Skip to content

fix(scripts): default remote GPU helpers to index 5 (#33)#137

Closed
RealDiligent wants to merge 1 commit into
mini-router:mainfrom
RealDiligent:fix/critical-issue-gpu5-default-scripts
Closed

fix(scripts): default remote GPU helpers to index 5 (#33)#137
RealDiligent wants to merge 1 commit into
mini-router:mainfrom
RealDiligent:fix/critical-issue-gpu5-default-scripts

Conversation

@RealDiligent

Copy link
Copy Markdown
Contributor

Summary

Fixes #33.

Remote helper scripts defaulted TRINITY_GPU_INDEX to 0 when unset, while validator config and secrets.env.example standardize on GPU 5. Jobs could land on the wrong GPU on shared boxes.

This PR changes the shell fallbacks to :-5 in setup_remote.sh, run_remote.sh, and remote_env.sh, and adds a regression test.

Test plan

  • test_setup_remote_defaults_to_gpu_five_when_unset added
  • CI test-router passes

@tmimmanuel

Copy link
Copy Markdown
Collaborator

MiniRouter evaluation result

Submission for RealDiligent on math500 failed during evaluation.

field value
submission d60d7516-ed0e-480e-8b15-a2b9be8ba9db
PR #137
repo mini-router/minirouter
benchmark math500
status failed
phase failed
progress 0/1
score pending
started 2026-07-13T02:45:19.105641+00:00
finished 2026-07-13T02:45:19.111781+00:00

Metrics

metric value
duration_seconds 0.01s
cost_usd $0.0000
cost_missing True
cost_ledger /home/ubuntu/codes/tmimmanuel/tiny-router-works/minirouter/data/workspaces/submissions/d60d7516-ed0e-480e-8b15-a2b9be8ba9db/cost_ledger.jsonl
missing_checkpoint True

Error

submission d60d7516-ed0e-480e-8b15-a2b9be8ba9db does not have a checkpoint to evaluate

@github-actions github-actions Bot added docs Documentation changes eval Evaluation changes train Training or routing changes labels Jul 13, 2026
@tmimmanuel tmimmanuel closed this Jul 13, 2026
@tmimmanuel

Copy link
Copy Markdown
Collaborator

MiniRouter evaluation result

Submission for tmimmanuel on math500 failed during evaluation.

field value
submission d60d7516-ed0e-480e-8b15-a2b9be8ba9db
PR #137
repo mini-router/minirouter
benchmark math500
status failed
phase failed
progress 0/1
score pending
started 2026-07-13T02:45:36.302559+00:00
finished 2026-07-13T02:45:36.309996+00:00

Metrics

metric value
duration_seconds 0.01s
cost_usd $0.0000
cost_missing True
cost_ledger /home/ubuntu/codes/tmimmanuel/tiny-router-works/minirouter/data/workspaces/submissions/d60d7516-ed0e-480e-8b15-a2b9be8ba9db/cost_ledger.jsonl
missing_checkpoint True

Error

submission d60d7516-ed0e-480e-8b15-a2b9be8ba9db does not have a checkpoint to evaluate

Align setup_remote, run_remote, and remote_env fallbacks with
DEFAULT_TRINITY_GPU_INDEX so unset TRINITY_GPU_INDEX no longer pins jobs to GPU 0.

Co-authored-by: Cursor <cursoragent@cursor.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

docs Documentation changes eval Evaluation changes train Training or routing changes

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Default remote helper scripts to GPU 5 instead of GPU 0

2 participants