gitadmin/foxhunt
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(ci): update CUDA_COMPUTE_CAP to 90 for H100 gpu-training pool

Browse Source 
The CI runner now runs on H100 (compute cap 9.0), not L4 (8.9).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
jgrusewski
2026-02-26 11:55:55 +01:00
parent 365470fc3b
commit 5853a704ce
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
.gitlab-ci.yml
View File

@@ -14,8 +14,8 @@ stages:
variables:
SQLX_OFFLINE: "true"
CARGO_TERM_COLOR: always
# L4 = compute capability 8.9 — tells bindgen_cuda to skip nvidia-smi detection
CUDA_COMPUTE_CAP: "89"
# H100 = compute capability 9.0 — tells bindgen_cuda to skip nvidia-smi detection
CUDA_COMPUTE_CAP: "90"
# Stable JWT secret for api_gateway tests (prevents env-var race between parallel tests)
JWT_SECRET: "CiTestSecret_Kx7mP9nR2sW5vY8bC3fG6jH1kL4pQ7tZ0uN9dM5eV8xS2wT6yA4zB_64chars_min"
# Training data on block storage PVC (mounted at /mnt/training-data by runner)