add fp8 gemm path choice for rocm_aiter_gemm_w8a8_blockscale · ROCm/vllm@0755e0c

Triggered via pull request September 9, 2025 01:44

zhuyuhua-v

synchronize #659

yuhua/355_gemm_choic

Status Failure

Total duration 7m 1s

Artifacts –

pre-commit.yml

on: pull_request

10 errors

pre-commit: ROCm_performance.md#L1

Ruff (E501): vllm/v1/attention/backends/triton_attn.py#L346

vllm/v1/attention/backends/triton_attn.py:346:81: E501 Line too long (89 > 80)

Ruff (E501): vllm/v1/attention/backends/triton_attn.py#L338

vllm/v1/attention/backends/triton_attn.py:338:81: E501 Line too long (103 > 80)

Ruff (E501): vllm/v1/attention/backends/triton_attn.py#L31

vllm/v1/attention/backends/triton_attn.py:31:81: E501 Line too long (83 > 80)

Ruff (E501): vllm/v1/attention/backends/triton_attn.py#L29

vllm/v1/attention/backends/triton_attn.py:29:81: E501 Line too long (145 > 80)

Ruff (E501): vllm/model_executor/models/llama.py#L61

vllm/model_executor/models/llama.py:61:81: E501 Line too long (145 > 80)

Ruff (E501): vllm/envs.py#L1189

vllm/envs.py:1189:81: E501 Line too long (94 > 80)

Ruff (E501): vllm/attention/layer.py#L522

vllm/attention/layer.py:522:81: E501 Line too long (99 > 80)

Ruff (E501): vllm/attention/layer.py#L521

vllm/attention/layer.py:521:81: E501 Line too long (147 > 80)

Ruff (E501): vllm/attention/layer.py#L32

vllm/attention/layer.py:32:81: E501 Line too long (145 > 80)