xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-09 04:24:58 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update Flashinfer from v0.4.1 to v0.5.2 (#27952)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor 2025-11-07 16:24:42 -08:00 committed by GitHub
parent 67a2da890e
commit 811df41ee9
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 11 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
docker/Dockerfile
View File

@ -132,9 +132,7 @@ WORKDIR /workspace
COPY requirements/common.txt requirements/common.txt
COPY requirements/cuda.txt requirements/cuda.txt
RUN --mount=type=cache,target=/root/.cache/uv \
# TODO: remove apache-tvm-ffi once FlashInfer is fixed https://github.com/flashinfer-ai/flashinfer/issues/1962
uv pip install --python /opt/venv/bin/python3 --pre apache-tvm-ffi==0.1.0b15 \
&& uv pip install --python /opt/venv/bin/python3 -r requirements/cuda.txt \
uv pip install --python /opt/venv/bin/python3 -r requirements/cuda.txt \
--extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
# cuda arch list used by torch
@ -356,16 +354,14 @@ RUN --mount=type=cache,target=/root/.cache/uv \
# Install vllm wheel first, so that torch etc will be installed.
RUN --mount=type=bind,from=build,src=/workspace/dist,target=/vllm-workspace/dist \
--mount=type=cache,target=/root/.cache/uv \
# TODO: remove apache-tvm-ffi once FlashInfer is fixed https://github.com/flashinfer-ai/flashinfer/issues/1962
uv pip install --system --pre apache-tvm-ffi==0.1.0b15 \
&& uv pip install --system dist/*.whl --verbose \
uv pip install --system dist/*.whl --verbose \
--extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
# Install FlashInfer pre-compiled kernel cache and binaries
# https://docs.flashinfer.ai/installation.html
RUN --mount=type=cache,target=/root/.cache/uv \
uv pip install --system flashinfer-cubin==0.4.1 \
&& uv pip install --system flashinfer-jit-cache==0.4.1 \
uv pip install --system flashinfer-cubin==0.5.2 \
&& uv pip install --system flashinfer-jit-cache==0.5.2 \
--extra-index-url https://flashinfer.ai/whl/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.') \
&& flashinfer show-config

4
docker/Dockerfile.nightly_torch
View File

@ -246,7 +246,7 @@ RUN pip install setuptools==75.6.0 packaging==23.2 ninja==1.11.1.3 build==1.2.2.
# build flashinfer for torch nightly from source around 10 mins
# release version: v0.4.1
# release version: v0.5.2
# todo(elainewy): cache flashinfer build result for faster build
ENV CCACHE_DIR=/root/.cache/ccache
RUN --mount=type=cache,target=/root/.cache/ccache \
@ -254,7 +254,7 @@ RUN --mount=type=cache,target=/root/.cache/ccache \
echo "git clone flashinfer..." \
&& git clone --recursive https://github.com/flashinfer-ai/flashinfer.git \
&& cd flashinfer \
&& git checkout v0.4.1\
&& git checkout v0.5.2 \
&& git submodule update --init --recursive \
&& echo "finish git clone flashinfer..." \
&& rm -rf build \

2
requirements/cuda.txt
View File

@ -12,4 +12,4 @@ torchvision==0.24.0 # Required for phi3v processor. See https://github.com/pytor
# Build from https://github.com/facebookresearch/xformers/releases/tag/v0.0.32.post1
xformers==0.0.33+5d4b92a5.d20251029; platform_system == 'Linux' and platform_machine == 'x86_64' # Requires PyTorch >= 2.9
# FlashInfer should be updated together with the Dockerfile
flashinfer-python==0.4.1
flashinfer-python==0.5.2

6
tests/kernels/attention/test_flashinfer_trtllm_attention.py
View File

@ -238,9 +238,11 @@ def test_flashinfer_trtllm_decode_with_baseline(
if q_quant_dtype == FP8_DTYPE and o_quant_dtype == FP4_DTYPE:
rtol, atol = 7e-2, 9e-2
elif q_quant_dtype == FP8_DTYPE and o_quant_dtype == FP8_DTYPE:
rtol, atol = 2e-2, 4e-2
rtol, atol = 3e-2, 4e-2
elif q_quant_dtype == FP8_DTYPE and o_quant_dtype == dtype:
rtol, atol = 1e-2, 2e-2
rtol, atol = 2e-2, 2e-2
elif kv_quant_dtype == FP8_DTYPE:
rtol, atol = 4e-2, 6e-2
else:
rtol, atol = 1e-2, 1e-2