blob: a1c796ec6ac7ec273b37d9b9d6df886b4d9c3136 [file] [log] [blame]
/**
* 3mm.c: This file is part of the PolyBench/C 3.2 test suite.
*
*
* Contact: Louis-Noel Pouchet <pouchet@cse.ohio-state.edu>
* Web address: http://polybench.sourceforge.net
*/
#include <stdio.h>
#include <unistd.h>
#include <string.h>
#include <math.h>
/* Include polybench common header. */
#include <polybench.h>
/* Include benchmark-specific header. */
/* Default data type is double, default size is 4000. */
#include "3mm.h"
/* Array initialization. */
static
void init_array(int ni, int nj, int nk, int nl, int nm,
DATA_TYPE POLYBENCH_2D(A,NI,NK,ni,nk),
DATA_TYPE POLYBENCH_2D(B,NK,NJ,nk,nj),
DATA_TYPE POLYBENCH_2D(C,NJ,NM,nj,nm),
DATA_TYPE POLYBENCH_2D(D,NM,NL,nm,nl))
{
#pragma STDC FP_CONTRACT OFF
int i, j;
for (i = 0; i < ni; i++)
for (j = 0; j < nk; j++)
A[i][j] = ((DATA_TYPE) i*j) / ni;
for (i = 0; i < nk; i++)
for (j = 0; j < nj; j++)
B[i][j] = ((DATA_TYPE) i*(j+1)) / nj;
for (i = 0; i < nj; i++)
for (j = 0; j < nm; j++)
C[i][j] = ((DATA_TYPE) i*(j+3)) / nl;
for (i = 0; i < nm; i++)
for (j = 0; j < nl; j++)
D[i][j] = ((DATA_TYPE) i*(j+2)) / nk;
}
/* DCE code. Must scan the entire live-out data.
Can be used also to check the correctness of the output. */
static
void print_array(int ni, int nl,
DATA_TYPE POLYBENCH_2D(G,NI,NL,ni,nl))
{
int i, j;
char *printmat = malloc(nl*16 + 1); printmat[nl*16] = 0;
for (i = 0; i < ni; i++) {
for (j = 0; j < nl; j++)
print_element(G[i][j], j*16, printmat);
fputs(printmat, stderr);
}
free(printmat);
}
/* Main computational kernel. The whole function will be timed,
including the call and return. */
static
void kernel_3mm(int ni, int nj, int nk, int nl, int nm,
DATA_TYPE POLYBENCH_2D(E,NI,NJ,ni,nj),
DATA_TYPE POLYBENCH_2D(A,NI,NK,ni,nk),
DATA_TYPE POLYBENCH_2D(B,NK,NJ,nk,nj),
DATA_TYPE POLYBENCH_2D(F,NJ,NL,nj,nl),
DATA_TYPE POLYBENCH_2D(C,NJ,NM,nj,nm),
DATA_TYPE POLYBENCH_2D(D,NM,NL,nm,nl),
DATA_TYPE POLYBENCH_2D(G,NI,NL,ni,nl))
{
int i, j, k;
#pragma scop
/* E := A*B */
for (i = 0; i < _PB_NI; i++)
for (j = 0; j < _PB_NJ; j++)
{
E[i][j] = 0;
for (k = 0; k < _PB_NK; ++k)
E[i][j] += A[i][k] * B[k][j];
}
/* F := C*D */
for (i = 0; i < _PB_NJ; i++)
for (j = 0; j < _PB_NL; j++)
{
F[i][j] = 0;
for (k = 0; k < _PB_NM; ++k)
F[i][j] += C[i][k] * D[k][j];
}
/* G := E*F */
for (i = 0; i < _PB_NI; i++)
for (j = 0; j < _PB_NL; j++)
{
G[i][j] = 0;
for (k = 0; k < _PB_NJ; ++k)
G[i][j] += E[i][k] * F[k][j];
}
#pragma endscop
}
#if !FMA_DISABLED
// NOTE: FMA_DISABLED is true for targets where FMA contraction causes
// discrepancies which cause the accuracy checks to fail.
// In this case, the test runs with the option -ffp-contract=off
static
void kernel_3mm_StrictFP(int ni, int nj, int nk, int nl, int nm,
DATA_TYPE POLYBENCH_2D(E,NI,NJ,ni,nj),
DATA_TYPE POLYBENCH_2D(A,NI,NK,ni,nk),
DATA_TYPE POLYBENCH_2D(B,NK,NJ,nk,nj),
DATA_TYPE POLYBENCH_2D(F,NJ,NL,nj,nl),
DATA_TYPE POLYBENCH_2D(C,NJ,NM,nj,nm),
DATA_TYPE POLYBENCH_2D(D,NM,NL,nm,nl),
DATA_TYPE POLYBENCH_2D(G,NI,NL,ni,nl))
{
#pragma STDC FP_CONTRACT OFF
int i, j, k;
/* E := A*B */
for (i = 0; i < _PB_NI; i++)
for (j = 0; j < _PB_NJ; j++)
{
E[i][j] = 0;
for (k = 0; k < _PB_NK; ++k)
E[i][j] += A[i][k] * B[k][j];
}
/* F := C*D */
for (i = 0; i < _PB_NJ; i++)
for (j = 0; j < _PB_NL; j++)
{
F[i][j] = 0;
for (k = 0; k < _PB_NM; ++k)
F[i][j] += C[i][k] * D[k][j];
}
/* G := E*F */
for (i = 0; i < _PB_NI; i++)
for (j = 0; j < _PB_NL; j++)
{
G[i][j] = 0;
for (k = 0; k < _PB_NJ; ++k)
G[i][j] += E[i][k] * F[k][j];
}
}
/* Return 0 when one of the elements of arrays A and B do not match within the
allowed FP_ABSTOLERANCE. Return 1 when all elements match. */
static inline int
check_FP(int ni, int nl,
DATA_TYPE POLYBENCH_2D(A,NI,NL,ni,nl),
DATA_TYPE POLYBENCH_2D(B,NI,NL,ni,nl)) {
int i, j;
double AbsTolerance = FP_ABSTOLERANCE;
for (i = 0; i < _PB_NI; i++)
for (j = 0; j < _PB_NL; j++)
{
double V1 = A[i][j];
double V2 = B[i][j];
double Diff = fabs(V1 - V2);
if (Diff > AbsTolerance) {
fprintf(stderr, "A[%d][%d] = %lf and B[%d][%d] = %lf differ more than"
" FP_ABSTOLERANCE = %lf\n", i, j, V1, i, j, V2, AbsTolerance);
return 0;
}
}
return 1;
}
#endif
int main(int argc, char** argv)
{
/* Retrieve problem size. */
int ni = NI;
int nj = NJ;
int nk = NK;
int nl = NL;
int nm = NM;
/* Variable declaration/allocation. */
POLYBENCH_2D_ARRAY_DECL(E, DATA_TYPE, NI, NJ, ni, nj);
POLYBENCH_2D_ARRAY_DECL(A, DATA_TYPE, NI, NK, ni, nk);
POLYBENCH_2D_ARRAY_DECL(B, DATA_TYPE, NK, NJ, nk, nj);
POLYBENCH_2D_ARRAY_DECL(F, DATA_TYPE, NJ, NL, nj, nl);
POLYBENCH_2D_ARRAY_DECL(C, DATA_TYPE, NJ, NM, nj, nm);
POLYBENCH_2D_ARRAY_DECL(D, DATA_TYPE, NM, NL, nm, nl);
POLYBENCH_2D_ARRAY_DECL(G, DATA_TYPE, NI, NL, ni, nl);
#if !FMA_DISABLED
POLYBENCH_2D_ARRAY_DECL(G_StrictFP, DATA_TYPE, NI, NL, ni, nl);
#endif
/* Initialize array(s). */
init_array (ni, nj, nk, nl, nm,
POLYBENCH_ARRAY(A),
POLYBENCH_ARRAY(B),
POLYBENCH_ARRAY(C),
POLYBENCH_ARRAY(D));
/* Start timer. */
polybench_start_instruments;
/* Run kernel. */
kernel_3mm (ni, nj, nk, nl, nm,
POLYBENCH_ARRAY(E),
POLYBENCH_ARRAY(A),
POLYBENCH_ARRAY(B),
POLYBENCH_ARRAY(F),
POLYBENCH_ARRAY(C),
POLYBENCH_ARRAY(D),
POLYBENCH_ARRAY(G));
/* Stop and print timer. */
polybench_stop_instruments;
polybench_print_instruments;
#if FMA_DISABLED
/* Prevent dead-code elimination. All live-out data must be printed
by the function call in argument. */
polybench_prevent_dce(print_array(ni, nl, POLYBENCH_ARRAY(G)));
#else
kernel_3mm_StrictFP(ni, nj, nk, nl, nm,
POLYBENCH_ARRAY(E),
POLYBENCH_ARRAY(A),
POLYBENCH_ARRAY(B),
POLYBENCH_ARRAY(F),
POLYBENCH_ARRAY(C),
POLYBENCH_ARRAY(D),
POLYBENCH_ARRAY(G_StrictFP));
if (!check_FP(ni, nl, POLYBENCH_ARRAY(G), POLYBENCH_ARRAY(G_StrictFP)))
return 1;
/* Prevent dead-code elimination. All live-out data must be printed
by the function call in argument. */
polybench_prevent_dce(print_array(ni, nl, POLYBENCH_ARRAY(G_StrictFP)));
#endif
/* Be clean. */
POLYBENCH_FREE_ARRAY(E);
POLYBENCH_FREE_ARRAY(A);
POLYBENCH_FREE_ARRAY(B);
POLYBENCH_FREE_ARRAY(F);
POLYBENCH_FREE_ARRAY(C);
POLYBENCH_FREE_ARRAY(D);
POLYBENCH_FREE_ARRAY(G);
#if !FMA_DISABLED
POLYBENCH_FREE_ARRAY(G_StrictFP);
#endif
return 0;
}