intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2026-01-08 05:56:36 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: correct suggested number of work groups for concurrent kernels on PVC

Browse Source 
value depends on CCS count:
- single CCS mode (default) - no limitations
- two CCS mode - 25% available
- four CCS mode - 12.5% available

Related-To: NEO-8377
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
This commit is contained in:
Mateusz Jablonski
2023-09-14 13:53:58 +00:00
committed by Compute-Runtime-Automation
parent 513045bd90
commit 6fc673b0fe
8 changed files with 31 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
shared/source/helpers/gfx_core_helper_pvc_and_later.inl
View File

@@ -63,10 +63,10 @@ uint32_t GfxCoreHelperHw<Family>::adjustMaxWorkGroupCount(uint32_t maxWorkGroupC
bool requiresLimitation = productHelper.isCooperativeEngineSupported(hwInfo) &&
(engineGroupType != EngineGroupType::CooperativeCompute) &&
(!isEngineInstanced);
auto numberOfpartsInTileForConcurrentKernels = productHelper.getNumberOfPartsInTileForConcurrentKernel();
if (requiresLimitation) {
auto ccsCount = hwInfo.gtSystemInfo.CCSInfo.NumberOfCCSEnabled;
auto ccsCount = hwInfo.gtSystemInfo.CCSInfo.NumberOfCCSEnabled;
auto numberOfpartsInTileForConcurrentKernels = productHelper.getNumberOfPartsInTileForConcurrentKernel(ccsCount);
if (requiresLimitation) {
UNRECOVERABLE_IF(ccsCount == 0);
numberOfpartsInTileForConcurrentKernels = std::max(numberOfpartsInTileForConcurrentKernels, ccsCount);
}