[ROCm] Remove stale SDPA and skinny GEMM workarounds (#50907)

Signed-off-by: Andreas Karatzas <Andreas.Karatzas@amd.com>
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
This commit is contained in:
Andreas Karatzas
2026-08-11 14:19:16 -04:00
committed by GitHub
co-authored by OpenAI Codex
parent 4f2f31b82e
commit 1ab2801dde
56 changed files with 500 additions and 558 deletions
@@ -9,8 +9,6 @@ processor in the pytest parent process to keep os.fork() in sibling
tests (e.g. test_weight_transfer_llm.py) deadlock-free.
"""
import os
import httpx
import pytest
import pytest_asyncio
@@ -41,10 +39,7 @@ def server():
"--no-enable-prefix-caching",
]
envs = os.environ.copy()
envs["VLLM_ROCM_USE_SKINNY_GEMM"] = "0"
with RemoteOpenAIServer(MODEL_NAME, args, env_dict=envs) as remote_server:
with RemoteOpenAIServer(MODEL_NAME, args) as remote_server:
yield remote_server
@@ -2,7 +2,6 @@
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import json
import os
import httpx
import pytest
@@ -79,11 +78,7 @@ def server(request):
else [str(extra_args)]
)
envs = os.environ.copy()
# See: https://github.com/vllm-project/vllm/pull/33493#issuecomment-3888060787
envs["VLLM_ROCM_USE_SKINNY_GEMM"] = "0"
with RemoteOpenAIServer(MODEL_NAME, args, env_dict=envs) as remote_server:
with RemoteOpenAIServer(MODEL_NAME, args) as remote_server:
yield remote_server