Skip to content

Jimwu.fix backend sampler hip multi output - #26878

Merged
ggerganov merged 3 commits into
ggml-org:masterfrom
AMD-Ecosystem:jimwu.fix-backend-sampler-hip-multi-output
Aug 11, 2026
Merged

Jimwu.fix backend sampler hip multi output#26878
ggerganov merged 3 commits into
ggml-org:masterfrom
AMD-Ecosystem:jimwu.fix-backend-sampler-hip-multi-output

Conversation

@jimw567

@jimw567 jimw567 commented Aug 11, 2026

Copy link
Copy Markdown
Contributor

Overview

  • Fixed a regression on gpu-rocm introduced by Backend sampling multi output #25532
  • Added hooks to create per run log directory while we are fixing the issue with log uploading from the rocm runner.

Requirements

  • I have read and agree with the contributing guidelines
  • AI usage disclosure: AI helped locate the PR leading to the regression

Jim Wu added 2 commits August 10, 2026 19:05
The new multi_output_sampling_chain test uses top_k, whose backend probs
path needs CUB (unavailable on HIP), so sampled_probs is null and the test
aborts. Add it to the existing HIP skip list alongside the other TOP_K tests.
The self-hosted gpu-rocm runner can't upload logs to Azure blob (egress
firewalled), so a run's logs were wiped by the next run. Write each run's
logs to $OUT/run-<run_id>-<attempt>/ so an Actions run URL maps to its logs.
@jimw567
jimw567 requested a review from ggerganov as a code owner August 11, 2026 01:09
@jimw567
jimw567 requested review from taronaeo and removed request for ggerganov August 11, 2026 01:09
@github-actions github-actions Bot added testing Everything test related devops improvements to build systems and github actions labels Aug 11, 2026
@taronaeo

Copy link
Copy Markdown
Member

Looks like its still failing.

Like the other TOP_K-based subtests, multi_output_cpu's backend sampler
never initializes on HIP (no CUB TOP_K), so it aborts. Add it to the skip list.
@jimw567

jimw567 commented Aug 11, 2026

Copy link
Copy Markdown
Contributor Author

Looks like its still failing.

It is green now. Please help review and merge.

@taronaeo taronaeo added the merge ready A maintainer can use this label to indicate that they consider the changes final and ready to merge. label Aug 11, 2026
@ggerganov
ggerganov merged commit 4801e3c into ggml-org:master Aug 11, 2026
23 of 28 checks passed
huaxel pushed a commit to huaxel/CachyLLama that referenced this pull request Aug 12, 2026
* test-backend-sampler: skip multi_output_sampling_chain on HIP

The new multi_output_sampling_chain test uses top_k, whose backend probs
path needs CUB (unavailable on HIP), so sampled_probs is null and the test
aborts. Add it to the existing HIP skip list alongside the other TOP_K tests.

* ci: keep gpu-rocm logs in a per-run dir keyed by GitHub run id

The self-hosted gpu-rocm runner can't upload logs to Azure blob (egress
firewalled), so a run's logs were wiped by the next run. Write each run's
logs to $OUT/run-<run_id>-<attempt>/ so an Actions run URL maps to its logs.

* test-backend-sampler: also skip multi_output_cpu on HIP

Like the other TOP_K-based subtests, multi_output_cpu's backend sampler
never initializes on HIP (no CUB TOP_K), so it aborts. Add it to the skip list.

---------

Co-authored-by: Jim Wu <ywu@xilinx.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

devops improvements to build systems and github actions merge ready A maintainer can use this label to indicate that they consider the changes final and ready to merge. testing Everything test related

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants