Skip to content

[HIP] Implement urKernelSuggestMaxCooperativeGroupCountExp for HIP #2617

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
jinz2014 wants to merge 2 commits into from
Closed

[HIP] Implement urKernelSuggestMaxCooperativeGroupCountExp for HIP #2617

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b332114
[HIP] Implement urKernelSuggestMaxCooperativeGroupCountExp for HIP
zjin-lcf Jan 24, 2025
387d818
[HIP] update the comments with the reviewer suggestions
zjin-lcf Jan 27, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 1 addition & 5 deletions source/adapters/hip/device.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,11 +58,7 @@ UR_APIEXPORT ur_result_t UR_APICALL urDeviceGetInfo(ur_device_handle_t hDevice,
return ReturnValue(VendorId);
}
case UR_DEVICE_INFO_MAX_COMPUTE_UNITS: {
int ComputeUnits = 0;
UR_CHECK_ERROR(hipDeviceGetAttribute(
&ComputeUnits, hipDeviceAttributeMultiprocessorCount, hDevice->get()));
detail::ur::assertion(ComputeUnits >= 0);
return ReturnValue(static_cast<uint32_t>(ComputeUnits));
return ReturnValue(hDevice->getNumComputeUnits());
}
case UR_DEVICE_INFO_MAX_WORK_ITEM_DIMENSIONS: {
return ReturnValue(MaxWorkItemDimensions);
Expand Down
6 changes: 6 additions & 0 deletions source/adapters/hip/device.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,6 +26,7 @@ struct ur_device_handle_t_ {
ur_platform_handle_t Platform;
hipEvent_t EvBase; // HIP event used as base counter
uint32_t DeviceIndex;
uint32_t NumComputeUnits{0};

int MaxWorkGroupSize{0};
int MaxBlockDimX{0};
Expand All @@ -41,6 +42,9 @@ struct ur_device_handle_t_ {
: HIPDevice(HipDevice), RefCount{1}, Platform(Platform), EvBase(EvBase),
DeviceIndex(DeviceIndex) {

UR_CHECK_ERROR(hipDeviceGetAttribute(
reinterpret_cast<int *>(&NumComputeUnits),
hipDeviceAttributeMultiprocessorCount, HIPDevice));
UR_CHECK_ERROR(hipDeviceGetAttribute(
&MaxWorkGroupSize, hipDeviceAttributeMaxThreadsPerBlock, HIPDevice));
UR_CHECK_ERROR(hipDeviceGetAttribute(
Expand Down Expand Up @@ -85,6 +89,8 @@ struct ur_device_handle_t_ {

int getManagedMemSupport() const noexcept { return ManagedMemSupport; };

uint32_t getNumComputeUnits() const noexcept { return NumComputeUnits; };

int getConcurrentManagedAccess() const noexcept {
return ConcurrentManagedAccess;
};
Expand Down
58 changes: 45 additions & 13 deletions source/adapters/hip/kernel.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -161,24 +161,56 @@ urKernelRelease(ur_kernel_handle_t hKernel) {
return UR_RESULT_SUCCESS;
}

// TODO(ur): Not implemented on hip atm. Also, need to add tests for this
// feature.
UR_APIEXPORT ur_result_t UR_APICALL
urKernelGetNativeHandle(ur_kernel_handle_t, ur_native_handle_t *) {
return UR_RESULT_ERROR_UNSUPPORTED_FEATURE;
}

UR_APIEXPORT ur_result_t UR_APICALL urKernelSuggestMaxCooperativeGroupCountExp(
ur_kernel_handle_t hKernel, ur_device_handle_t hDevice, uint32_t workDim,
const size_t *pLocalWorkSize, size_t dynamicSharedMemorySize,
uint32_t *pGroupCountRet) {
std::ignore = hKernel;
UR_ASSERT(hKernel, UR_RESULT_ERROR_INVALID_KERNEL);

std::ignore = hDevice;
std::ignore = workDim;
std::ignore = pLocalWorkSize;
std::ignore = dynamicSharedMemorySize;
std::ignore = pGroupCountRet;
return UR_RESULT_ERROR_UNSUPPORTED_FEATURE;

size_t localWorkSize = pLocalWorkSize[0];
localWorkSize *= (workDim >= 2 ? pLocalWorkSize[1] : 1);
localWorkSize *= (workDim == 3 ? pLocalWorkSize[2] : 1);

// We need to set the active current device for this kernel explicitly here,
// because the occupancy querying API does not take device parameter.
ur_device_handle_t Device = hKernel->getProgram()->getDevice();
ScopedDevice Active(Device);
try {
// We need to calculate max num of work-groups using per-device semantics.

int MaxNumActiveGroupsPerCU{0};
UR_CHECK_ERROR(hipOccupancyMaxActiveBlocksPerMultiprocessor(
&MaxNumActiveGroupsPerCU, hKernel->get(), localWorkSize,
dynamicSharedMemorySize));
detail::ur::assertion(MaxNumActiveGroupsPerCU >= 0);
// Handle the case where we can't have all SMs active with at least 1 group
Copy link
Copy Markdown
Contributor

@GeorgeWeb GeorgeWeb Jan 27, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think AMD calls these WGPs (Work-Group Processors). SM is for Nvidia architectures.

// per SM. In that case, the device is still able to run 1 work-group, hence
// we will manually check if it is possible with the available HW resources.
if (MaxNumActiveGroupsPerCU == 0) {
size_t MaxWorkGroupSize{};
urKernelGetGroupInfo(
hKernel, Device, UR_KERNEL_GROUP_INFO_WORK_GROUP_SIZE,
sizeof(MaxWorkGroupSize), &MaxWorkGroupSize, nullptr);
size_t MaxLocalSizeBytes{};
urDeviceGetInfo(Device, UR_DEVICE_INFO_LOCAL_MEM_SIZE,
sizeof(MaxLocalSizeBytes), &MaxLocalSizeBytes, nullptr);
if (localWorkSize > MaxWorkGroupSize ||
dynamicSharedMemorySize > MaxLocalSizeBytes)
*pGroupCountRet = 0;
else
*pGroupCountRet = 1;
} else {
// Multiply by the number of SMs (CUs = compute units) on the device in
// order to retreive the total number of groups/blocks that can be
// launched.
*pGroupCountRet = Device->getNumComputeUnits() * MaxNumActiveGroupsPerCU;
}
} catch (ur_result_t Err) {
return Err;
}
return UR_RESULT_SUCCESS;
}

UR_APIEXPORT ur_result_t UR_APICALL urKernelSetArgValue(
Expand Down