Repository navigation
Conversation
rjodinchr
requested changes
Oct 20, 2026
Comment on lines
+72
to
+86
| error = clGetKernelWorkGroupInfo(kernel, device, CL_KERNEL_WORK_GROUP_SIZE, | ||
| sizeof(maxWorkGroupSize), | ||
| &maxWorkGroupSize, NULL); | ||
| test_error(error, "Unable to get max work group size from kernel"); | ||
|
|
||
| cl_uint maxThreadDim = 0; | ||
| error = clGetDeviceInfo(device, CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS, | ||
| sizeof(maxThreadDim), &maxThreadDim, NULL); | ||
| test_error(error, "Unable to get max work item dimensions from device"); | ||
|
|
||
| std::vector<size_t> workGroupMaxSizes(maxThreadDim, 0); | ||
| error = clGetDeviceInfo(device, CL_DEVICE_MAX_WORK_GROUP_SIZES, | ||
| sizeof(size_t) * workGroupMaxSizes.size(), | ||
| workGroupMaxSizes.data(), NULL); | ||
| test_error(error, "Unable to get max work group sizes from device"); |
Collaborator
There was a problem hiding this comment.
I think we can use get_max_allowed_1d_work_group_size_on_device(device, kernel, &maxWorkGroupSize) from harness/kernelHelpers.h here instead of querying CL_KERNEL_WORK_GROUP_SIZE, CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS, and CL_DEVICE_MAX_WORK_GROUP_SIZES manually.
Suggested change
| error = clGetKernelWorkGroupInfo(kernel, device, CL_KERNEL_WORK_GROUP_SIZE, | |
| sizeof(maxWorkGroupSize), | |
| &maxWorkGroupSize, NULL); | |
| test_error(error, "Unable to get max work group size from kernel"); | |
| cl_uint maxThreadDim = 0; | |
| error = clGetDeviceInfo(device, CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS, | |
| sizeof(maxThreadDim), &maxThreadDim, NULL); | |
| test_error(error, "Unable to get max work item dimensions from device"); | |
| std::vector<size_t> workGroupMaxSizes(maxThreadDim, 0); | |
| error = clGetDeviceInfo(device, CL_DEVICE_MAX_WORK_GROUP_SIZES, | |
| sizeof(size_t) * workGroupMaxSizes.size(), | |
| workGroupMaxSizes.data(), NULL); | |
| test_error(error, "Unable to get max work group sizes from device"); | |
| error = get_max_allowed_1d_work_group_size_on_device(device, kernel, | |
| &maxWorkGroupSize); | |
| test_error(error, "get_max_allowed_1d_work_group_size_on_device failed"); |
Comment on lines
+34
to
+54
| const Version version = get_device_cl_version(device); | ||
| cl_bool nonUniformSupport = CL_FALSE; | ||
| if (version >= Version(3, 0)) | ||
| { | ||
| error = clGetDeviceInfo( | ||
| device, CL_DEVICE_NON_UNIFORM_WORK_GROUP_SUPPORT, | ||
| sizeof(nonUniformSupport), &nonUniformSupport, NULL); | ||
| test_error(error, | ||
| "Unable to query CL_DEVICE_NON_UNIFORM_WORK_GROUP_SUPPORT"); | ||
| } | ||
| else if (version >= Version(2, 0)) | ||
| { | ||
| nonUniformSupport = CL_TRUE; | ||
| } | ||
|
|
||
| if (nonUniformSupport == CL_FALSE) | ||
| { | ||
| log_info("Device does not support non-uniform work-groups, skipping " | ||
| "test\n"); | ||
| return TEST_SKIPPED_ITSELF; | ||
| } |
Collaborator
There was a problem hiding this comment.
Do we actually need to skip when nonUniformSupport == CL_FALSE? Even on OpenCL 1.2 devices (or OpenCL 3.0 devices without non-uniform work-group support), clEnqueueNDRangeKernel is still required to return CL_INVALID_WORK_GROUP_SIZE for this enqueue. Removing this check would simplify the test and also cover OpenCL 1.2 devices (which aren't covered by consistency_non_uniform_work_group or options_uniform_work_group_size).
Contributor
Author
There was a problem hiding this comment.
yep, good point - corrected
rjodinchr
previously approved these changes
Oct 20, 2026
Collaborator
|
Please rebase |
shajder
force-pushed
the
api_negative_non_uniform_invalid_work_size
branch
from
October 21, 2026 07:58
d490fa4 to
a852383
Compare
rjodinchr
approved these changes
Oct 21, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Related to #2790.
Checks that
clEnqueueNDRangeKernelreturnsCL_INVALID_WORK_GROUP_SIZEfor a non-uniform work-group when the program is built from source without -cl-std, so it is compiled for OpenCL C 1.x, on a device that supports non-uniform work-groups.[run-test: test_api negative_enqueue_non_uniform_work_group_size]