Skip to content

Conversation

@joeatodd
Copy link
Contributor

This new API returns the maximum work-groups which can occupy a single compute unit of the given device.

@joeatodd joeatodd requested a review from a team as a code owner October 22, 2024 11:59
size_t max_wgs = kernel.template ext_oneapi_get_info<
syclex::info::kernel_queue_specific::max_num_work_groups>(
q, sycl::range<3>{syclcompat::dim3{wg_range}}, local_mem_size);
assert(max_per_cu == max_wgs / max_compute_units);
Copy link
Contributor

@AD2605 AD2605 Oct 22, 2024

Choose a reason for hiding this comment

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

do we expect NDEBUG to not to defined during testing ?
Does the CI build and test DPCPP with a specific configuration or is the build type string left empty ?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yes, asserts work as expected during testing. llvm-lit will test on as many backends as are available, AFAIK.

Copy link
Contributor

Choose a reason for hiding this comment

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

thanks

Copy link
Contributor

@Alcpz Alcpz left a comment

Choose a reason for hiding this comment

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

LGTM 🚢

@joeatodd
Copy link
Contributor Author

@intel/llvm-gatekeepers this is ready to merge 🙏

@martygrant martygrant merged commit 4cc64bd into intel:sycl Oct 23, 2024
13 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants