Replace runs-on.com with ec2-github-runner for GPU CI#7
Closed
jack-champagne wants to merge 16 commits into
Closed
Replace runs-on.com with ec2-github-runner for GPU CI#7jack-champagne wants to merge 16 commits into
jack-champagne wants to merge 16 commits into
Conversation
…line Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add Aqua and JET to [extras]/[targets], create test/aqua.jl with all checks enabled (ambiguities disabled for CUDA.jl noise), add compat entries for all extras, and suppress known stale-dep false positives. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…m 0 to 1 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace anonymous NamedTuple refs with a concrete mutable CallbackRef type that registers a finalizer to automatically unregister both forward and gradient callbacks from the global registry when GC'd or explicitly closed. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…/catch finalizer safety Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace Any-typed GC anchor fields in ElementaryOperator, MatrixOperator, OperatorTerm, Operator, and WorkStream with concrete typed fields. Move callbacks.jl include before operators.jl so CallbackRef is in scope. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…meterize dtype/batch_size tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ase 6 gap Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…sion Added docstring for CUDENSITYMATError and all destroy_* functions that lacked them. Added export statements for the utility functions in state.jl, the batch/append/prepare/compute functions in operators.jl, and the prepare/compute functions in expectation.jl and spectrum.jl. Removed warnonly = [:missing_docs] from docs/make.jl now that all exported symbols are documented. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
… sync explicitly Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add .runs-on.yml defining a 'gpu' runner profile (g4dn.xlarge, T4, ubuntu22-gpu-x64, spot, 45 min timeout) and update the gpu-test CI job to use it via the runs-on label syntax. Add JULIA_CUDA_MEMORY_POOL=none and CUDA_VISIBLE_DEVICES='0' env vars. Job remains gated to main/tag pushes. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Delete .runs-on.yml (no longer using runs-on.com service) - Rewrite gpu-test as three-job pattern: start-gpu-runner, gpu-test, stop-gpu-runner - GPU tests now run on PRs (non-fork) and main pushes, not just main - Add gpu-benchmark.yml for on-demand benchmarks (T4 through H100) - Fork PR protection via repo name check on start-gpu-runner
Member
Author
|
Superseded by direct merge from clean branch off main |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
.runs-on.yml(no longer using runs-on.com service)gpu-benchmark.ymlfor on-demand benchmarks (T4, A10G, A100, H100)Dependencies
GH_RUNNER_PAT,AWS_GPU_RUNNER_ROLE_ARN,AWS_GPU_RUNNER_SUBNET_ID,AWS_GPU_RUNNER_SG_ID,AWS_GPU_RUNNER_AMI_ID,AWS_GPU_RUNNER_INSTANCE_PROFILETest plan
reposcopeif: always()cleanup on failure