intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2026-01-10 07:08:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: to always use grfs count in calculateNumThreadsPerThreadGroup

Browse Source 
grf size != grf count

Related-To: GSD-8437
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
This commit is contained in:
Katarzyna Cencelewska
2024-03-22 09:39:15 +00:00
committed by Compute-Runtime-Automation
parent df54d67f40
commit da7b03dd15
27 changed files with 86 additions and 68 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
shared/source/xe_hpg_core/command_encoder_xe_hpg_core.cpp
View File

@@ -1,5 +1,5 @@
/*
* Copyright (C) 2021-2023 Intel Corporation
* Copyright (C) 2021-2024 Intel Corporation
*
* SPDX-License-Identifier: MIT
*
@@ -87,7 +87,7 @@ void EncodeDispatchKernel<Family>::appendAdditionalIDDFields(InterfaceDescriptor
template <>
template <typename WalkerType, typename InterfaceDescriptorType>
void EncodeDispatchKernel<Family>::adjustInterfaceDescriptorData(InterfaceDescriptorType &interfaceDescriptor, const Device &device, const HardwareInfo &hwInfo, const uint32_t threadGroupCount, const uint32_t numGrf, WalkerType &walkerCmd) {
void EncodeDispatchKernel<Family>::adjustInterfaceDescriptorData(InterfaceDescriptorType &interfaceDescriptor, const Device &device, const HardwareInfo &hwInfo, const uint32_t threadGroupCount, const uint32_t grfCount, WalkerType &walkerCmd) {
const auto &productHelper = device.getProductHelper();
if (productHelper.isDisableOverdispatchAvailable(hwInfo)) {
if (interfaceDescriptor.getNumberOfThreadsInGpgpuThreadGroup() == 1) {