Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Update main post 24.06 #135

Merged
merged 3 commits into from
Jun 28, 2024
Merged

Update main post 24.06 #135

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
b157bbf
Add cusparseLt in the installation to support 24.06 (#132)
tanmayv25 Jun 18, 2024
cf8c195
Fix library paths on ARM-SBSA (#133)
tanmayv25 Jun 19, 2024
ddfbe3a
Fix location of the dependency library. (#134)
mc-nv Jun 24, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -229,6 +229,8 @@ if (${TRITON_PYTORCH_DOCKER_BUILD})
COMMAND docker cp pytorch_backend_ptlib:/usr/local/lib/python3.10/dist-packages/torch/lib/libtorch_cuda_linalg.so libtorch_cuda_linalg.so
COMMAND docker cp pytorch_backend_ptlib:/usr/local/lib/python3.10/dist-packages/torch/lib/libtorch_global_deps.so libtorch_global_deps.so
COMMAND docker cp pytorch_backend_ptlib:/usr/local/lib/python3.10/dist-packages/torch/lib/libcaffe2_nvrtc.so libcaffe2_nvrtc.so
# TODO: Revisit when not needed by making it part of cuda base container.
COMMAND docker cp -L pytorch_backend_ptlib:/usr/local/cuda/lib64/libcusparseLt.so libcusparseLt.so
COMMAND docker cp pytorch_backend_ptlib:/usr/local/lib/libtorchvision.so libtorchvision.so
COMMAND /bin/sh -c "if [ ${TRITON_PYTORCH_ENABLE_TORCHTRT} = 'ON' ]; then docker cp pytorch_backend_ptlib:/usr/local/lib/python3.10/dist-packages/torch_tensorrt/lib/libtorchtrt_runtime.so libtorchtrt_runtime.so; fi"
COMMAND docker cp pytorch_backend_ptlib:/usr/local/lib/python3.10/dist-packages/torch_tensorrt/bin/torchtrtc torchtrtc || echo "error ignored..." || true
Expand Down Expand Up @@ -434,6 +436,7 @@ if (${TRITON_PYTORCH_DOCKER_BUILD})
install(
FILES
${PT_LIB_PATHS}
${CMAKE_CURRENT_BINARY_DIR}/libcusparseLt.so
${CMAKE_CURRENT_BINARY_DIR}/LICENSE.pytorch
DESTINATION ${CMAKE_INSTALL_PREFIX}/backends/pytorch
)
Expand Down Expand Up @@ -474,6 +477,7 @@ if (${TRITON_PYTORCH_DOCKER_BUILD})
COMMAND ln -sf libopencv_flann.so libopencv_flann.so.${OPENCV_VERSION}
COMMAND ln -sf libpng16.so libpng16.so.16
COMMAND ln -sf libjpeg.so libjpeg.so.8
COMMAND ln -sf libcusparseLt.so libcusparseLt.so.0
RESULT_VARIABLE LINK_STATUS
WORKING_DIRECTORY ${CMAKE_INSTALL_PREFIX}/backends/pytorch)
if(LINK_STATUS AND NOT LINK_STATUS EQUAL 0)
Expand Down
4 changes: 2 additions & 2 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -146,11 +146,11 @@ key: "INFERENCE_MODE"

* `DISABLE_CUDNN`: Boolean flag to disable the cuDNN library. By default, cuDNN is enabled.

[cuDNN](https://developer.nvidia.com/cudnn) is a GPU-accelerated library of primitives for
[cuDNN](https://developer.nvidia.com/cudnn) is a GPU-accelerated library of primitives for
deep neural networks. cuDNN provides highly tuned implementations for standard routines.

Typically, models run with cuDNN enabled are faster. However there are some exceptions
where using cuDNN can be slower, cause higher memory usage or result in errors.
where using cuDNN can be slower, cause higher memory usage or result in errors.


The section of model config file specifying this parameter will look like:
Expand Down
Loading