frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

Migrate the rest of CUDA 12.1 jobs to 12.4 (#144118)

Browse Source 
CUDA 12.4 is the default now and we don't build nightly 12.1 anymore, so it's time to move the rest of CI jobs to 12.4.  I also clean up some redundant CI jobs on periodic and inductor-periodic.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/144118
Approved by: https://github.com/atalman
This commit is contained in:
Huy Do
2025-01-03 17:45:41 +00:00
committed by PyTorch MergeBot
parent cbdc70ae07
commit f3968373c1
6 changed files with 84 additions and 186 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
.ci/docker/build.sh
View File

@ -208,20 +208,6 @@ case "$image" in
CONDA_CMAKE=yes CONDA_CMAKE=yes
TRITON=yes TRITON=yes
;; ;;
pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9)
CUDA_VERSION=12.4.1
CUDNN_VERSION=9
ANACONDA_PYTHON_VERSION=3.10
GCC_VERSION=9
PROTOBUF=yes
DB=yes
VISION=yes
KATEX=yes
UCX_COMMIT=${_UCX_COMMIT}
UCC_COMMIT=${_UCC_COMMIT}
CONDA_CMAKE=yes
TRITON=yes
;;
pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9) pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9)
CUDA_VERSION=12.1.1 CUDA_VERSION=12.1.1
CUDNN_VERSION=9 CUDNN_VERSION=9
@ -236,20 +222,6 @@ case "$image" in
CONDA_CMAKE=yes CONDA_CMAKE=yes
TRITON=yes TRITON=yes
;; ;;
pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9)
CUDA_VERSION=12.4.1
CUDNN_VERSION=9
ANACONDA_PYTHON_VERSION=3.10
GCC_VERSION=9
PROTOBUF=yes
DB=yes
VISION=yes
KATEX=yes
UCX_COMMIT=${_UCX_COMMIT}
UCC_COMMIT=${_UCC_COMMIT}
CONDA_CMAKE=yes
TRITON=yes
;;
pytorch-linux-focal-py3-clang10-onnx) pytorch-linux-focal-py3-clang10-onnx)
ANACONDA_PYTHON_VERSION=3.9 ANACONDA_PYTHON_VERSION=3.9
CLANG_VERSION=10 CLANG_VERSION=10

20
.github/workflows/inductor-micro-benchmark.yml vendored
View File

@ -37,16 +37,16 @@ jobs:
curr_ref_type: ${{ github.ref_type }} curr_ref_type: ${{ github.ref_type }}
check_experiments: "awsa100" check_experiments: "awsa100"
linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-build: linux-focal-cuda12_4-py3_10-gcc9-inductor-micro-benchmark-build:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
needs: needs:
- get-default-label-prefix - get-default-label-prefix
- get-a100-test-label-type - get-a100-test-label-type
with: with:
runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}" runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}"
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
cuda-arch-list: '8.0' cuda-arch-list: '8.0'
test-matrix: | test-matrix: |
{ include: [ { include: [
@ -54,14 +54,14 @@ jobs:
]} ]}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-test: linux-focal-cuda12_4-py3_10-gcc9-inductor-micro-benchmark-test:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-build needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-micro-benchmark-build
with: with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-micro-benchmark-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-micro-benchmark-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-micro-benchmark-build.outputs.test-matrix }}
use-gha: anything-non-empty-to-use-gha use-gha: anything-non-empty-to-use-gha
timeout-minutes: 720 timeout-minutes: 720
secrets: inherit secrets: inherit

20
.github/workflows/inductor-perf-compare.yml vendored
View File

@ -34,16 +34,16 @@ jobs:
curr_ref_type: ${{ github.ref_type }} curr_ref_type: ${{ github.ref_type }}
check_experiments: "awsa100" check_experiments: "awsa100"
linux-focal-cuda12_1-py3_10-gcc9-inductor-build: linux-focal-cuda12_4-py3_10-gcc9-inductor-build:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
needs: needs:
- get-default-label-prefix - get-default-label-prefix
- get-test-label-type - get-test-label-type
with: with:
runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}" runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}"
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
cuda-arch-list: '8.0' cuda-arch-list: '8.0'
test-matrix: | test-matrix: |
{ include: [ { include: [
@ -54,14 +54,14 @@ jobs:
]} ]}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-test: linux-focal-cuda12_4-py3_10-gcc9-inductor-test:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_1-py3_10-gcc9-inductor-build needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
with: with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.test-matrix }}
use-gha: anything-non-empty-to-use-gha use-gha: anything-non-empty-to-use-gha
# disable monitor in perf tests for more investigation # disable monitor in perf tests for more investigation
disable-monitor: true disable-monitor: true

44
.github/workflows/inductor-perf-test-nightly.yml vendored
View File

@ -77,14 +77,14 @@ jobs:
curr_ref_type: ${{ github.ref_type }} curr_ref_type: ${{ github.ref_type }}
# NB: Keep this in sync with trunk.yml # NB: Keep this in sync with trunk.yml
linux-focal-cuda12_1-py3_10-gcc9-inductor-build: linux-focal-cuda12_4-py3_10-gcc9-inductor-build:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
needs: get-label-type needs: get-label-type
with: with:
runner_prefix: "${{ needs.get-label-type.outputs.label-type }}" runner_prefix: "${{ needs.get-label-type.outputs.label-type }}"
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
cuda-arch-list: '8.0' cuda-arch-list: '8.0'
test-matrix: | test-matrix: |
{ include: [ { include: [
@ -104,48 +104,48 @@ jobs:
selected-test-configs: ${{ inputs.benchmark_configs }} selected-test-configs: ${{ inputs.benchmark_configs }}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-test-nightly: linux-focal-cuda12_4-py3_10-gcc9-inductor-test-nightly:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_1-py3_10-gcc9-inductor-build needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
if: github.event.schedule == '0 7 * * 1-6' if: github.event.schedule == '0 7 * * 1-6'
with: with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
dashboard-tag: training-true-inference-true-default-true-dynamic-true-cudagraphs-true-aotinductor-true-freezing_cudagraphs-true-cudagraphs_low_precision-true dashboard-tag: training-true-inference-true-default-true-dynamic-true-cudagraphs-true-aotinductor-true-freezing_cudagraphs-true-cudagraphs_low_precision-true
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.test-matrix }}
use-gha: anything-non-empty-to-use-gha use-gha: anything-non-empty-to-use-gha
timeout-minutes: 720 timeout-minutes: 720
# disable monitor in perf tests for more investigation # disable monitor in perf tests for more investigation
disable-monitor: true disable-monitor: true
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-test-weekly: linux-focal-cuda12_4-py3_10-gcc9-inductor-test-weekly:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_1-py3_10-gcc9-inductor-build needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
if: github.event.schedule == '0 7 * * 0' if: github.event.schedule == '0 7 * * 0'
with: with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
dashboard-tag: training-true-inference-true-default-true-dynamic-true-cudagraphs-true-aotinductor-true-freezing_cudagraphs-true-maxautotune-true-freeze_autotune_cudagraphs-true-cudagraphs_low_precision-true dashboard-tag: training-true-inference-true-default-true-dynamic-true-cudagraphs-true-aotinductor-true-freezing_cudagraphs-true-maxautotune-true-freeze_autotune_cudagraphs-true-cudagraphs_low_precision-true
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.test-matrix }}
use-gha: anything-non-empty-to-use-gha use-gha: anything-non-empty-to-use-gha
timeout-minutes: 1440 timeout-minutes: 1440
# disable monitor in perf tests for more investigation # disable monitor in perf tests for more investigation
disable-monitor: true disable-monitor: true
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-test: linux-focal-cuda12_4-py3_10-gcc9-inductor-test:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_1-py3_10-gcc9-inductor-build needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
if: github.event_name == 'workflow_dispatch' if: github.event_name == 'workflow_dispatch'
with: with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
dashboard-tag: training-${{ inputs.training }}-inference-${{ inputs.inference }}-default-${{ inputs.default }}-dynamic-${{ inputs.dynamic }}-cudagraphs-${{ inputs.cudagraphs }}-cppwrapper-false-aotinductor-${{ inputs.aotinductor }}-maxautotune-${{ inputs.maxautotune }}-freezing_cudagraphs-${{ inputs.freezing_cudagraphs }}-cudagraphs_low_precision-${{ inputs.cudagraphs }} dashboard-tag: training-${{ inputs.training }}-inference-${{ inputs.inference }}-default-${{ inputs.default }}-dynamic-${{ inputs.dynamic }}-cudagraphs-${{ inputs.cudagraphs }}-cppwrapper-false-aotinductor-${{ inputs.aotinductor }}-maxautotune-${{ inputs.maxautotune }}-freezing_cudagraphs-${{ inputs.freezing_cudagraphs }}-cudagraphs_low_precision-${{ inputs.cudagraphs }}
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.test-matrix }}
use-gha: anything-non-empty-to-use-gha use-gha: anything-non-empty-to-use-gha
timeout-minutes: 720 timeout-minutes: 720
# disable monitor in perf tests for more investigation # disable monitor in perf tests for more investigation

82
.github/workflows/inductor-periodic.yml vendored
View File

@ -39,14 +39,14 @@ jobs:
curr_ref_type: ${{ github.ref_type }} curr_ref_type: ${{ github.ref_type }}
check_experiments: "awsa100" check_experiments: "awsa100"
linux-focal-cuda12_1-py3_10-gcc9-periodic-dynamo-benchmarks-build: linux-focal-cuda12_4-py3_10-gcc9-periodic-dynamo-benchmarks-build:
name: cuda12.1-py3.10-gcc9-sm86-periodic-dynamo-benchmarks name: cuda12.4-py3.10-gcc9-sm86-periodic-dynamo-benchmarks
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
needs: get-default-label-prefix needs: get-default-label-prefix
with: with:
runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}" runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}"
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm86 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
cuda-arch-list: '8.6' cuda-arch-list: '8.6'
test-matrix: | test-matrix: |
{ include: [ { include: [
@ -68,26 +68,26 @@ jobs:
]} ]}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-periodic-dynamo-benchmarks-test: linux-focal-cuda12_4-py3_10-gcc9-periodic-dynamo-benchmarks-test:
name: cuda12.1-py3.10-gcc9-sm86-periodic-dynamo-benchmarks name: cuda12.4-py3.10-gcc9-sm86-periodic-dynamo-benchmarks
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_1-py3_10-gcc9-periodic-dynamo-benchmarks-build needs: linux-focal-cuda12_4-py3_10-gcc9-periodic-dynamo-benchmarks-build
with: with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm86 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-periodic-dynamo-benchmarks-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-periodic-dynamo-benchmarks-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-periodic-dynamo-benchmarks-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-periodic-dynamo-benchmarks-build.outputs.test-matrix }}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-build-gcp: linux-focal-cuda12_4-py3_10-gcc9-inductor-build-gcp:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
needs: needs:
- get-default-label-prefix - get-default-label-prefix
- get-a100-test-label-type - get-a100-test-label-type
with: with:
runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}" runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}"
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
cuda-arch-list: '8.0' cuda-arch-list: '8.0'
test-matrix: | test-matrix: |
{ include: [ { include: [
@ -95,14 +95,14 @@ jobs:
]} ]}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-test-gcp: linux-focal-cuda12_4-py3_10-gcc9-inductor-test-gcp:
name: cuda12.1-py3.10-gcc9-sm80 name: cuda12.4-py3.10-gcc9-sm80
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_1-py3_10-gcc9-inductor-build-gcp needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build-gcp
with: with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm80 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm80
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build-gcp.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build-gcp.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build-gcp.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build-gcp.outputs.test-matrix }}
use-gha: anything-non-empty-to-use-gha use-gha: anything-non-empty-to-use-gha
# disable monitor in smoke perf tests for more investigation # disable monitor in smoke perf tests for more investigation
disable-monitor: true disable-monitor: true
@ -143,52 +143,16 @@ jobs:
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-build:
name: cuda12.1-py3.10-gcc9-sm86
uses: ./.github/workflows/_linux-build.yml
needs: get-default-label-prefix
with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm86
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9-inductor-benchmarks
cuda-arch-list: '8.6'
runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}"
sync-tag: linux-focal-cuda12_1-py3_10-gcc9-inductor-build
test-matrix: |
{ include: [
{ config: "dynamic_inductor_huggingface", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "dynamic_inductor_timm", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "dynamic_inductor_timm", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "dynamic_inductor_torchbench", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "dynamic_inductor_torchbench", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "aot_inductor_huggingface", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "aot_inductor_timm", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "aot_inductor_timm", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "aot_inductor_torchbench", shard: 1, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
{ config: "aot_inductor_torchbench", shard: 2, num_shards: 2, runner: "linux.g5.4xlarge.nvidia.gpu" },
]}
secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-inductor-test:
name: cuda12.1-py3.10-gcc9-sm86
uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_1-py3_10-gcc9-inductor-build
with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-sm86
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-inductor-build.outputs.test-matrix }}
secrets: inherit
linux-focal-cuda12_4-py3_10-gcc9-inductor-build: linux-focal-cuda12_4-py3_10-gcc9-inductor-build:
# Should be synced with the benchmark tests in inductor.yml, but this doesn't run inductor_timm
name: cuda12.4-py3.10-gcc9-sm86 name: cuda12.4-py3.10-gcc9-sm86
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
needs: get-default-label-prefix needs: get-default-label-prefix
with: with:
runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}"
sync-tag: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86
docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9-inductor-benchmarks
cuda-arch-list: '8.6' cuda-arch-list: '8.6'
runner_prefix: "${{ needs.get-default-label-prefix.outputs.label-type }}"
sync-tag: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
test-matrix: | test-matrix: |
{ include: [ { include: [
{ config: "dynamic_inductor_huggingface", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" }, { config: "dynamic_inductor_huggingface", shard: 1, num_shards: 1, runner: "linux.g5.4xlarge.nvidia.gpu" },
@ -209,13 +173,11 @@ jobs:
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build needs: linux-focal-cuda12_4-py3_10-gcc9-inductor-build
with: with:
sync-tag: linux-focal-cuda12_4-py3_10-gcc9-inductor-test
build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86 build-environment: linux-focal-cuda12.4-py3.10-gcc9-sm86
docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-inductor-build.outputs.test-matrix }}
secrets: inherit secrets: inherit
linux-jammy-cpu-py3_9-gcc11-inductor-build: linux-jammy-cpu-py3_9-gcc11-inductor-build:
name: linux-jammy-cpu-py3.9-gcc11-inductor name: linux-jammy-cpu-py3.9-gcc11-inductor
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml

76
.github/workflows/periodic.yml vendored
View File

@ -49,36 +49,6 @@ jobs:
curr_branch: ${{ github.head_ref || github.ref_name }} curr_branch: ${{ github.head_ref || github.ref_name }}
curr_ref_type: ${{ github.ref_type }} curr_ref_type: ${{ github.ref_type }}
linux-focal-cuda12_1-py3_10-gcc9-build:
name: linux-focal-cuda12.1-py3.10-gcc9
uses: ./.github/workflows/_linux-build.yml
needs: get-label-type
with:
runner_prefix: "${{ needs.get-label-type.outputs.label-type }}"
build-environment: linux-focal-cuda12.1-py3.10-gcc9
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9
test-matrix: |
{ include: [
{ config: "nogpu_AVX512", shard: 1, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" },
{ config: "nogpu_AVX512", shard: 2, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" },
{ config: "nogpu_NO_AVX2", shard: 1, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" },
{ config: "nogpu_NO_AVX2", shard: 2, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" },
{ config: "jit_legacy", shard: 1, num_shards: 1, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge.nvidia.gpu" },
]}
secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-test:
name: linux-focal-cuda12.1-py3.10-gcc9
uses: ./.github/workflows/_linux-test.yml
needs:
- linux-focal-cuda12_1-py3_10-gcc9-build
- target-determination
with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-build.outputs.test-matrix }}
secrets: inherit
linux-focal-cuda12_4-py3_10-gcc9-build: linux-focal-cuda12_4-py3_10-gcc9-build:
name: linux-focal-cuda12.4-py3.10-gcc9 name: linux-focal-cuda12.4-py3.10-gcc9
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
@ -89,11 +59,6 @@ jobs:
docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9 docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9
test-matrix: | test-matrix: |
{ include: [ { include: [
{ config: "default", shard: 1, num_shards: 5, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge.nvidia.gpu" },
{ config: "default", shard: 2, num_shards: 5, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge.nvidia.gpu" },
{ config: "default", shard: 3, num_shards: 5, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge.nvidia.gpu" },
{ config: "default", shard: 4, num_shards: 5, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge.nvidia.gpu" },
{ config: "default", shard: 5, num_shards: 5, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge.nvidia.gpu" },
{ config: "nogpu_AVX512", shard: 1, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" }, { config: "nogpu_AVX512", shard: 1, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" },
{ config: "nogpu_AVX512", shard: 2, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" }, { config: "nogpu_AVX512", shard: 2, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" },
{ config: "nogpu_NO_AVX2", shard: 1, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" }, { config: "nogpu_NO_AVX2", shard: 1, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" },
@ -109,7 +74,6 @@ jobs:
- linux-focal-cuda12_4-py3_10-gcc9-build - linux-focal-cuda12_4-py3_10-gcc9-build
- target-determination - target-determination
with: with:
timeout-minutes: 360
build-environment: linux-focal-cuda12.4-py3.10-gcc9 build-environment: linux-focal-cuda12.4-py3.10-gcc9
docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-build.outputs.test-matrix }}
@ -206,16 +170,16 @@ jobs:
test-matrix: ${{ needs.linux-focal-rocm6_2-py3_10-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-rocm6_2-py3_10-build.outputs.test-matrix }}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-experimental-split-build: linux-focal-cuda12_4-py3_10-gcc9-experimental-split-build:
name: linux-focal-cuda12.1-py3.10-gcc9-experimental-split-build name: linux-focal-cuda12.4-py3.10-gcc9-experimental-split-build
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
needs: get-label-type needs: get-label-type
if: false # See https://github.com/pytorch/pytorch/issues/138750 if: false # See https://github.com/pytorch/pytorch/issues/138750
with: with:
runner_prefix: "${{ needs.get-label-type.outputs.label-type }}" runner_prefix: "${{ needs.get-label-type.outputs.label-type }}"
use_split_build: true use_split_build: true
build-environment: linux-focal-cuda12.1-py3.10-gcc9 build-environment: linux-focal-cuda12.4-py3.10-gcc9
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9 docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9
test-matrix: | test-matrix: |
{ include: [ { include: [
{ config: "nogpu_AVX512", shard: 1, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" }, { config: "nogpu_AVX512", shard: 1, num_shards: 2, runner: "${{ needs.get-label-type.outputs.label-type }}linux.4xlarge" },
@ -226,16 +190,16 @@ jobs:
]} ]}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3_10-gcc9-experimental-split-build-test: linux-focal-cuda12_4-py3_10-gcc9-experimental-split-build-test:
name: linux-focal-cuda12.1-py3.10-gcc9-experimental-split-build name: linux-focal-cuda12.4-py3.10-gcc9-experimental-split-build
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: needs:
- linux-focal-cuda12_1-py3_10-gcc9-experimental-split-build - linux-focal-cuda12_4-py3_10-gcc9-experimental-split-build
- target-determination - target-determination
with: with:
build-environment: linux-focal-cuda12.1-py3.10-gcc9-experimental-split-build build-environment: linux-focal-cuda12.4-py3.10-gcc9-experimental-split-build
docker-image: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-experimental-split-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-experimental-split-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3_10-gcc9-experimental-split-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3_10-gcc9-experimental-split-build.outputs.test-matrix }}
secrets: inherit secrets: inherit
@ -301,14 +265,14 @@ jobs:
test-matrix: ${{ needs.linux-focal-cuda11_8-py3_10-gcc9-experimental-split-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda11_8-py3_10-gcc9-experimental-split-build.outputs.test-matrix }}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build: linux-focal-cuda12_4-py3-gcc9-slow-gradcheck-build:
name: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck name: linux-focal-cuda12.4-py3-gcc9-slow-gradcheck
uses: ./.github/workflows/_linux-build.yml uses: ./.github/workflows/_linux-build.yml
needs: get-label-type needs: get-label-type
with: with:
runner_prefix: "${{ needs.get-label-type.outputs.label-type }}" runner_prefix: "${{ needs.get-label-type.outputs.label-type }}"
build-environment: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck build-environment: linux-focal-cuda12.4-py3-gcc9-slow-gradcheck
docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9 docker-image-name: pytorch-linux-focal-cuda12.4-cudnn9-py3-gcc9
cuda-arch-list: 8.6 cuda-arch-list: 8.6
test-matrix: | test-matrix: |
{ include: [ { include: [
@ -323,16 +287,16 @@ jobs:
]} ]}
secrets: inherit secrets: inherit
linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-test: linux-focal-cuda12_4-py3-gcc9-slow-gradcheck-test:
name: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck name: linux-focal-cuda12.4-py3-gcc9-slow-gradcheck
uses: ./.github/workflows/_linux-test.yml uses: ./.github/workflows/_linux-test.yml
needs: needs:
- linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build - linux-focal-cuda12_4-py3-gcc9-slow-gradcheck-build
- target-determination - target-determination
with: with:
build-environment: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck build-environment: linux-focal-cuda12.4-py3-gcc9-slow-gradcheck
docker-image: ${{ needs.linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build.outputs.docker-image }} docker-image: ${{ needs.linux-focal-cuda12_4-py3-gcc9-slow-gradcheck-build.outputs.docker-image }}
test-matrix: ${{ needs.linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build.outputs.test-matrix }} test-matrix: ${{ needs.linux-focal-cuda12_4-py3-gcc9-slow-gradcheck-build.outputs.test-matrix }}
timeout-minutes: 300 timeout-minutes: 300
secrets: inherit secrets: inherit