Skip to content

Commit ebb7983

Browse files
XuehaiPanXuehaiPan
committed
Update
[ghstack-poisoned]
2 parents d81f4b6 + c157104 commit ebb7983

File tree

1,101 files changed

+47408
-22320
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

1,101 files changed

+47408
-22320
lines changed

‎.ci/aarch64_linux/build_aarch64_wheel.py

Lines changed: 4 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -438,9 +438,7 @@ def build_torchvision(
438438
)
439439
build_vars += f"BUILD_VERSION={version}.dev{build_date}"
440440
elif build_version is not None:
441-
build_vars += (
442-
f"BUILD_VERSION={build_version} PYTORCH_VERSION={branch[1:].split('-')[0]}"
443-
)
441+
build_vars += f"BUILD_VERSION={build_version} PYTORCH_VERSION={branch[1:].split('-', maxsplit=1)[0]}"
444442
if host.using_docker():
445443
build_vars += " CMAKE_SHARED_LINKER_FLAGS=-Wl,-z,max-page-size=0x10000"
446444

@@ -495,9 +493,7 @@ def build_torchdata(
495493
)
496494
build_vars += f"BUILD_VERSION={version}.dev{build_date}"
497495
elif build_version is not None:
498-
build_vars += (
499-
f"BUILD_VERSION={build_version} PYTORCH_VERSION={branch[1:].split('-')[0]}"
500-
)
496+
build_vars += f"BUILD_VERSION={build_version} PYTORCH_VERSION={branch[1:].split('-', maxsplit=1)[0]}"
501497
if host.using_docker():
502498
build_vars += " CMAKE_SHARED_LINKER_FLAGS=-Wl,-z,max-page-size=0x10000"
503499

@@ -553,9 +549,7 @@ def build_torchtext(
553549
)
554550
build_vars += f"BUILD_VERSION={version}.dev{build_date}"
555551
elif build_version is not None:
556-
build_vars += (
557-
f"BUILD_VERSION={build_version} PYTORCH_VERSION={branch[1:].split('-')[0]}"
558-
)
552+
build_vars += f"BUILD_VERSION={build_version} PYTORCH_VERSION={branch[1:].split('-', maxsplit=1)[0]}"
559553
if host.using_docker():
560554
build_vars += " CMAKE_SHARED_LINKER_FLAGS=-Wl,-z,max-page-size=0x10000"
561555

@@ -613,9 +607,7 @@ def build_torchaudio(
613607
)
614608
build_vars += f"BUILD_VERSION={version}.dev{build_date}"
615609
elif build_version is not None:
616-
build_vars += (
617-
f"BUILD_VERSION={build_version} PYTORCH_VERSION={branch[1:].split('-')[0]}"
618-
)
610+
build_vars += f"BUILD_VERSION={build_version} PYTORCH_VERSION={branch[1:].split('-', maxsplit=1)[0]}"
619611
if host.using_docker():
620612
build_vars += " CMAKE_SHARED_LINKER_FLAGS=-Wl,-z,max-page-size=0x10000"
621613

‎.ci/docker/README.md

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -104,7 +104,6 @@ If your new Docker image needs a library installed from a specific pinned commit
104104
```bash
105105
pytorch-linux-jammy-cuda12.8-cudnn9-py3.12-gcc11-new1)
106106
CUDA_VERSION=12.8.1
107-
CUDNN_VERSION=9
108107
ANACONDA_PYTHON_VERSION=3.12
109108
GCC_VERSION=11
110109
VISION=yes

‎.ci/docker/build.sh

Lines changed: 4 additions & 74 deletions
Original file line numberDiff line numberDiff line change
@@ -93,7 +93,6 @@ tag=$(echo $image | awk -F':' '{print $2}')
9393
case "$tag" in
9494
pytorch-linux-jammy-cuda12.4-cudnn9-py3-gcc11)
9595
CUDA_VERSION=12.4
96-
CUDNN_VERSION=9
9796
ANACONDA_PYTHON_VERSION=3.10
9897
GCC_VERSION=11
9998
VISION=yes
@@ -104,7 +103,6 @@ case "$tag" in
104103
;;
105104
pytorch-linux-jammy-cuda12.8-cudnn9-py3-gcc11)
106105
CUDA_VERSION=12.8.1
107-
CUDNN_VERSION=9
108106
ANACONDA_PYTHON_VERSION=3.10
109107
GCC_VERSION=11
110108
VISION=yes
@@ -115,7 +113,6 @@ case "$tag" in
115113
;;
116114
pytorch-linux-jammy-cuda12.8-cudnn9-py3-gcc9-inductor-benchmarks)
117115
CUDA_VERSION=12.8.1
118-
CUDNN_VERSION=9
119116
ANACONDA_PYTHON_VERSION=3.10
120117
GCC_VERSION=9
121118
VISION=yes
@@ -127,7 +124,6 @@ case "$tag" in
127124
;;
128125
pytorch-linux-jammy-cuda12.8-cudnn9-py3.12-gcc9-inductor-benchmarks)
129126
CUDA_VERSION=12.8.1
130-
CUDNN_VERSION=9
131127
ANACONDA_PYTHON_VERSION=3.12
132128
GCC_VERSION=9
133129
VISION=yes
@@ -139,7 +135,6 @@ case "$tag" in
139135
;;
140136
pytorch-linux-jammy-cuda12.8-cudnn9-py3.13-gcc9-inductor-benchmarks)
141137
CUDA_VERSION=12.8.1
142-
CUDNN_VERSION=9
143138
ANACONDA_PYTHON_VERSION=3.13
144139
GCC_VERSION=9
145140
VISION=yes
@@ -149,20 +144,8 @@ case "$tag" in
149144
TRITON=yes
150145
INDUCTOR_BENCHMARKS=yes
151146
;;
152-
pytorch-linux-jammy-cuda12.6-cudnn9-py3-gcc9)
153-
CUDA_VERSION=12.6.3
154-
CUDNN_VERSION=9
155-
ANACONDA_PYTHON_VERSION=3.10
156-
GCC_VERSION=9
157-
VISION=yes
158-
KATEX=yes
159-
UCX_COMMIT=${_UCX_COMMIT}
160-
UCC_COMMIT=${_UCC_COMMIT}
161-
TRITON=yes
162-
;;
163147
pytorch-linux-jammy-cuda12.8-cudnn9-py3.12-gcc11-vllm)
164148
CUDA_VERSION=12.8.1
165-
CUDNN_VERSION=9
166149
ANACONDA_PYTHON_VERSION=3.12
167150
GCC_VERSION=11
168151
VISION=yes
@@ -171,45 +154,8 @@ case "$tag" in
171154
UCC_COMMIT=${_UCC_COMMIT}
172155
TRITON=yes
173156
;;
174-
pytorch-linux-jammy-cuda12.6-cudnn9-py3-gcc9-inductor-benchmarks)
175-
CUDA_VERSION=12.6
176-
CUDNN_VERSION=9
177-
ANACONDA_PYTHON_VERSION=3.10
178-
GCC_VERSION=9
179-
VISION=yes
180-
KATEX=yes
181-
UCX_COMMIT=${_UCX_COMMIT}
182-
UCC_COMMIT=${_UCC_COMMIT}
183-
TRITON=yes
184-
INDUCTOR_BENCHMARKS=yes
185-
;;
186-
pytorch-linux-jammy-cuda12.6-cudnn9-py3.12-gcc9-inductor-benchmarks)
187-
CUDA_VERSION=12.6
188-
CUDNN_VERSION=9
189-
ANACONDA_PYTHON_VERSION=3.12
190-
GCC_VERSION=9
191-
VISION=yes
192-
KATEX=yes
193-
UCX_COMMIT=${_UCX_COMMIT}
194-
UCC_COMMIT=${_UCC_COMMIT}
195-
TRITON=yes
196-
INDUCTOR_BENCHMARKS=yes
197-
;;
198-
pytorch-linux-jammy-cuda12.6-cudnn9-py3.13-gcc9-inductor-benchmarks)
199-
CUDA_VERSION=12.6
200-
CUDNN_VERSION=9
201-
ANACONDA_PYTHON_VERSION=3.13
202-
GCC_VERSION=9
203-
VISION=yes
204-
KATEX=yes
205-
UCX_COMMIT=${_UCX_COMMIT}
206-
UCC_COMMIT=${_UCC_COMMIT}
207-
TRITON=yes
208-
INDUCTOR_BENCHMARKS=yes
209-
;;
210157
pytorch-linux-jammy-cuda12.8-cudnn9-py3-gcc9)
211158
CUDA_VERSION=12.8.1
212-
CUDNN_VERSION=9
213159
ANACONDA_PYTHON_VERSION=3.10
214160
GCC_VERSION=9
215161
VISION=yes
@@ -230,19 +176,7 @@ case "$tag" in
230176
VISION=yes
231177
TRITON=yes
232178
;;
233-
pytorch-linux-jammy-py3.11-clang12)
234-
ANACONDA_PYTHON_VERSION=3.11
235-
CLANG_VERSION=12
236-
VISION=yes
237-
TRITON=yes
238-
;;
239-
pytorch-linux-jammy-py3.9-gcc9)
240-
ANACONDA_PYTHON_VERSION=3.9
241-
GCC_VERSION=9
242-
VISION=yes
243-
TRITON=yes
244-
;;
245-
pytorch-linux-jammy-rocm-n-py3 | pytorch-linux-noble-rocm-n-py3)
179+
pytorch-linux-jammy-rocm-n-py3 | pytorch-linux-jammy-rocm-n-py3-benchmarks | pytorch-linux-noble-rocm-n-py3)
246180
if [[ $tag =~ "jammy" ]]; then
247181
ANACONDA_PYTHON_VERSION=3.10
248182
else
@@ -256,7 +190,9 @@ case "$tag" in
256190
KATEX=yes
257191
UCX_COMMIT=${_UCX_COMMIT}
258192
UCC_COMMIT=${_UCC_COMMIT}
259-
INDUCTOR_BENCHMARKS=yes
193+
if [[ $tag =~ "benchmarks" ]]; then
194+
INDUCTOR_BENCHMARKS=yes
195+
fi
260196
;;
261197
pytorch-linux-noble-rocm-alpha-py3)
262198
ANACONDA_PYTHON_VERSION=3.12
@@ -268,7 +204,6 @@ case "$tag" in
268204
KATEX=yes
269205
UCX_COMMIT=${_UCX_COMMIT}
270206
UCC_COMMIT=${_UCC_COMMIT}
271-
INDUCTOR_BENCHMARKS=yes
272207
PYTORCH_ROCM_ARCH="gfx90a;gfx942;gfx950"
273208
;;
274209
pytorch-linux-jammy-xpu-2025.0-py3)
@@ -299,7 +234,6 @@ case "$tag" in
299234
pytorch-linux-jammy-cuda12.8-cudnn9-py3.9-clang12)
300235
ANACONDA_PYTHON_VERSION=3.9
301236
CUDA_VERSION=12.8.1
302-
CUDNN_VERSION=9
303237
CLANG_VERSION=12
304238
VISION=yes
305239
TRITON=yes
@@ -378,7 +312,6 @@ case "$tag" in
378312
fi
379313
if [[ "$image" == *cuda* ]]; then
380314
extract_version_from_image_name cuda CUDA_VERSION
381-
extract_version_from_image_name cudnn CUDNN_VERSION
382315
fi
383316
if [[ "$image" == *rocm* ]]; then
384317
extract_version_from_image_name rocm ROCM_VERSION
@@ -430,9 +363,6 @@ docker build \
430363
--build-arg "PYTHON_VERSION=${PYTHON_VERSION}" \
431364
--build-arg "GCC_VERSION=${GCC_VERSION}" \
432365
--build-arg "CUDA_VERSION=${CUDA_VERSION}" \
433-
--build-arg "CUDNN_VERSION=${CUDNN_VERSION}" \
434-
--build-arg "TENSORRT_VERSION=${TENSORRT_VERSION}" \
435-
--build-arg "GRADLE_VERSION=${GRADLE_VERSION}" \
436366
--build-arg "NINJA_VERSION=${NINJA_VERSION:-}" \
437367
--build-arg "KATEX=${KATEX:-}" \
438368
--build-arg "ROCM_VERSION=${ROCM_VERSION:-}" \

‎.github/ci_commit_pins/torchbench.txt renamed to ‎.ci/docker/ci_commit_pins/torchbench.txt

File renamed without changes.

‎.ci/docker/ci_commit_pins/triton.txt

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1 +1 @@
1-
11ec6354315768a85da41032535e3b7b99c5f706
1+
f7888497a1eb9e98d4c07537f0d0bcfe180d1363

‎.ci/docker/common/install_cpython.sh

Lines changed: 3 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -66,8 +66,9 @@ function do_cpython_build {
6666
ln -s pip3 ${prefix}/bin/pip
6767
fi
6868
# install setuptools since python 3.12 is required to use distutils
69-
${prefix}/bin/pip install wheel==0.45.1 setuptools==80.9.0
70-
local abi_tag=$(${prefix}/bin/python -c "from wheel.pep425tags import get_abbr_impl, get_impl_ver, get_abi_tag; print('{0}{1}-{2}'.format(get_abbr_impl(), get_impl_ver(), get_abi_tag()))")
69+
# packaging is needed to create symlink since wheel no longer provides needed information
70+
${prefix}/bin/pip install packaging==25.0 wheel==0.45.1 setuptools==80.9.0
71+
local abi_tag=$(${prefix}/bin/python -c "from packaging.tags import interpreter_name, interpreter_version; import sysconfig ; from sysconfig import get_config_var; print('{0}{1}-{0}{1}{2}'.format(interpreter_name(), interpreter_version(), 't' if sysconfig.get_config_var('Py_GIL_DISABLED') else ''))")
7172
ln -sf ${prefix} /opt/python/${abi_tag}
7273
}
7374

‎.ci/docker/common/install_cuda.sh

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -68,8 +68,8 @@ function install_nvshmem {
6868
# download, unpack, install
6969
wget -q "${url}"
7070
tar xf "${filename}.tar.gz"
71-
cp -a "libnvshmem/include/"* /usr/local/include/
72-
cp -a "libnvshmem/lib/"* /usr/local/lib/
71+
cp -a "libnvshmem/include/"* /usr/local/cuda/include/
72+
cp -a "libnvshmem/lib/"* /usr/local/cuda/lib64/
7373

7474
# cleanup
7575
cd ..

‎.ci/docker/common/install_cudnn.sh

Lines changed: 0 additions & 26 deletions
This file was deleted.

‎.ci/docker/common/install_inductor_benchmark_deps.sh

Lines changed: 28 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -15,11 +15,37 @@ function install_timm() {
1515
commit=$(get_pinned_commit timm)
1616

1717
pip_install "git+https://github.com/huggingface/pytorch-image-models@${commit}"
18-
# Clean up
19-
conda_run pip uninstall -y torch torchvision triton
18+
}
19+
20+
function install_torchbench() {
21+
local commit
22+
commit=$(get_pinned_commit torchbench)
23+
git clone https://github.com/pytorch/benchmark torchbench
24+
pushd torchbench
25+
git checkout "$commit"
26+
27+
python install.py --continue_on_fail
28+
29+
# TODO (huydhn): transformers-4.44.2 added by https://github.com/pytorch/benchmark/pull/2488
30+
# is regressing speedup metric. This needs to be investigated further
31+
pip install transformers==4.38.1
32+
33+
echo "Print all dependencies after TorchBench is installed"
34+
python -mpip freeze
35+
popd
36+
37+
chown -R jenkins torchbench
2038
}
2139

2240
# Pango is needed for weasyprint which is needed for doctr
2341
conda_install pango
42+
43+
# Stable packages are ok here, just to satisfy TorchBench check
44+
pip_install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128
45+
46+
install_torchbench
2447
install_huggingface
2548
install_timm
49+
50+
# Clean up
51+
conda_run pip uninstall -y torch torchvision torchaudio triton

‎.ci/docker/common/install_xpu.sh

Lines changed: 25 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -34,18 +34,27 @@ function install_ubuntu() {
3434

3535
# The xpu-smi packages
3636
apt-get install -y flex bison xpu-smi
37-
# Compute and Media Runtimes
38-
apt-get install -y \
39-
intel-opencl-icd intel-level-zero-gpu level-zero \
40-
intel-media-va-driver-non-free libmfx1 libmfxgen1 libvpl2 \
41-
libegl-mesa0 libegl1-mesa libegl1-mesa-dev libgbm1 libgl1-mesa-dev libgl1-mesa-dri \
42-
libglapi-mesa libgles2-mesa-dev libglx-mesa0 libigdgmm12 libxatracker2 mesa-va-drivers \
43-
mesa-vdpau-drivers mesa-vulkan-drivers va-driver-all vainfo hwinfo clinfo
44-
if [[ "${XPU_DRIVER_TYPE,,}" == "rolling" ]]; then
45-
apt-get install -y intel-ocloc
37+
38+
if [[ "${XPU_DRIVER_TYPE,,}" == "lts" ]]; then
39+
# Compute and Media Runtimes
40+
apt-get install -y \
41+
intel-opencl-icd intel-level-zero-gpu level-zero \
42+
intel-media-va-driver-non-free libmfx1 libmfxgen1 libvpl2 \
43+
libegl-mesa0 libegl1-mesa libegl1-mesa-dev libgbm1 libgl1-mesa-dev libgl1-mesa-dri \
44+
libglapi-mesa libgles2-mesa-dev libglx-mesa0 libigdgmm12 libxatracker2 mesa-va-drivers \
45+
mesa-vdpau-drivers mesa-vulkan-drivers va-driver-all vainfo hwinfo clinfo
46+
# Development Packages
47+
apt-get install -y libigc-dev intel-igc-cm libigdfcl-dev libigfxcmrt-dev level-zero-dev
48+
else # rolling driver
49+
apt-get install -y \
50+
intel-opencl-icd libze-intel-gpu1 libze1 \
51+
intel-media-va-driver-non-free libmfx-gen1 libvpl2 \
52+
libegl-mesa0 libegl1-mesa libegl1-mesa-dev libgbm1 libgl1-mesa-dev libgl1-mesa-dri \
53+
libglapi-mesa libglx-mesa0 libigdgmm12 libxatracker2 mesa-va-drivers \
54+
mesa-vdpau-drivers mesa-vulkan-drivers va-driver-all vainfo hwinfo clinfo intel-ocloc
55+
apt-get install -y libigc-dev intel-igc-cm libigdfcl-dev libigfxcmrt-dev libze-dev
4656
fi
47-
# Development Packages
48-
apt-get install -y libigc-dev intel-igc-cm libigdfcl-dev libigfxcmrt-dev level-zero-dev
57+
4958
# Install Intel Support Packages
5059
apt-get install -y ${XPU_PACKAGES}
5160

@@ -130,11 +139,11 @@ function install_sles() {
130139

131140
}
132141

133-
# Default use GPU driver LTS releases
134-
XPU_DRIVER_VERSION="/lts/2350"
135-
if [[ "${XPU_DRIVER_TYPE,,}" == "rolling" ]]; then
136-
# Use GPU driver rolling releases
137-
XPU_DRIVER_VERSION=""
142+
# Default use GPU driver rolling releases
143+
XPU_DRIVER_VERSION=""
144+
if [[ "${XPU_DRIVER_TYPE,,}" == "lts" ]]; then
145+
# Use GPU driver LTS releases
146+
XPU_DRIVER_VERSION="/lts/2350"
138147
fi
139148

140149
# Default use Intel® oneAPI Deep Learning Essentials 2025.0

0 commit comments

Comments
 (0)