Skip to content

[UR][OpenCL][EnqueueCommandBuffer] use sync points even if command buffer's internal queue is in-order #18147

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
wenju-he wants to merge 2 commits into from
Closed

[UR][OpenCL][EnqueueCommandBuffer] use sync points even if command buffer's internal queue is in-order #18147

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9ece879
[UR][OpenCL][EnqueueCommandBuffer] Use internal in-order queue if com…
wenju-he Apr 23, 2025
f0b410a
revert 9ece879, always use pSyncPoint
wenju-he Apr 23, 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
40 changes: 15 additions & 25 deletions unified-runtime/source/adapters/opencl/command_buffer.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -171,11 +171,9 @@ UR_APIEXPORT ur_result_t UR_APICALL urCommandBufferAppendKernelLaunchExp(
cl_command_properties_khr *Properties =
hCommandBuffer->IsUpdatable ? UpdateProperties : nullptr;

const bool IsInOrder = hCommandBuffer->IsInOrder;
cl_sync_point_khr *RetSyncPoint = IsInOrder ? nullptr : pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList =
IsInOrder ? nullptr : pSyncPointWaitList;
uint32_t WaitListSize = IsInOrder ? 0 : numSyncPointsInWaitList;
cl_sync_point_khr *RetSyncPoint = pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList = pSyncPointWaitList;
uint32_t WaitListSize = numSyncPointsInWaitList;
CL_RETURN_ON_FAILURE(clCommandNDRangeKernelKHR(
hCommandBuffer->CLCommandBuffer, nullptr, Properties, hKernel->CLKernel,
workDim, pGlobalWorkOffset, pGlobalWorkSize, pLocalWorkSize, WaitListSize,
Expand Down Expand Up @@ -246,11 +244,9 @@ UR_APIEXPORT ur_result_t UR_APICALL urCommandBufferAppendMemBufferCopyExp(
CLContext, ur::cl::getAdapter()->fnCache.clCommandCopyBufferKHRCache,
cl_ext::CommandCopyBufferName, &clCommandCopyBufferKHR));

const bool IsInOrder = hCommandBuffer->IsInOrder;
cl_sync_point_khr *RetSyncPoint = IsInOrder ? nullptr : pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList =
IsInOrder ? nullptr : pSyncPointWaitList;
uint32_t WaitListSize = IsInOrder ? 0 : numSyncPointsInWaitList;
cl_sync_point_khr *RetSyncPoint = pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList = pSyncPointWaitList;
uint32_t WaitListSize = numSyncPointsInWaitList;
CL_RETURN_ON_FAILURE(clCommandCopyBufferKHR(
hCommandBuffer->CLCommandBuffer, nullptr, nullptr, hSrcMem->CLMemory,
hDstMem->CLMemory, srcOffset, dstOffset, size, WaitListSize,
Expand Down Expand Up @@ -289,11 +285,9 @@ UR_APIEXPORT ur_result_t UR_APICALL urCommandBufferAppendMemBufferCopyRectExp(
ur::cl::getAdapter()->fnCache.clCommandCopyBufferRectKHRCache,
cl_ext::CommandCopyBufferRectName, &clCommandCopyBufferRectKHR));

const bool IsInOrder = hCommandBuffer->IsInOrder;
cl_sync_point_khr *RetSyncPoint = IsInOrder ? nullptr : pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList =
IsInOrder ? nullptr : pSyncPointWaitList;
uint32_t WaitListSize = IsInOrder ? 0 : numSyncPointsInWaitList;
cl_sync_point_khr *RetSyncPoint = pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList = pSyncPointWaitList;
uint32_t WaitListSize = numSyncPointsInWaitList;
CL_RETURN_ON_FAILURE(clCommandCopyBufferRectKHR(
hCommandBuffer->CLCommandBuffer, nullptr, nullptr, hSrcMem->CLMemory,
hDstMem->CLMemory, OpenCLOriginRect, OpenCLDstRect, OpenCLRegion,
Expand Down Expand Up @@ -397,11 +391,9 @@ UR_APIEXPORT ur_result_t UR_APICALL urCommandBufferAppendMemBufferFillExp(
CLContext, ur::cl::getAdapter()->fnCache.clCommandFillBufferKHRCache,
cl_ext::CommandFillBufferName, &clCommandFillBufferKHR));

const bool IsInOrder = hCommandBuffer->IsInOrder;
cl_sync_point_khr *RetSyncPoint = IsInOrder ? nullptr : pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList =
IsInOrder ? nullptr : pSyncPointWaitList;
uint32_t WaitListSize = IsInOrder ? 0 : numSyncPointsInWaitList;
cl_sync_point_khr *RetSyncPoint = pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList = pSyncPointWaitList;
uint32_t WaitListSize = numSyncPointsInWaitList;
CL_RETURN_ON_FAILURE(clCommandFillBufferKHR(
hCommandBuffer->CLCommandBuffer, nullptr, nullptr, hBuffer->CLMemory,
pPattern, patternSize, offset, size, WaitListSize, SyncPointWaitList,
Expand Down Expand Up @@ -766,11 +758,9 @@ ur_result_t UR_APICALL urCommandBufferAppendNativeCommandExp(
cl_ext::CommandBarrierWithWaitListName,
&clCommandBarrierWithWaitListKHR));

const bool IsInOrder = hCommandBuffer->IsInOrder;
cl_sync_point_khr *RetSyncPoint = IsInOrder ? nullptr : pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList =
IsInOrder ? nullptr : pSyncPointWaitList;
uint32_t WaitListSize = IsInOrder ? 0 : numSyncPointsInWaitList;
cl_sync_point_khr *RetSyncPoint = pSyncPoint;
const cl_sync_point_khr *SyncPointWaitList = pSyncPointWaitList;
uint32_t WaitListSize = numSyncPointsInWaitList;
CL_RETURN_ON_FAILURE(clCommandBarrierWithWaitListKHR(
hCommandBuffer->CLCommandBuffer, nullptr, nullptr, WaitListSize,
SyncPointWaitList, nullptr, nullptr));
Expand Down
Loading