SWDEV-524746 - Part-II Add multi device support for hip graph. Updated kernel arg manager for each device (#813)

- Updated kernel arg manager to support allocating kernel args on multiple devices for single graph.
- Updated AQL path to capture on the device where graph node is added.

Co-authored-by: Anusha GodavarthySurya <Anusha.GodavarthySurya@amd.com>
这个提交包含在:
Godavarthy Surya, Anusha
2025-09-25 20:38:18 +05:30
提交者 GitHub
父节点 4a31affb76
当前提交 fb72d7f851
修改 5 个文件,包含 120 行新增72 行删除
@@ -3691,7 +3691,7 @@ bool VirtualGPU::submitKernelInternal(const amd::NDRangeContainer& sizes, const
// Allocate buffer to hold kernel arguments
if (isGraphCapture) {
argBuffer = command_->getGraphKernArg(gpuKernel.KernargSegmentByteSize(),
gpuKernel.KernargSegmentAlignment());
gpuKernel.KernargSegmentAlignment(), dev().index());
command_->SetKernelName(gpuKernel.getDemangledName().c_str());
} else {
ClPrint(amd::LOG_DETAIL_DEBUG, amd::LOG_KERN,