Skip to content
Cuda - Fast tests on H100

chore(ci): use new heuristic for throughput benchmarks #1748

Sign in to view logs

chore(ci): use new heuristic for throughput benchmarks

chore(ci): use new heuristic for throughput benchmarks #1748

Triggered via pull request March 6, 2025 10:28
@zama-botzama-bot
labeled #2112
dt/bench/throughput_heuristic
Status Success
Total duration 32m 0s
Artifacts –

gpu_fast_h100_tests.yml

on: pull_request
should-run
6s
should-run
Setup instance (cuda-h100-tests)
48s
Setup instance (cuda-h100-tests)
Matrix: CUDA H100 tests
Slack Notification
0s
Slack Notification
Teardown instance (cuda-h100-tests)
0s
Teardown instance (cuda-h100-tests)
Fit to window
Zoom out
Zoom in

Annotations

4 errors
Setup instance (cuda-h100-tests)
Instance task failed (details: api call failed (reason: bad_request, message: Not Enough Stock of H100-80G-PCIe. Unable to launch virtual-machines.))
Setup instance (cuda-h100-tests)
Failure occurred while waiting for instance.
Setup instance (cuda-h100-tests)
Instance stop request has failed (HTTP status code: 500, body: Failed to stop backend instance (reason: backend provider not found with instance key: InstanceHashMapKey("159917a3769c6f527a2d6ed2227510449891fc43", "faca6f4b-fa95-447e-a359-2a592d12d97b")))
Setup instance (cuda-h100-tests)
instance stop request failed