Reduce MAX_JOBS by half to reduce QEMU memory pressure

- xformers: 6 -> 3
- flash-attention: 8 -> 4
- vllm: 8 -> 4

Testing if lower parallelism helps avoid segfaults under emulation

vllm/Dockerfile

@@ -85,7 +85,7 @@ RUN cd xformers && \
 #    git checkout ${XFORMERS_REF} && \
     git submodule sync && \
     git submodule update --init --recursive -j 8 && \
-    MAX_JOBS=6 uv build --wheel --no-build-isolation -o /wheels
+    MAX_JOBS=3 uv build --wheel --no-build-isolation -o /wheels
 
 FROM build-base AS build-flashinfer
 ARG FLASHINFER_ENABLE_AOT=1
@@ -115,7 +115,7 @@ 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=8 && \
+    export MAX_JOBS=4 && \
     export NVCC_THREADS=1 && \
     export CMAKE_BUILD_PARALLEL_LEVEL=$MAX_JOBS && \
     MAX_JOBS=$MAX_JOBS \
@@ -136,7 +136,7 @@ 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=8 && \
+    export MAX_JOBS=4 && \
     export CMAKE_BUILD_PARALLEL_LEVEL=$MAX_JOBS && \
     python use_existing_torch.py && \
     uv pip install -r requirements/build.txt && \