intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2026-01-05 09:09:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

refactor: change additional walker fields encoder 4/n

Browse Source 
- move post sync system fence into dedicated encoder

Related-To: NEO-12639

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
This commit is contained in:
Zbigniew Zdanowicz
2024-10-31 09:06:05 +00:00
committed by Compute-Runtime-Automation
parent 64061b623b
commit 32fd00e150
10 changed files with 38 additions and 60 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
shared/source/command_container/command_encoder_bdw_and_later.inl
View File

@@ -278,6 +278,15 @@ void EncodeDispatchKernel<Family>::encode(CommandContainer &container, EncodeDis
auto threadGroupCount = cmd.getThreadGroupIdXDimension() * cmd.getThreadGroupIdYDimension() * cmd.getThreadGroupIdZDimension();
EncodeDispatchKernel<Family>::encodeThreadGroupDispatch(idd, *args.device, hwInfo, threadGroupDims, threadGroupCount, kernelDescriptor.kernelAttributes.numGrfRequired, numThreadsPerThreadGroup, cmd);
EncodeWalkerArgs walkerArgs{
KernelExecutionType::defaultType,
args.requiresSystemMemoryFence(),
kernelDescriptor,
args.requiredDispatchWalkOrder,
args.additionalSizeParam,
args.device->getDeviceInfo().maxFrontEndThreads};
EncodeDispatchKernel<Family>::encodeAdditionalWalkerFields(rootDeviceEnvironment, cmd, walkerArgs);
EncodeDispatchKernel<Family>::encodeWalkerPostSyncFields(cmd, walkerArgs);
memcpy_s(iddPtr, sizeof(idd), &idd, sizeof(idd));
@@ -404,6 +413,10 @@ template <typename Family>
template <typename WalkerType>
inline void EncodeDispatchKernel<Family>::encodeAdditionalWalkerFields(const RootDeviceEnvironment &rootDeviceEnvironment, WalkerType &walkerCmd, const EncodeWalkerArgs &walkerArgs) {}
template <typename Family>
template <typename WalkerType>
inline void NEO::EncodeDispatchKernel<Family>::encodeWalkerPostSyncFields(WalkerType &walkerCmd, const EncodeWalkerArgs &walkerArgs) {}
template <typename Family>
template <typename InterfaceDescriptorType>
void EncodeDispatchKernel<Family>::setupPreferredSlmSize(InterfaceDescriptorType *pInterfaceDescriptor, const RootDeviceEnvironment &rootDeviceEnvironment, const uint32_t threadsPerThreadGroup, uint32_t slmTotalSize, SlmPolicy slmPolicy) {}