Skip to content

Commit 6eb80df

Browse files
huydhnhuydhn
committed
Move the GPU workflows to linux_job_v3
cuda.yml, cuda-perf.yml, cuda-windows.yml and vulkan.yml. The jobs that already ran on v3 only lose use-custom-docker-registry, which v3 ignores. Authored with Claude Code.
1 parent 872158d commit 6eb80df

4 files changed

Lines changed: 20 additions & 28 deletions

File tree

‎.github/workflows/cuda-perf.yml‎

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -138,7 +138,6 @@ jobs:
138138
runner: ${{ contains(matrix.model, 'Qwen3.5-35B-A3B') && 'mt-l-x86iavx512-11-125-a100' || 'mt-l-x86aavx2-29-113-a10g' }}
139139
gpu-arch-type: cuda
140140
gpu-arch-version: "13.0"
141-
use-custom-docker-registry: false
142141
submodules: recursive
143142
upload-artifact: model-${{ matrix.model_safe }}-${{ matrix.quant }}
144143
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
@@ -218,7 +217,6 @@ jobs:
218217
runner: ${{ contains(matrix.model, 'Qwen3.5-35B-A3B') && 'mt-l-x86iavx512-11-125-a100' || 'mt-l-x86aavx2-29-113-a10g' }}
219218
gpu-arch-type: cuda
220219
gpu-arch-version: "13.0"
221-
use-custom-docker-registry: false
222220
submodules: recursive
223221
download-artifact: model-${{ matrix.model_safe }}-${{ matrix.quant }}
224222
upload-artifact: results-${{ matrix.model_safe }}-${{ matrix.quant }}

‎.github/workflows/cuda-windows.yml‎

Lines changed: 8 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -27,6 +27,10 @@ permissions:
2727
contents: read
2828

2929
jobs:
30+
docker-image:
31+
name: Resolve CI docker image
32+
uses: ./.github/workflows/_docker-image.yml
33+
3034
changed-files:
3135
name: Get changed files
3236
uses: ./.github/workflows/_get-changed-files.yml
@@ -44,7 +48,7 @@ jobs:
4448
# filter so push commits that don't touch CUDA-relevant paths skip
4549
# this job on non-sampled commits. See _ci-run-decision.yml for
4650
# the sampling policy.
47-
needs: [changed-files, run-decision]
51+
needs: [docker-image, changed-files, run-decision]
4852
if: |
4953
(github.event.pull_request.head.repo.full_name == github.repository || github.event_name != 'pull_request') &&
5054
(
@@ -54,7 +58,7 @@ jobs:
5458
contains(needs.changed-files.outputs.changed-files, '.github/workflows/cuda-windows.yml') ||
5559
needs.run-decision.outputs.is-full-run == 'true'
5660
)
57-
uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
61+
uses: pytorch/test-infra/.github/workflows/linux_job_v3.yml@main
5862
permissions:
5963
id-token: write
6064
contents: read
@@ -89,10 +93,10 @@ jobs:
8993
with:
9094
timeout: 90
9195
secrets-env: EXECUTORCH_HF_TOKEN
92-
runner: linux.g5.4xlarge.nvidia.gpu
96+
runner: mt-l-x86aavx2-29-113-a10g
9397
gpu-arch-type: cuda
9498
gpu-arch-version: "13.0"
95-
docker-image: ci-image:executorch-ubuntu-22.04-cuda-windows
99+
docker-image: ${{ needs.docker-image.outputs.docker-registry }}/ci-image:executorch-ubuntu-22.04-cuda-windows-${{ needs.docker-image.outputs.ci-docker-hash }}
96100
submodules: recursive
97101
upload-artifact: ${{ matrix.model_repo }}-${{ matrix.model_name }}-cuda-windows-${{ matrix.quant }}
98102
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}

‎.github/workflows/cuda.yml‎

Lines changed: 10 additions & 19 deletions
Original file line numberDiff line numberDiff line change
@@ -67,16 +67,15 @@ jobs:
6767
cuda-version: ["12.6", "13.0"]
6868

6969
name: test-executorch-cuda-build-${{ matrix.cuda-version }}
70-
uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
70+
uses: pytorch/test-infra/.github/workflows/linux_job_v3.yml@main
7171
permissions:
7272
id-token: write
7373
contents: read
7474
with:
7575
timeout: 90
76-
runner: linux.g5.4xlarge.nvidia.gpu
76+
runner: mt-l-x86aavx2-29-113-a10g
7777
gpu-arch-type: cuda
7878
gpu-arch-version: ${{ matrix.cuda-version }}
79-
use-custom-docker-registry: false
8079
submodules: recursive
8180
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
8281
script: |
@@ -126,16 +125,15 @@ jobs:
126125
contains(needs.changed-files.outputs.changed-files, '.ci/scripts/export_model_artifact.sh') ||
127126
contains(needs.changed-files.outputs.changed-files, '.ci/scripts/test_model_e2e.sh') ||
128127
needs.run-decision.outputs.is-full-run == 'true'
129-
uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
128+
uses: pytorch/test-infra/.github/workflows/linux_job_v3.yml@main
130129
permissions:
131130
id-token: write
132131
contents: read
133132
with:
134133
timeout: 90
135-
runner: linux.g5.4xlarge.nvidia.gpu
134+
runner: mt-l-x86aavx2-29-113-a10g
136135
gpu-arch-type: cuda
137136
gpu-arch-version: "13.0"
138-
use-custom-docker-registry: false
139137
submodules: recursive
140138
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
141139
script: |
@@ -176,7 +174,6 @@ jobs:
176174
runner: mt-l-x86iavx512-11-125-a100
177175
gpu-arch-type: cuda
178176
gpu-arch-version: "13.0"
179-
use-custom-docker-registry: false
180177
submodules: recursive
181178
upload-artifact: cuda-target-smem-cross-arch
182179
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
@@ -220,7 +217,6 @@ jobs:
220217
runner: mt-l-x86aavx2-29-113-a10g
221218
gpu-arch-type: cuda
222219
gpu-arch-version: "13.0"
223-
use-custom-docker-registry: false
224220
submodules: recursive
225221
download-artifact: cuda-target-smem-cross-arch
226222
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
@@ -272,16 +268,15 @@ jobs:
272268
contains(needs.changed-files.outputs.changed-files, 'extension/pybindings') ||
273269
contains(needs.changed-files.outputs.changed-files, 'runtime/__init__.py') ||
274270
needs.run-decision.outputs.is-full-run == 'true'
275-
uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
271+
uses: pytorch/test-infra/.github/workflows/linux_job_v3.yml@main
276272
permissions:
277273
id-token: write
278274
contents: read
279275
with:
280276
timeout: 90
281-
runner: linux.g5.4xlarge.nvidia.gpu
277+
runner: mt-l-x86aavx2-29-113-a10g
282278
gpu-arch-type: cuda
283279
gpu-arch-version: "13.0"
284-
use-custom-docker-registry: false
285280
submodules: recursive
286281
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
287282
script: |
@@ -352,16 +347,15 @@ jobs:
352347
contains(needs.changed-files.outputs.changed-files, '.ci/scripts/export_model_artifact.sh') ||
353348
contains(needs.changed-files.outputs.changed-files, '.ci/scripts/test_model_e2e.sh') ||
354349
needs.run-decision.outputs.is-full-run == 'true'
355-
uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
350+
uses: pytorch/test-infra/.github/workflows/linux_job_v3.yml@main
356351
permissions:
357352
id-token: write
358353
contents: read
359354
with:
360355
timeout: 90
361-
runner: linux.g5.4xlarge.nvidia.gpu
356+
runner: mt-l-x86aavx2-29-113-a10g
362357
gpu-arch-type: cuda
363358
gpu-arch-version: "13.0"
364-
use-custom-docker-registry: false
365359
submodules: recursive
366360
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
367361
script: |
@@ -510,7 +504,6 @@ jobs:
510504
runner: ${{ (matrix.model.name == 'Qwen3.5-35B-A3B-HQQ-INT4' || matrix.model.name == 'gemma-4-31B-it-GGUF') && 'mt-l-x86iavx512-11-125-a100' || 'mt-l-x86aavx2-29-113-a10g' }}
511505
gpu-arch-type: cuda
512506
gpu-arch-version: "13.0"
513-
use-custom-docker-registry: false
514507
submodules: recursive
515508
upload-artifact: ${{ matrix.model.repo }}-${{ matrix.model.name }}-cuda-${{ matrix.quant }}
516509
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
@@ -585,7 +578,6 @@ jobs:
585578
runner: mt-l-x86iavx512-11-125-a100
586579
gpu-arch-type: cuda
587580
gpu-arch-version: "13.0"
588-
use-custom-docker-registry: false
589581
submodules: recursive
590582
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
591583
script: |
@@ -639,7 +631,7 @@ jobs:
639631
contains(needs.changed-files.outputs.changed-files, '.ci/scripts/test_model_e2e.sh') ||
640632
needs.run-decision.outputs.is-full-run == 'true'
641633
)
642-
uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
634+
uses: pytorch/test-infra/.github/workflows/linux_job_v3.yml@main
643635
permissions:
644636
id-token: write
645637
contents: read
@@ -661,10 +653,9 @@ jobs:
661653
timeout: 120
662654
secrets-env: EXECUTORCH_HF_TOKEN
663655
download-artifact: ${{ matrix.artifact }}
664-
runner: linux.g5.4xlarge.nvidia.gpu
656+
runner: mt-l-x86aavx2-29-113-a10g
665657
gpu-arch-type: cuda
666658
gpu-arch-version: "13.0"
667-
use-custom-docker-registry: false
668659
submodules: recursive
669660
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
670661
script: |

‎.github/workflows/vulkan.yml‎

Lines changed: 2 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -50,16 +50,15 @@ jobs:
5050
contains(needs.changed-files.outputs.changed-files, '.ci/scripts/setup-vulkan-linux-deps.sh') ||
5151
contains(needs.changed-files.outputs.changed-files, '.github/workflows/vulkan.yml') ||
5252
needs.run-decision.outputs.is-full-run == 'true'
53-
uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
53+
uses: pytorch/test-infra/.github/workflows/linux_job_v3.yml@main
5454
permissions:
5555
id-token: write
5656
contents: read
5757
with:
5858
timeout: 120
59-
runner: linux.g5.4xlarge.nvidia.gpu
59+
runner: mt-l-x86aavx2-29-113-a10g
6060
gpu-arch-type: cuda
6161
gpu-arch-version: "12.6"
62-
use-custom-docker-registry: false
6362
submodules: recursive
6463
ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
6564
script: |

0 commit comments

Comments
 (0)