[GPU] accuracy issue fix with fp16 for gemm_tiled_opt kernel #33940

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

Jump to bottom

Open

sungeunk wants to merge 1 commit into openvinotoolkit:master from sungeunk:179229_gemm_tiled_opt_f16_acc_issue

+31 −12

Contributor

sungeunk commented Feb 3, 2026

Description of the issue(symptom, root-cause, how it was resolved)

Fixed the use of float accumulators for intermediate calculations.
Corrected the use of float inputs for MAD operations.

The code and line that caused this issue (if it is not changed directly)

src\plugins\intel_gpu\src\kernel_selector\cl_kernels\gemm_tiled_opt.cl

Reproduction step and snapshot (if applicable. Do not attach for customer model)

reproducer is attached in the ticket.

Checklist

Is it a proper fix? (not a workaround)
Did you include test case for this fix, if necessary?
Did you review existing test that can be extended to cover this scenario? Which test did you review?

Tickets:

179229


          accuracy issue fix for fp16

685ccb1

- fixed float accumulator for intermediate calc
- fixed float for mad inputs

sungeunk added the category: GPU label

sungeunk requested review from a team as code owners

February 3, 2026 08:31

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels