mirror of
https://github.com/saymrwulf/pytorch.git
synced 2026-05-14 20:57:59 +00:00
933b6d9830
Adds `cpp_wrapper` mode to the nightly inductor benchmark runs, as well as optionally for manually triggered runs. This is justified by `aot_inductor` already being in those runs. Additionally, re-enables `aot_inductor` in the nightly aarch64 runs. It was disabled 5 months ago to deal with a performance instability, which has likely gone away at this point. Pull Request resolved: https://github.com/pytorch/pytorch/pull/145791 Approved by: https://github.com/desertfire
143 lines
8.3 KiB
YAML
143 lines
8.3 KiB
YAML
name: inductor-perf-nightly-aarch64
|
|
|
|
on:
|
|
schedule:
|
|
# Does not perform max_autotune on CPU, so skip the weekly run setup
|
|
- cron: 0 7 * * *
|
|
# NB: GitHub has an upper limit of 10 inputs here
|
|
workflow_dispatch:
|
|
inputs:
|
|
training:
|
|
# CPU for training is not typical, but leave the option open here
|
|
description: Run training (off by default)?
|
|
required: false
|
|
type: boolean
|
|
default: false
|
|
inference:
|
|
description: Run inference (on by default)?
|
|
required: false
|
|
type: boolean
|
|
default: true
|
|
default:
|
|
description: Run inductor_default?
|
|
required: false
|
|
type: boolean
|
|
default: true
|
|
dynamic:
|
|
description: Run inductor_dynamic_shapes?
|
|
required: false
|
|
type: boolean
|
|
default: false
|
|
cppwrapper:
|
|
description: Run inductor_cpp_wrapper?
|
|
required: false
|
|
type: boolean
|
|
default: false
|
|
aotinductor:
|
|
description: Run aot_inductor for inference?
|
|
required: false
|
|
type: boolean
|
|
default: false
|
|
benchmark_configs:
|
|
description: The list of configs used the benchmark
|
|
required: false
|
|
type: string
|
|
default: inductor_huggingface_perf_cpu_aarch64,inductor_timm_perf_cpu_aarch64,inductor_torchbench_perf_cpu_aarch64
|
|
|
|
concurrency:
|
|
group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref_name }}-${{ github.ref_type == 'branch' && github.sha }}-${{ github.event_name == 'workflow_dispatch' }}-${{ github.event_name == 'schedule' }}
|
|
cancel-in-progress: true
|
|
|
|
permissions: read-all
|
|
|
|
jobs:
|
|
get-label-type:
|
|
name: get-label-type
|
|
uses: pytorch/pytorch/.github/workflows/_runner-determinator.yml@main
|
|
if: ${{ (github.event_name != 'schedule' || github.repository == 'pytorch/pytorch') && github.repository_owner == 'pytorch' }}
|
|
with:
|
|
triggering_actor: ${{ github.triggering_actor }}
|
|
issue_owner: ${{ github.event.pull_request.user.login || github.event.issue.user.login }}
|
|
curr_branch: ${{ github.head_ref || github.ref_name }}
|
|
curr_ref_type: ${{ github.ref_type }}
|
|
|
|
linux-jammy-aarch64-py3_10-inductor-build:
|
|
name: linux-jammy-aarch64-py3.10-inductor
|
|
uses: ./.github/workflows/_linux-build.yml
|
|
needs: get-label-type
|
|
with:
|
|
runner_prefix: "${{ needs.get-label-type.outputs.label-type }}"
|
|
runner: linux.arm64.m7g.4xlarge
|
|
build-environment: linux-jammy-aarch64-py3.10
|
|
docker-image-name: pytorch-linux-jammy-aarch64-py3.10-gcc11-inductor-benchmarks
|
|
test-matrix: |
|
|
{ include: [
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 1, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 2, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 3, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 4, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 5, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 6, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 7, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 8, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_huggingface_perf_cpu_aarch64", shard: 9, num_shards: 9, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 1, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 2, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 3, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 4, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 5, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 6, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 7, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 8, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 9, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 10, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 11, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 12, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 13, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 14, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_timm_perf_cpu_aarch64", shard: 15, num_shards: 15, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 1, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 2, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 3, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 4, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 5, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 6, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 7, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 8, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 9, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 10, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 11, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
{ config: "inductor_torchbench_perf_cpu_aarch64", shard: 12, num_shards: 12, runner: "linux.arm64.m7g.metal" },
|
|
]}
|
|
selected-test-configs: ${{ inputs.benchmark_configs }}
|
|
secrets: inherit
|
|
|
|
|
|
linux-jammy-aarch64-py3_10-inductor-test-nightly:
|
|
name: linux-jammy-aarch64-py3.10-inductor
|
|
uses: ./.github/workflows/_linux-test.yml
|
|
needs: linux-jammy-aarch64-py3_10-inductor-build
|
|
if: github.event.schedule == '0 7 * * *'
|
|
with:
|
|
build-environment: linux-jammy-aarch64-py3.10
|
|
dashboard-tag: training-false-inference-true-default-true-dynamic-true-cppwrapper-true-aotinductor-true
|
|
docker-image: ${{ needs.linux-jammy-aarch64-py3_10-inductor-build.outputs.docker-image }}
|
|
test-matrix: ${{ needs.linux-jammy-aarch64-py3_10-inductor-build.outputs.test-matrix }}
|
|
timeout-minutes: 720
|
|
# disable monitor in perf tests for more investigation
|
|
disable-monitor: true
|
|
secrets: inherit
|
|
|
|
|
|
linux-jammy-aarch64-py3_10-inductor-test:
|
|
name: linux-jammy-aarch64-py3.10-inductor
|
|
uses: ./.github/workflows/_linux-test.yml
|
|
needs: linux-jammy-aarch64-py3_10-inductor-build
|
|
if: github.event_name == 'workflow_dispatch'
|
|
with:
|
|
build-environment: linux-jammy-aarch64-py3.10
|
|
dashboard-tag: training-${{ inputs.training }}-inference-${{ inputs.inference }}-default-${{ inputs.default }}-dynamic-${{ inputs.dynamic }}-cppwrapper-${{ inputs.cppwrapper }}-aotinductor-${{ inputs.aotinductor }}
|
|
docker-image: ${{ needs.linux-jammy-aarch64-py3_10-inductor-build.outputs.docker-image }}
|
|
test-matrix: ${{ needs.linux-jammy-aarch64-py3_10-inductor-build.outputs.test-matrix }}
|
|
timeout-minutes: 720
|
|
secrets: inherit
|