Skip to content

Commit

Permalink
Merge branch 'jgfouca/final_scream_downstream_2024_11_21' into next (PR
Browse files Browse the repository at this point in the history
#6761)

The scream/eamxx fork has been archived (frozen). All new scream/eamxx development will happen in the E3SM repo.

[BFB]
  • Loading branch information
rljacob committed Nov 22, 2024
2 parents e8bc496 + 0d3b7da commit 2513412
Show file tree
Hide file tree
Showing 123 changed files with 3,961 additions and 3,596 deletions.
39 changes: 37 additions & 2 deletions .github/workflows/eamxx-sa-coverage.yml
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,11 @@ name: eamxx-sa-coverage

on:
workflow_dispatch:
inputs:
submit:
description: 'Force cdash submission'
required: true
type: boolean

# Add schedule trigger for nightly runs at midnight MT (Standard Time)
schedule:
Expand All @@ -13,7 +18,8 @@ concurrency:
cancel-in-progress: true

env:
submit: ${{ github.event_name == 'schedule' && 'true' || 'false' }} # Submit to cdash only for nightlies
# Submit to cdash only for nightlies or if the user explicitly forced a submission via workflow dispatch
submit: ${{ github.event_name == 'schedule' || (github.event_name == 'workflow_dispatch' && inputs.submit) }}

jobs:
gcc-openmp:
Expand Down Expand Up @@ -48,11 +54,40 @@ jobs:
submodules: recursive
- name: Show action trigger
uses: ./.github/actions/show-workflow-trigger
- name: Get CUDA Arch
run: |
# Ensure nvidia-smi is available
if ! command -v nvidia-smi &> /dev/null; then
echo "nvidia-smi could not be found. Please ensure you have Nvidia drivers installed."
exit 1
fi
# Get the GPU model from nvidia-smi, and set env for next step
gpu_model=$(nvidia-smi --query-gpu=name --format=csv,noheader | head -n 1)
case "$gpu_model" in
*"H100"*)
echo "Hopper=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=90" >> $GITHUB_ENV
ARCH=90
;;
*"A100"*)
echo "Ampere=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=80" >> $GITHUB_ENV
;;
*"V100"*)
echo "Volta=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=70" >> $GITHUB_ENV
;;
*)
echo "Unsupported GPU model: $gpu_model"
exit 1
;;
esac
- name: Run tests
uses: ./.github/actions/test-all-scream
with:
build_type: cov
machine: ghci-snl-cuda
generate: false
submit: ${{ env.submit }}
cmake-configs: Kokkos_ARCH_VOLTA70=ON;CMAKE_CUDA_ARCHITECTURES=70
cmake-configs: Kokkos_ARCH_HOPPER90=${{ env.Hopper }};Kokkos_ARCH_AMPERE80=${{ env.Ampere }};Kokkos_ARCH_VOLTA70=${{ env.Volta }};CMAKE_CUDA_ARCHITECTURES=${{ env.CUDA_ARCH }}
41 changes: 38 additions & 3 deletions .github/workflows/eamxx-sa-sanitizer.yml
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,11 @@ name: eamxx-sa-sanitizer

on:
workflow_dispatch:
inputs:
submit:
description: 'Force cdash submission'
required: true
type: boolean

# Add schedule trigger for nightly runs at midnight MT (Standard Time)
schedule:
Expand All @@ -13,12 +18,13 @@ concurrency:
cancel-in-progress: true

env:
submit: ${{ github.event_name == 'schedule' && 'true' || 'false' }} # Submit to cdash only for nightlies
# Submit to cdash only for nightlies or if the user explicitly forced a submission via workflow dispatch
submit: ${{ github.event_name == 'schedule' || (github.event_name == 'workflow_dispatch' && inputs.submit) }}

jobs:
gcc-openmp:
runs-on: [self-hosted, ghci-snl-cpu, gcc]
name: gcc-openmp / cov
name: gcc-openmp / valg
steps:
- name: Check out the repository
uses: actions/checkout@v4
Expand Down Expand Up @@ -52,11 +58,40 @@ jobs:
submodules: recursive
- name: Show action trigger
uses: ./.github/actions/show-workflow-trigger
- name: Get CUDA Arch
run: |
# Ensure nvidia-smi is available
if ! command -v nvidia-smi &> /dev/null; then
echo "nvidia-smi could not be found. Please ensure you have Nvidia drivers installed."
exit 1
fi
# Get the GPU model from nvidia-smi, and set env for next step
gpu_model=$(nvidia-smi --query-gpu=name --format=csv,noheader | head -n 1)
case "$gpu_model" in
*"H100"*)
echo "Hopper=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=90" >> $GITHUB_ENV
ARCH=90
;;
*"A100"*)
echo "Ampere=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=80" >> $GITHUB_ENV
;;
*"V100"*)
echo "Volta=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=70" >> $GITHUB_ENV
;;
*)
echo "Unsupported GPU model: $gpu_model"
exit 1
;;
esac
- name: Run tests
uses: ./.github/actions/test-all-scream
with:
build_type: ${{ matrix.build_type }}
machine: ghci-snl-cuda
generate: false
submit: ${{ env.submit }}
cmake-configs: Kokkos_ARCH_VOLTA70=ON;CMAKE_CUDA_ARCHITECTURES=70
cmake-configs: Kokkos_ARCH_HOPPER90=${{ env.Hopper }};Kokkos_ARCH_AMPERE80=${{ env.Ampere }};Kokkos_ARCH_VOLTA70=${{ env.Volta }};CMAKE_CUDA_ARCHITECTURES=${{ env.CUDA_ARCH }}
145 changes: 60 additions & 85 deletions .github/workflows/eamxx-sa-testing.yml
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,17 @@ on:
pull_request:
branches: [ master ]
types: [opened, synchronize, ready_for_review, reopened]
paths:
# first, yes to these
- '.github/workflows/eamxx-sa-testing.yml'
- 'cime_config/machine/config_machines.xml'
- 'components/eamxx/**'
- 'components/homme/**'
- 'externals/ekat'
- 'externals/scorpio'
# second, no to these
- '!components/eamxx/docs/**'
- '!components/eamxx/mkdocs.yml'

# Manual run is used to bless
workflow_dispatch:
Expand All @@ -21,6 +32,10 @@ on:
description: 'Generate baselines'
required: true
type: boolean
submit:
description: 'Force cdash submission'
required: true
type: boolean

# Add schedule trigger for nightly runs at midnight MT (Standard Time)
schedule:
Expand All @@ -34,66 +49,20 @@ concurrency:
cancel-in-progress: true

env:
submit: ${{ github.event_name == 'schedule' && 'true' || 'false' }} # Submit to cdash only for nightlies
# Submit to cdash only for nightlies or if the user explicitly forced a submission via workflow dispatch
submit: ${{ github.event_name == 'schedule' || (github.event_name == 'workflow_dispatch' && inputs.submit) }}
generate: ${{ github.event_name == 'workflow_dispatch' && inputs.bless }}

jobs:
pre_process_pr:
if: ${{ github.event_name == 'pull_request' }}
runs-on: ubuntu-latest # This job can run anywhere
outputs:
relevant_paths: ${{ steps.check_paths.outputs.value }}
labels: ${{ steps.get_labels.outputs.labels }}
steps:
- id: check_paths
run: |
paths=(
components/eamxx
components/eam/src/physics/rrtmgp
components/eam/src/physics/p3/scream
components/eam/src/physics/cam
components/eam/src/physics/rrtmgp/external
externals/ekat
externals/scorpio
externals/haero
externals/YAKL
.github/workflows/eamxx-sa-testing.yml
)
pattern=$(IFS=\|; echo "${paths[*]}")
# Use the GitHub API to get the list of changed files
response=$(curl -s -H "Authorization: token ${{ secrets.GITHUB_TOKEN }}" \
"https://api.github.com/repos/${{ github.repository }}/pulls/${{ github.event.number }}/files")
changed_files=$(echo "$response" | grep -o '"filename": *"[^"]*"' | sed 's/"filename": *//; s/"//g')
# Check for matches and echo the matching files (or "" if none)
matching_files=$(echo "$changed_files" | grep -E "^($pattern)" || echo "")
if [[ -n "$matching_files" ]]; then
echo "Found relevant files: $matching_files"
echo "value=true" >> $GITHUB_OUTPUT
else
echo "No relevant files touched by this PR."
echo "value=false" >> $GITHUB_OUTPUT
fi
- id: get_labels
run: |
labels="${{ join(github.event.pull_request.labels.*.name, ',') }}"
echo "labels=${labels}" >> $GITHUB_OUTPUT
gcc-openmp:
needs: [pre_process_pr]
if: |
github.event_name == 'schedule' ||
(
github.event_name == 'pull_request' &&
needs.pre_process_pr.outputs.relevant_paths=='true' &&
!contains(needs.pre_process_pr.outputs.labels,'CI: skip gcc') &&
!contains(needs.pre_process_pr.outputs.labels,'CI: skip openmp') &&
!contains(needs.pre_process_pr.outputs.labels,'CI: skip eamxx-sa') &&
!contains(needs.pre_process_pr.outputs.labels,'CI: skip eamxx-all')
) || (
github.event_name == 'workflow_dispatch' &&
github.event.inputs.job_to_run == 'gcc-openmp' ||
github.event.inputs.job_to_run == 'all'
)
${{
github.event_name != 'workflow_dispatch' ||
(
github.event.inputs.job_to_run == 'gcc-openmp' ||
github.event.inputs.job_to_run == 'all'
)
}}
runs-on: [self-hosted, ghci-snl-cpu, gcc]
strategy:
fail-fast: false
Expand All @@ -109,14 +78,6 @@ jobs:
submodules: recursive
- name: Show action trigger
uses: ./.github/actions/show-workflow-trigger
- name: Set test-all inputs based on event specs
run: |
echo "generate=false" >> $GITHUB_ENV
if [ "${{ github.event_name }}" == "workflow_dispatch" ]; then
if [ "${{ inputs.bless }}" == "true" ]; then
echo "generate=true" >> $GITHUB_ENV
fi
fi
- name: Run tests
uses: ./.github/actions/test-all-scream
with:
Expand All @@ -126,21 +87,14 @@ jobs:
submit: ${{ env.submit }}
cmake-configs: Kokkos_ENABLE_OPENMP=ON
gcc-cuda:
needs: [pre_process_pr]
if: |
github.event_name == 'schedule' ||
(
github.event_name == 'pull_request' &&
needs.pre_process_pr.outputs.relevant_paths=='true' &&
!contains(needs.pre_process_pr.outputs.labels,'CI: skip gcc') &&
!contains(needs.pre_process_pr.outputs.labels,'CI: skip cuda') &&
!contains(needs.pre_process_pr.outputs.labels,'CI: skip eamxx-sa') &&
!contains(needs.pre_process_pr.outputs.labels,'CI: skip eamxx-all')
) || (
github.event_name == 'workflow_dispatch' &&
github.event.inputs.job_to_run == 'gcc-cuda' ||
github.event.inputs.job_to_run == 'all'
)
${{
github.event_name != 'workflow_dispatch' ||
(
github.event.inputs.job_to_run == 'gcc-cuda' ||
github.event.inputs.job_to_run == 'all'
)
}}
runs-on: [self-hosted, ghci-snl-cuda, cuda, gcc]
strategy:
fail-fast: false
Expand All @@ -156,19 +110,40 @@ jobs:
submodules: recursive
- name: Show action trigger
uses: ./.github/actions/show-workflow-trigger
- name: Set test-all inputs based on event specs
- name: Get CUDA Arch
run: |
echo "generate=false" >> $GITHUB_ENV
if [ "${{ github.event_name }}" == "workflow_dispatch" ]; then
if [ "${{ inputs.bless }}" == "true" ]; then
echo "generate=true" >> $GITHUB_ENV
fi
# Ensure nvidia-smi is available
if ! command -v nvidia-smi &> /dev/null; then
echo "nvidia-smi could not be found. Please ensure you have Nvidia drivers installed."
exit 1
fi
# Get the GPU model from nvidia-smi, and set env for next step
gpu_model=$(nvidia-smi --query-gpu=name --format=csv,noheader | head -n 1)
case "$gpu_model" in
*"H100"*)
echo "Hopper=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=90" >> $GITHUB_ENV
ARCH=90
;;
*"A100"*)
echo "Ampere=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=80" >> $GITHUB_ENV
;;
*"V100"*)
echo "Volta=ON" >> $GITHUB_ENV
echo "CUDA_ARCH=70" >> $GITHUB_ENV
;;
*)
echo "Unsupported GPU model: $gpu_model"
exit 1
;;
esac
- name: Run tests
uses: ./.github/actions/test-all-scream
with:
build_type: ${{ matrix.build_type }}
machine: ghci-snl-cuda
generate: ${{ env.generate }}
submit: ${{ env.submit }}
cmake-configs: Kokkos_ARCH_VOLTA70=ON;CMAKE_CUDA_ARCHITECTURES=70
cmake-configs: Kokkos_ARCH_HOPPER90=${{ env.Hopper }};Kokkos_ARCH_AMPERE80=${{ env.Ampere }};Kokkos_ARCH_VOLTA70=${{ env.Volta }};CMAKE_CUDA_ARCHITECTURES=${{ env.CUDA_ARCH }}
Loading

0 comments on commit 2513412

Please sign in to comment.