[Decode] Correct compute walker params

Thread number should be 32 for RP1 by GroupW*GroupH
diff --git a/media_driver/media_driver_next/agnostic/gen12/codec/hal/dec/av1/packet/decode_filmgrain_gennoise_rp1_packet_g12.cpp b/media_driver/media_driver_next/agnostic/gen12/codec/hal/dec/av1/packet/decode_filmgrain_gennoise_rp1_packet_g12.cpp
index 23dc14d..a9e8de2 100644
--- a/media_driver/media_driver_next/agnostic/gen12/codec/hal/dec/av1/packet/decode_filmgrain_gennoise_rp1_packet_g12.cpp
+++ b/media_driver/media_driver_next/agnostic/gen12/codec/hal/dec/av1/packet/decode_filmgrain_gennoise_rp1_packet_g12.cpp
@@ -319,7 +319,7 @@
     m_renderData.KernelParam.block_width        = CODECHAL_MACROBLOCK_WIDTH;
     m_renderData.KernelParam.block_height       = CODECHAL_MACROBLOCK_HEIGHT;
     m_renderData.KernelParam.blocks_x           = 1;
-    m_renderData.KernelParam.blocks_y           = 1;
+    m_renderData.KernelParam.blocks_y           = 32;
 
     m_renderData.iCurbeOffset                   = m_renderHal->pMhwStateHeap->GetSizeofCmdInterfaceDescriptorData();