Repository navigation
GPU runners: no sm_89 label; a fallback card may be an A10G - #15
Merged
Merged
Conversation
The ladder can now retype a GPU box to a g5 (A10G, sm_86), and the CUDA code under test builds sm_86, which runs on both. Callers ask for `gpu` alone, and the preflight warning no longer claims a fallback card has different memory. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The GPU runners' capacity ladder in aws-dev-infra is gaining g5 (A10G, sm_86) shapes. dps and snarkVM-cuda gpu-test already build
CUDA_ARCH=sm_86(upstream's arch, which runs on the L4 too) and already ask for[self-hosted, linux, x64, gpu], so thesm_89label goes: a box retyped to g5 would carry it falsely.start-runner.ymlandREADME.mdexamples:runs-on: [self-hosted, linux, x64, gpu].gpu-runner-preflight: the comment names the A10G as a possible fallback, and the off-card warning no longer says "different GPU memory" (an A10G has the same as an L4). New test: an A10G warns.Needs
v2moved once merged. aws-dev-infra's test reads these docs, so this merges first.🤖 Generated with Claude Code