[CI][Perf] migrate diffusion DFX benches to vllm bench serve --omni (#7737)

Signed-off-by: wangyu <410167048@qq.com>
This commit is contained in:
wangyu authored and GitHub committed 2026-09-21 19:16:39 -07:00
1 parent a4ad73e08b
commit 0570b0da7e
30 files changed
+2040 -729

No files matched your search

+2 -1
View File
@@ -159,7 +159,8 @@ When `mark` is present, it must be an **array** with exactly one ``hardware_mark
}
```
- Local bulk load: `pytest -sv tests/dfx/perf/scripts/run_diffusion_benchmark.py -m "full_model and diffusion and H100"`
- Local bulk load: `pytest -sv tests/dfx/perf/scripts/run_benchmark.py -m "full_model and H100"` (omni/TTS and `/v1/images/*` + `/v1/videos` diffusion)
- Diffusion chat-completions remaining cases: `pytest -sv tests/dfx/perf/scripts/run_diffusion_benchmark.py -m "full_model and diffusion and H100"`
- Nightly CI perf steps: `--test-config-file tests/dfx/perf/tests/test_<model>_vllm_omni.json` (file selects cases; no `-m`)
- Result filenames use **runtime** GPU detection (`get_runtime_resource_label`); `H100` is omitted on the default CI pool