Fix xformers TORCH_STABLE_ONLY issue + ramp up MAX_JOBS for native GH200

- Switch to official facebook/xformers (johnnynunez fork has TORCH_STABLE_ONLY requiring PyTorch headers not in 2.9.0)
- Increase MAX_JOBS from 2-4 to 8 for all builds (native GH200 has 97GB HBM3)
- Increase NVCC_THREADS from 1 to 4 for flash-attention

vllm/Dockerfile

@@ -73,13 +73,11 @@ FROM build-base AS build-xformers
 #ARG XFORMERS_REF=v0.0.32.post2
 #ARG XFORMERS_BUILD_VERSION=0.0.30+cu130
 #ENV BUILD_VERSION=${XFORMERS_BUILD_VERSION:-${XFORMERS_REF#v}}
-RUN git clone https://github.com/johnnynunez/xformers.git
-# https://github.com/facebookresearch/xformers.git
+RUN git clone https://github.com/facebookresearch/xformers.git
 RUN cd xformers && \
-#    git checkout ${XFORMERS_REF} && \
     git submodule sync && \
     git submodule update --init --recursive -j 8 && \
-    MAX_JOBS=3 uv build --wheel --no-build-isolation -o /wheels
+    MAX_JOBS=8 uv build --wheel --no-build-isolation -o /wheels
 
 FROM build-base AS build-flashinfer
 ARG FLASHINFER_ENABLE_AOT=1
@@ -101,7 +99,7 @@ RUN git clone https://github.com/LMCache/LMCache.git && \
 # PyTorch version is dated in LMCache
     sed -i '/torch/d' pyproject.toml && \
     uv pip install setuptools_scm && \
-    MAX_JOBS=2 python -m build --wheel --no-isolation && \
+    MAX_JOBS=8 python -m build --wheel --no-isolation && \
     cp dist/*.whl /wheels/
 
 
@@ -110,8 +108,8 @@ RUN apt-get update && apt-get install -y build-essential cmake gcc && \
     git clone https://github.com/Dao-AILab/flash-attention flash-attention && \
     cd flash-attention/hopper && \
     mkdir wheels && \
-    export MAX_JOBS=4 && \
-    export NVCC_THREADS=1 && \
+    export MAX_JOBS=8 && \
+    export NVCC_THREADS=4 && \
     export CMAKE_BUILD_PARALLEL_LEVEL=$MAX_JOBS && \
     MAX_JOBS=$MAX_JOBS \
     CMAKE_BUILD_PARALLEL_LEVEL=$MAX_JOBS \
@@ -137,7 +135,7 @@ RUN cd vllm && \
     git submodule sync && \
     git submodule update --init --recursive -j 8 && \
     sed -i 's/GIT_TAG [a-f0-9]\{40\}/GIT_TAG main/' cmake/external_projects/vllm_flash_attn.cmake && \
-    export MAX_JOBS=4 && \
+    export MAX_JOBS=8 && \
     export CMAKE_BUILD_PARALLEL_LEVEL=$MAX_JOBS && \
     python use_existing_torch.py && \
     uv pip install -r requirements/build.txt && \