
| ; CHECK-NOV-NEXT: addi a3, a3, -1 |
| ; CHECK-NOV-NEXT: and a0, a3, a0 |
| ; CHECK-NOV-NEXT: slti a2, a2, 0 |
| ; CHECK-NOV-NEXT: addi a2, a2, -1 |
| ; CHECK-NOV-NEXT: and a1, a2, a1 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f64i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixdfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e64, m1, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixdfti |
| ; CHECK-V-NEXT: mv a2, a1 |
| ; CHECK-V-NEXT: blez a1, .LBB47_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: li a2, 1 |
| ; CHECK-V-NEXT: .LBB47_2: # %entry |
| ; CHECK-V-NEXT: mv a3, s1 |
| ; CHECK-V-NEXT: blez s1, .LBB47_4 |
| ; CHECK-V-NEXT: # %bb.3: # %entry |
| ; CHECK-V-NEXT: li a3, 1 |
| ; CHECK-V-NEXT: .LBB47_4: # %entry |
| ; CHECK-V-NEXT: slti a1, a1, 1 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: slti a1, s1, 1 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a1, a1, s0 |
| ; CHECK-V-NEXT: slti a3, a3, 0 |
| ; CHECK-V-NEXT: addi a3, a3, -1 |
| ; CHECK-V-NEXT: and a1, a3, a1 |
| ; CHECK-V-NEXT: slti a2, a2, 0 |
| ; CHECK-V-NEXT: addi a2, a2, -1 |
| ; CHECK-V-NEXT: and a0, a2, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, a1 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x double> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.smin.v2i128(<2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616>) |
| %spec.store.select7 = call <2 x i128> @llvm.smax.v2i128(<2 x i128> %spec.store.select, <2 x i128> zeroinitializer) |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @stest_f32i64_mm(<2 x float> %x) { |
| ; CHECK-NOV-LABEL: stest_f32i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.s fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv a2, a0 |
| ; CHECK-NOV-NEXT: li a0, -1 |
| ; CHECK-NOV-NEXT: srli a3, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB48_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: slti a4, a1, 0 |
| ; CHECK-NOV-NEXT: beqz a4, .LBB48_3 |
| ; CHECK-NOV-NEXT: j .LBB48_4 |
| ; CHECK-NOV-NEXT: .LBB48_2: |
| ; CHECK-NOV-NEXT: sltu a4, a2, a3 |
| ; CHECK-NOV-NEXT: bnez a4, .LBB48_4 |
| ; CHECK-NOV-NEXT: .LBB48_3: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB48_4: # %entry |
| ; CHECK-NOV-NEXT: beqz s1, .LBB48_6 |
| ; CHECK-NOV-NEXT: # %bb.5: # %entry |
| ; CHECK-NOV-NEXT: slti a6, s1, 0 |
| ; CHECK-NOV-NEXT: j .LBB48_7 |
| ; CHECK-NOV-NEXT: .LBB48_6: |
| ; CHECK-NOV-NEXT: sltu a6, s0, a3 |
| ; CHECK-NOV-NEXT: .LBB48_7: # %entry |
| ; CHECK-NOV-NEXT: neg a5, a6 |
| ; CHECK-NOV-NEXT: and a5, a5, s1 |
| ; CHECK-NOV-NEXT: bnez a6, .LBB48_9 |
| ; CHECK-NOV-NEXT: # %bb.8: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB48_9: # %entry |
| ; CHECK-NOV-NEXT: neg a4, a4 |
| ; CHECK-NOV-NEXT: slli a3, a0, 63 |
| ; CHECK-NOV-NEXT: beq a5, a0, .LBB48_11 |
| ; CHECK-NOV-NEXT: # %bb.10: # %entry |
| ; CHECK-NOV-NEXT: slti a5, a5, 0 |
| ; CHECK-NOV-NEXT: xori a5, a5, 1 |
| ; CHECK-NOV-NEXT: and a1, a4, a1 |
| ; CHECK-NOV-NEXT: beqz a5, .LBB48_12 |
| ; CHECK-NOV-NEXT: j .LBB48_13 |
| ; CHECK-NOV-NEXT: .LBB48_11: |
| ; CHECK-NOV-NEXT: sltu a5, a3, s0 |
| ; CHECK-NOV-NEXT: and a1, a4, a1 |
| ; CHECK-NOV-NEXT: bnez a5, .LBB48_13 |
| ; CHECK-NOV-NEXT: .LBB48_12: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB48_13: # %entry |
| ; CHECK-NOV-NEXT: beq a1, a0, .LBB48_15 |
| ; CHECK-NOV-NEXT: # %bb.14: # %entry |
| ; CHECK-NOV-NEXT: slti a0, a1, 0 |
| ; CHECK-NOV-NEXT: xori a0, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a0, .LBB48_16 |
| ; CHECK-NOV-NEXT: j .LBB48_17 |
| ; CHECK-NOV-NEXT: .LBB48_15: |
| ; CHECK-NOV-NEXT: sltu a0, a3, a2 |
| ; CHECK-NOV-NEXT: bnez a0, .LBB48_17 |
| ; CHECK-NOV-NEXT: .LBB48_16: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB48_17: # %entry |
| ; CHECK-NOV-NEXT: mv a0, s0 |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f32i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: li a2, -1 |
| ; CHECK-V-NEXT: srli a3, a2, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB48_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 0 |
| ; CHECK-V-NEXT: beqz a4, .LBB48_3 |
| ; CHECK-V-NEXT: j .LBB48_4 |
| ; CHECK-V-NEXT: .LBB48_2: |
| ; CHECK-V-NEXT: sltu a4, a0, a3 |
| ; CHECK-V-NEXT: bnez a4, .LBB48_4 |
| ; CHECK-V-NEXT: .LBB48_3: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB48_4: # %entry |
| ; CHECK-V-NEXT: beqz s1, .LBB48_6 |
| ; CHECK-V-NEXT: # %bb.5: # %entry |
| ; CHECK-V-NEXT: slti a6, s1, 0 |
| ; CHECK-V-NEXT: j .LBB48_7 |
| ; CHECK-V-NEXT: .LBB48_6: |
| ; CHECK-V-NEXT: sltu a6, s0, a3 |
| ; CHECK-V-NEXT: .LBB48_7: # %entry |
| ; CHECK-V-NEXT: neg a5, a6 |
| ; CHECK-V-NEXT: and a5, a5, s1 |
| ; CHECK-V-NEXT: bnez a6, .LBB48_9 |
| ; CHECK-V-NEXT: # %bb.8: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB48_9: # %entry |
| ; CHECK-V-NEXT: neg a4, a4 |
| ; CHECK-V-NEXT: slli a3, a2, 63 |
| ; CHECK-V-NEXT: beq a5, a2, .LBB48_11 |
| ; CHECK-V-NEXT: # %bb.10: # %entry |
| ; CHECK-V-NEXT: slti a5, a5, 0 |
| ; CHECK-V-NEXT: xori a5, a5, 1 |
| ; CHECK-V-NEXT: and a1, a4, a1 |
| ; CHECK-V-NEXT: beqz a5, .LBB48_12 |
| ; CHECK-V-NEXT: j .LBB48_13 |
| ; CHECK-V-NEXT: .LBB48_11: |
| ; CHECK-V-NEXT: sltu a5, a3, s0 |
| ; CHECK-V-NEXT: and a1, a4, a1 |
| ; CHECK-V-NEXT: bnez a5, .LBB48_13 |
| ; CHECK-V-NEXT: .LBB48_12: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB48_13: # %entry |
| ; CHECK-V-NEXT: beq a1, a2, .LBB48_15 |
| ; CHECK-V-NEXT: # %bb.14: # %entry |
| ; CHECK-V-NEXT: slti a1, a1, 0 |
| ; CHECK-V-NEXT: xori a1, a1, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB48_16 |
| ; CHECK-V-NEXT: j .LBB48_17 |
| ; CHECK-V-NEXT: .LBB48_15: |
| ; CHECK-V-NEXT: sltu a1, a3, a0 |
| ; CHECK-V-NEXT: bnez a1, .LBB48_17 |
| ; CHECK-V-NEXT: .LBB48_16: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB48_17: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, s0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x float> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.smin.v2i128(<2 x i128> %conv, <2 x i128> <i128 9223372036854775807, i128 9223372036854775807>) |
| %spec.store.select7 = call <2 x i128> @llvm.smax.v2i128(<2 x i128> %spec.store.select, <2 x i128> <i128 -9223372036854775808, i128 -9223372036854775808>) |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @utest_f32i64_mm(<2 x float> %x) { |
| ; CHECK-NOV-LABEL: utest_f32i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa0 |
| ; CHECK-NOV-NEXT: fmv.s fa0, fa1 |
| ; CHECK-NOV-NEXT: call __fixunssfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.s fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixunssfti |
| ; CHECK-NOV-NEXT: snez a1, a1 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a0, a1, a0 |
| ; CHECK-NOV-NEXT: snez a1, s1 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a1, a1, s0 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utest_f32i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixunssfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vslidedown.vi v8, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixunssfti |
| ; CHECK-V-NEXT: snez a1, a1 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: snez a1, s1 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a1, a1, s0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a1 |
| ; CHECK-V-NEXT: vmv.s.x v9, a0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x float> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.umin.v2i128(<2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616>) |
| %conv6 = trunc <2 x i128> %spec.store.select to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @ustest_f32i64_mm(<2 x float> %x) { |
| ; CHECK-NOV-LABEL: ustest_f32i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset fs0, -32 |
| ; CHECK-NOV-NEXT: fmv.s fs0, fa1 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.s fa0, fs0 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: blez a1, .LBB50_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: li a2, 1 |
| ; CHECK-NOV-NEXT: .LBB50_2: # %entry |
| ; CHECK-NOV-NEXT: mv a3, s1 |
| ; CHECK-NOV-NEXT: blez s1, .LBB50_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: li a3, 1 |
| ; CHECK-NOV-NEXT: .LBB50_4: # %entry |
| ; CHECK-NOV-NEXT: slti a1, a1, 1 |
| ; CHECK-NOV-NEXT: neg a1, a1 |
| ; CHECK-NOV-NEXT: and a1, a1, a0 |
| ; CHECK-NOV-NEXT: slti a0, s1, 1 |
| ; CHECK-NOV-NEXT: neg a0, a0 |
| ; CHECK-NOV-NEXT: and a0, a0, s0 |
| ; CHECK-NOV-NEXT: slti a3, a3, 0 |
| ; CHECK-NOV-NEXT: addi a3, a3, -1 |
| ; CHECK-NOV-NEXT: and a0, a3, a0 |
| ; CHECK-NOV-NEXT: slti a2, a2, 0 |
| ; CHECK-NOV-NEXT: addi a2, a2, -1 |
| ; CHECK-NOV-NEXT: and a1, a2, a1 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f32i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -64 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 64 |
| ; CHECK-V-NEXT: sd ra, 56(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 48(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 40(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: sub sp, sp, a0 |
| ; CHECK-V-NEXT: .cfi_escape 0x0f, 0x0e, 0x72, 0x00, 0x11, 0xc0, 0x00, 0x22, 0x11, 0x02, 0x92, 0xa2, 0x38, 0x00, 0x1e, 0x22 # sp + 64 + 2 * vlenb |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vs1r.v v8, (a0) # Unknown-size Folded Spill |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: vslidedown.vi v9, v8, 1 |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v9 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: vsetivli zero, 1, e32, mf2, ta, ma |
| ; CHECK-V-NEXT: addi a0, sp, 32 |
| ; CHECK-V-NEXT: vl1r.v v8, (a0) # Unknown-size Folded Reload |
| ; CHECK-V-NEXT: vfmv.f.s fa0, v8 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv a2, a1 |
| ; CHECK-V-NEXT: blez a1, .LBB50_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: li a2, 1 |
| ; CHECK-V-NEXT: .LBB50_2: # %entry |
| ; CHECK-V-NEXT: mv a3, s1 |
| ; CHECK-V-NEXT: blez s1, .LBB50_4 |
| ; CHECK-V-NEXT: # %bb.3: # %entry |
| ; CHECK-V-NEXT: li a3, 1 |
| ; CHECK-V-NEXT: .LBB50_4: # %entry |
| ; CHECK-V-NEXT: slti a1, a1, 1 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: slti a1, s1, 1 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a1, a1, s0 |
| ; CHECK-V-NEXT: slti a3, a3, 0 |
| ; CHECK-V-NEXT: addi a3, a3, -1 |
| ; CHECK-V-NEXT: and a1, a3, a1 |
| ; CHECK-V-NEXT: slti a2, a2, 0 |
| ; CHECK-V-NEXT: addi a2, a2, -1 |
| ; CHECK-V-NEXT: and a0, a2, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vmv.s.x v9, a1 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: csrr a0, vlenb |
| ; CHECK-V-NEXT: slli a0, a0, 1 |
| ; CHECK-V-NEXT: add sp, sp, a0 |
| ; CHECK-V-NEXT: ld ra, 56(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 48(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 40(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 64 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x float> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.smin.v2i128(<2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616>) |
| %spec.store.select7 = call <2 x i128> @llvm.smax.v2i128(<2 x i128> %spec.store.select, <2 x i128> zeroinitializer) |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @stest_f16i64_mm(<2 x half> %x) { |
| ; CHECK-NOV-LABEL: stest_f16i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: mv s2, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv a2, a0 |
| ; CHECK-NOV-NEXT: li a0, -1 |
| ; CHECK-NOV-NEXT: srli a3, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a1, .LBB51_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: slti a4, a1, 0 |
| ; CHECK-NOV-NEXT: beqz a4, .LBB51_3 |
| ; CHECK-NOV-NEXT: j .LBB51_4 |
| ; CHECK-NOV-NEXT: .LBB51_2: |
| ; CHECK-NOV-NEXT: sltu a4, a2, a3 |
| ; CHECK-NOV-NEXT: bnez a4, .LBB51_4 |
| ; CHECK-NOV-NEXT: .LBB51_3: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB51_4: # %entry |
| ; CHECK-NOV-NEXT: beqz s1, .LBB51_6 |
| ; CHECK-NOV-NEXT: # %bb.5: # %entry |
| ; CHECK-NOV-NEXT: slti a6, s1, 0 |
| ; CHECK-NOV-NEXT: j .LBB51_7 |
| ; CHECK-NOV-NEXT: .LBB51_6: |
| ; CHECK-NOV-NEXT: sltu a6, s0, a3 |
| ; CHECK-NOV-NEXT: .LBB51_7: # %entry |
| ; CHECK-NOV-NEXT: neg a5, a6 |
| ; CHECK-NOV-NEXT: and a5, a5, s1 |
| ; CHECK-NOV-NEXT: bnez a6, .LBB51_9 |
| ; CHECK-NOV-NEXT: # %bb.8: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB51_9: # %entry |
| ; CHECK-NOV-NEXT: neg a4, a4 |
| ; CHECK-NOV-NEXT: slli a3, a0, 63 |
| ; CHECK-NOV-NEXT: beq a5, a0, .LBB51_11 |
| ; CHECK-NOV-NEXT: # %bb.10: # %entry |
| ; CHECK-NOV-NEXT: slti a5, a5, 0 |
| ; CHECK-NOV-NEXT: xori a5, a5, 1 |
| ; CHECK-NOV-NEXT: and a1, a4, a1 |
| ; CHECK-NOV-NEXT: beqz a5, .LBB51_12 |
| ; CHECK-NOV-NEXT: j .LBB51_13 |
| ; CHECK-NOV-NEXT: .LBB51_11: |
| ; CHECK-NOV-NEXT: sltu a5, a3, s0 |
| ; CHECK-NOV-NEXT: and a1, a4, a1 |
| ; CHECK-NOV-NEXT: bnez a5, .LBB51_13 |
| ; CHECK-NOV-NEXT: .LBB51_12: # %entry |
| ; CHECK-NOV-NEXT: mv s0, a3 |
| ; CHECK-NOV-NEXT: .LBB51_13: # %entry |
| ; CHECK-NOV-NEXT: beq a1, a0, .LBB51_15 |
| ; CHECK-NOV-NEXT: # %bb.14: # %entry |
| ; CHECK-NOV-NEXT: slti a0, a1, 0 |
| ; CHECK-NOV-NEXT: xori a0, a0, 1 |
| ; CHECK-NOV-NEXT: beqz a0, .LBB51_16 |
| ; CHECK-NOV-NEXT: j .LBB51_17 |
| ; CHECK-NOV-NEXT: .LBB51_15: |
| ; CHECK-NOV-NEXT: sltu a0, a3, a2 |
| ; CHECK-NOV-NEXT: bnez a0, .LBB51_17 |
| ; CHECK-NOV-NEXT: .LBB51_16: # %entry |
| ; CHECK-NOV-NEXT: mv a2, a3 |
| ; CHECK-NOV-NEXT: .LBB51_17: # %entry |
| ; CHECK-NOV-NEXT: mv a0, s0 |
| ; CHECK-NOV-NEXT: mv a1, a2 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: stest_f16i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -32 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-V-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: mv s2, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: li a2, -1 |
| ; CHECK-V-NEXT: srli a3, a2, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB51_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: slti a4, a1, 0 |
| ; CHECK-V-NEXT: beqz a4, .LBB51_3 |
| ; CHECK-V-NEXT: j .LBB51_4 |
| ; CHECK-V-NEXT: .LBB51_2: |
| ; CHECK-V-NEXT: sltu a4, a0, a3 |
| ; CHECK-V-NEXT: bnez a4, .LBB51_4 |
| ; CHECK-V-NEXT: .LBB51_3: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB51_4: # %entry |
| ; CHECK-V-NEXT: beqz s1, .LBB51_6 |
| ; CHECK-V-NEXT: # %bb.5: # %entry |
| ; CHECK-V-NEXT: slti a6, s1, 0 |
| ; CHECK-V-NEXT: j .LBB51_7 |
| ; CHECK-V-NEXT: .LBB51_6: |
| ; CHECK-V-NEXT: sltu a6, s0, a3 |
| ; CHECK-V-NEXT: .LBB51_7: # %entry |
| ; CHECK-V-NEXT: neg a5, a6 |
| ; CHECK-V-NEXT: and a5, a5, s1 |
| ; CHECK-V-NEXT: bnez a6, .LBB51_9 |
| ; CHECK-V-NEXT: # %bb.8: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB51_9: # %entry |
| ; CHECK-V-NEXT: neg a4, a4 |
| ; CHECK-V-NEXT: slli a3, a2, 63 |
| ; CHECK-V-NEXT: beq a5, a2, .LBB51_11 |
| ; CHECK-V-NEXT: # %bb.10: # %entry |
| ; CHECK-V-NEXT: slti a5, a5, 0 |
| ; CHECK-V-NEXT: xori a5, a5, 1 |
| ; CHECK-V-NEXT: and a1, a4, a1 |
| ; CHECK-V-NEXT: beqz a5, .LBB51_12 |
| ; CHECK-V-NEXT: j .LBB51_13 |
| ; CHECK-V-NEXT: .LBB51_11: |
| ; CHECK-V-NEXT: sltu a5, a3, s0 |
| ; CHECK-V-NEXT: and a1, a4, a1 |
| ; CHECK-V-NEXT: bnez a5, .LBB51_13 |
| ; CHECK-V-NEXT: .LBB51_12: # %entry |
| ; CHECK-V-NEXT: mv s0, a3 |
| ; CHECK-V-NEXT: .LBB51_13: # %entry |
| ; CHECK-V-NEXT: beq a1, a2, .LBB51_15 |
| ; CHECK-V-NEXT: # %bb.14: # %entry |
| ; CHECK-V-NEXT: slti a1, a1, 0 |
| ; CHECK-V-NEXT: xori a1, a1, 1 |
| ; CHECK-V-NEXT: beqz a1, .LBB51_16 |
| ; CHECK-V-NEXT: j .LBB51_17 |
| ; CHECK-V-NEXT: .LBB51_15: |
| ; CHECK-V-NEXT: sltu a1, a3, a0 |
| ; CHECK-V-NEXT: bnez a1, .LBB51_17 |
| ; CHECK-V-NEXT: .LBB51_16: # %entry |
| ; CHECK-V-NEXT: mv a0, a3 |
| ; CHECK-V-NEXT: .LBB51_17: # %entry |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v9, a0 |
| ; CHECK-V-NEXT: vmv.s.x v8, s0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 32 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x half> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.smin.v2i128(<2 x i128> %conv, <2 x i128> <i128 9223372036854775807, i128 9223372036854775807>) |
| %spec.store.select7 = call <2 x i128> @llvm.smax.v2i128(<2 x i128> %spec.store.select, <2 x i128> <i128 -9223372036854775808, i128 -9223372036854775808>) |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @utesth_f16i64_mm(<2 x half> %x) { |
| ; CHECK-NOV-LABEL: utesth_f16i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixunssfti |
| ; CHECK-NOV-NEXT: mv s1, a0 |
| ; CHECK-NOV-NEXT: mv s2, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixunssfti |
| ; CHECK-NOV-NEXT: snez a1, a1 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a0, a1, a0 |
| ; CHECK-NOV-NEXT: snez a1, s2 |
| ; CHECK-NOV-NEXT: addi a1, a1, -1 |
| ; CHECK-NOV-NEXT: and a1, a1, s1 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: utesth_f16i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -32 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-V-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: fmv.w.x fa0, a1 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixunssfti |
| ; CHECK-V-NEXT: mv s1, a0 |
| ; CHECK-V-NEXT: mv s2, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, s0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixunssfti |
| ; CHECK-V-NEXT: snez a1, a1 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: snez a1, s2 |
| ; CHECK-V-NEXT: addi a1, a1, -1 |
| ; CHECK-V-NEXT: and a1, a1, s1 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v9, a1 |
| ; CHECK-V-NEXT: vmv.s.x v8, a0 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 32 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptoui <2 x half> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.umin.v2i128(<2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616>) |
| %conv6 = trunc <2 x i128> %spec.store.select to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| define <2 x i64> @ustest_f16i64_mm(<2 x half> %x) { |
| ; CHECK-NOV-LABEL: ustest_f16i64_mm: |
| ; CHECK-NOV: # %bb.0: # %entry |
| ; CHECK-NOV-NEXT: addi sp, sp, -32 |
| ; CHECK-NOV-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-NOV-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-NOV-NEXT: .cfi_offset ra, -8 |
| ; CHECK-NOV-NEXT: .cfi_offset s0, -16 |
| ; CHECK-NOV-NEXT: .cfi_offset s1, -24 |
| ; CHECK-NOV-NEXT: .cfi_offset s2, -32 |
| ; CHECK-NOV-NEXT: mv s2, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv s0, a0 |
| ; CHECK-NOV-NEXT: mv s1, a1 |
| ; CHECK-NOV-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-NOV-NEXT: call __extendhfsf2 |
| ; CHECK-NOV-NEXT: call __fixsfti |
| ; CHECK-NOV-NEXT: mv a2, a1 |
| ; CHECK-NOV-NEXT: blez a1, .LBB53_2 |
| ; CHECK-NOV-NEXT: # %bb.1: # %entry |
| ; CHECK-NOV-NEXT: li a2, 1 |
| ; CHECK-NOV-NEXT: .LBB53_2: # %entry |
| ; CHECK-NOV-NEXT: mv a3, s1 |
| ; CHECK-NOV-NEXT: blez s1, .LBB53_4 |
| ; CHECK-NOV-NEXT: # %bb.3: # %entry |
| ; CHECK-NOV-NEXT: li a3, 1 |
| ; CHECK-NOV-NEXT: .LBB53_4: # %entry |
| ; CHECK-NOV-NEXT: slti a1, a1, 1 |
| ; CHECK-NOV-NEXT: neg a1, a1 |
| ; CHECK-NOV-NEXT: and a1, a1, a0 |
| ; CHECK-NOV-NEXT: slti a0, s1, 1 |
| ; CHECK-NOV-NEXT: neg a0, a0 |
| ; CHECK-NOV-NEXT: and a0, a0, s0 |
| ; CHECK-NOV-NEXT: slti a3, a3, 0 |
| ; CHECK-NOV-NEXT: addi a3, a3, -1 |
| ; CHECK-NOV-NEXT: and a0, a3, a0 |
| ; CHECK-NOV-NEXT: slti a2, a2, 0 |
| ; CHECK-NOV-NEXT: addi a2, a2, -1 |
| ; CHECK-NOV-NEXT: and a1, a2, a1 |
| ; CHECK-NOV-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-NOV-NEXT: addi sp, sp, 32 |
| ; CHECK-NOV-NEXT: ret |
| ; |
| ; CHECK-V-LABEL: ustest_f16i64_mm: |
| ; CHECK-V: # %bb.0: # %entry |
| ; CHECK-V-NEXT: addi sp, sp, -32 |
| ; CHECK-V-NEXT: .cfi_def_cfa_offset 32 |
| ; CHECK-V-NEXT: sd ra, 24(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s0, 16(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s1, 8(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: sd s2, 0(sp) # 8-byte Folded Spill |
| ; CHECK-V-NEXT: .cfi_offset ra, -8 |
| ; CHECK-V-NEXT: .cfi_offset s0, -16 |
| ; CHECK-V-NEXT: .cfi_offset s1, -24 |
| ; CHECK-V-NEXT: .cfi_offset s2, -32 |
| ; CHECK-V-NEXT: mv s2, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, a0 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv s0, a0 |
| ; CHECK-V-NEXT: mv s1, a1 |
| ; CHECK-V-NEXT: fmv.w.x fa0, s2 |
| ; CHECK-V-NEXT: call __extendhfsf2 |
| ; CHECK-V-NEXT: call __fixsfti |
| ; CHECK-V-NEXT: mv a2, a1 |
| ; CHECK-V-NEXT: blez a1, .LBB53_2 |
| ; CHECK-V-NEXT: # %bb.1: # %entry |
| ; CHECK-V-NEXT: li a2, 1 |
| ; CHECK-V-NEXT: .LBB53_2: # %entry |
| ; CHECK-V-NEXT: mv a3, s1 |
| ; CHECK-V-NEXT: blez s1, .LBB53_4 |
| ; CHECK-V-NEXT: # %bb.3: # %entry |
| ; CHECK-V-NEXT: li a3, 1 |
| ; CHECK-V-NEXT: .LBB53_4: # %entry |
| ; CHECK-V-NEXT: slti a1, a1, 1 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a0, a1, a0 |
| ; CHECK-V-NEXT: slti a1, s1, 1 |
| ; CHECK-V-NEXT: neg a1, a1 |
| ; CHECK-V-NEXT: and a1, a1, s0 |
| ; CHECK-V-NEXT: slti a3, a3, 0 |
| ; CHECK-V-NEXT: addi a3, a3, -1 |
| ; CHECK-V-NEXT: and a1, a3, a1 |
| ; CHECK-V-NEXT: slti a2, a2, 0 |
| ; CHECK-V-NEXT: addi a2, a2, -1 |
| ; CHECK-V-NEXT: and a0, a2, a0 |
| ; CHECK-V-NEXT: vsetivli zero, 2, e64, m1, ta, ma |
| ; CHECK-V-NEXT: vmv.s.x v9, a0 |
| ; CHECK-V-NEXT: vmv.s.x v8, a1 |
| ; CHECK-V-NEXT: vslideup.vi v8, v9, 1 |
| ; CHECK-V-NEXT: ld ra, 24(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s0, 16(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s1, 8(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: ld s2, 0(sp) # 8-byte Folded Reload |
| ; CHECK-V-NEXT: addi sp, sp, 32 |
| ; CHECK-V-NEXT: ret |
| entry: |
| %conv = fptosi <2 x half> %x to <2 x i128> |
| %spec.store.select = call <2 x i128> @llvm.smin.v2i128(<2 x i128> %conv, <2 x i128> <i128 18446744073709551616, i128 18446744073709551616>) |
| %spec.store.select7 = call <2 x i128> @llvm.smax.v2i128(<2 x i128> %spec.store.select, <2 x i128> zeroinitializer) |
| %conv6 = trunc <2 x i128> %spec.store.select7 to <2 x i64> |
| ret <2 x i64> %conv6 |
| } |
| |
| declare <2 x i32> @llvm.smin.v2i32(<2 x i32>, <2 x i32>) |
| declare <2 x i32> @llvm.smax.v2i32(<2 x i32>, <2 x i32>) |
| declare <2 x i32> @llvm.umin.v2i32(<2 x i32>, <2 x i32>) |
| declare <4 x i32> @llvm.smin.v4i32(<4 x i32>, <4 x i32>) |
| declare <4 x i32> @llvm.smax.v4i32(<4 x i32>, <4 x i32>) |
| declare <4 x i32> @llvm.umin.v4i32(<4 x i32>, <4 x i32>) |
| declare <8 x i32> @llvm.smin.v8i32(<8 x i32>, <8 x i32>) |
| declare <8 x i32> @llvm.smax.v8i32(<8 x i32>, <8 x i32>) |
| declare <8 x i32> @llvm.umin.v8i32(<8 x i32>, <8 x i32>) |
| declare <2 x i64> @llvm.smin.v2i64(<2 x i64>, <2 x i64>) |
| declare <2 x i64> @llvm.smax.v2i64(<2 x i64>, <2 x i64>) |
| declare <2 x i64> @llvm.umin.v2i64(<2 x i64>, <2 x i64>) |
| declare <4 x i64> @llvm.smin.v4i64(<4 x i64>, <4 x i64>) |
| declare <4 x i64> @llvm.smax.v4i64(<4 x i64>, <4 x i64>) |
| declare <4 x i64> @llvm.umin.v4i64(<4 x i64>, <4 x i64>) |
| declare <2 x i128> @llvm.smin.v2i128(<2 x i128>, <2 x i128>) |
| declare <2 x i128> @llvm.smax.v2i128(<2 x i128>, <2 x i128>) |
| declare <2 x i128> @llvm.umin.v2i128(<2 x i128>, <2 x i128>) |
