P4 to Git Change 1767752 by wchau@wchau_OCL_Linux on 2019/04/09 22:58:03

SWDEV-165259 - Update OpenCL runtime to support MsgPack metadata
	- Add support for the V3 code objects

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/devkernel.cpp#19 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/devkernel.hpp#14 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/devprogram.cpp#39 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/devprogram.hpp#24 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpukernel.cpp#336 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpukernel.hpp#134 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palbe/inc/core/palCmdBuffer.h#63 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palbe/src/core/hw/gfxip/gfx6/gfx6ComputeCmdBuffer.cpp#63 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palbe/src/core/hw/gfxip/gfx9/gfx9ComputeCmdBuffer.cpp#69 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palkernel.cpp#77 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palkernel.hpp#27 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palprogram.cpp#90 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palsettings.cpp#76 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palsettings.hpp#21 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palvirtual.cpp#130 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rockernel.cpp#52 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rockernel.hpp#27 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocprogram.cpp#103 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocprogram.hpp#47 edit
This commit is contained in:
foreman
2019-04-09 23:24:10 -04:00
parent 0fb499bb71
commit 36a5f2a85f
15 changed files with 951 additions and 417 deletions
+7
View File
@@ -553,6 +553,11 @@ void VirtualGPU::DmaFlushMgmt::resetCbWorkload(const Device& dev) {
void VirtualGPU::DmaFlushMgmt::findSplitSize(const Device& dev, uint64_t threads,
uint instructions) {
if (!dev.settings().splitSizeForWin7_) {
dispatchSplitSize_ = 0;
return;
}
uint64_t workload = threads * instructions;
if (maxDispatchWorkload_ < workload) {
dispatchSplitSize_ = static_cast<uint>(maxDispatchWorkload_ / instructions);
@@ -2258,6 +2263,8 @@ bool VirtualGPU::submitKernelInternal(const amd::NDRangeContainer& sizes, const
dispatchParam.wavesPerSh = (enqueueEvent != nullptr) ?
enqueueEvent->profilingInfo().waves_ : 0;
dispatchParam.useAtc = dev().settings().svmFineGrainSystem_ ? true : false;
dispatchParam.workitemPrivateSegmentSize = hsaKernel.spillSegSize();
dispatchParam.kernargSegmentSize = hsaKernel.argsBufferSize();
// Run AQL dispatch in HW
eventBegin(MainEngine);
iCmd()->CmdDispatchAql(dispatchParam);