Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 8 additions & 8 deletions .cirun.yml
Original file line number Diff line number Diff line change
@@ -1,23 +1,23 @@
runners:
# Primary: A10G. ~3-4x faster than the T4 (Ampere, tensor cores, ~600 GB/s).
# A10G for dev + prerelease (cu13). scverse-standard AMI, eu-west-1.
- name: aws-gpu-runner-g5
cloud: aws
instance_type: g5.xlarge
machine_image: ami-067a4ba2816407ee9
region: eu-north-1
machine_image: ami-0756bd122da9b2618
region: eu-west-1
preemptible:
- true
- false
labels:
- cirun-aws-gpu
# Fallback: T4. Cheapest; cirun picks this when g5 spot is dry.
- cirun-aws-gpu-a10
# T4 for stable (cu13). scverse-standard AMI, eu-west-1.
- name: aws-gpu-runner-g4dn
cloud: aws
instance_type: g4dn.xlarge
machine_image: ami-067a4ba2816407ee9
region: eu-north-1
machine_image: ami-0756bd122da9b2618
region: eu-west-1
preemptible:
- true
- false
labels:
- cirun-aws-gpu
- cirun-aws-gpu-t4
18 changes: 12 additions & 6 deletions .github/workflows/test-gpu.yml
Original file line number Diff line number Diff line change
Expand Up @@ -39,34 +39,40 @@ jobs:
enable-cache: false
- id: get-envs
run: |
# Stable and dev environments
# Stable + dev, all cu13 (stable -> T4, dev -> A10)
STABLE_DEV_JSON=$(uvx --from "hatch==1.16.5" hatch env show --json | jq -c '
to_entries
| map(
select(.key | startswith("hatch-test") and endswith("12") and (contains("prerelease") | not))
select(.key | startswith("hatch-test"))
| select(
((.key | contains("stable")) and (.key | endswith("13")))
or ((.key | contains("dev")) and (.key | endswith("13")))
)
| {
name: .key,
python: .value.python | sub("3[.]13"; "3.13.3"),
runner: (if (.key | contains("stable")) then "cirun-aws-gpu-t4" else "cirun-aws-gpu-a10" end),
}
)')
echo "stable-dev=${STABLE_DEV_JSON}" | tee -a $GITHUB_OUTPUT

# Prerelease environments
# Prerelease environments (cu13 -> A10)
PRERELEASE_JSON=$(uvx --from "hatch==1.16.5" hatch env show --json | jq -c '
to_entries
| map(
select(.key | startswith("hatch-test") and endswith("12") and contains("prerelease"))
select(.key | startswith("hatch-test") and endswith("13") and contains("prerelease"))
| {
name: .key,
python: .value.python | sub("3[.]13"; "3.13.3"),
runner: "cirun-aws-gpu-a10",
}
)')
echo "prerelease=${PRERELEASE_JSON}" | tee -a $GITHUB_OUTPUT

test-stable-dev:
name: Test (${{ matrix.env.name }})
needs: get-envs
runs-on: cirun-aws-gpu
runs-on: ${{ matrix.env.runner }}
timeout-minutes: 30
strategy:
fail-fast: false
Expand Down Expand Up @@ -128,7 +134,7 @@ jobs:
test-prerelease:
name: Test (${{ matrix.env.name }})
needs: get-envs
runs-on: cirun-aws-gpu
runs-on: ${{ matrix.env.runner }}
timeout-minutes: 30
continue-on-error: true # Prerelease failures don't fail the workflow
strategy:
Expand Down
Loading