[DRAFT] Added magma conda build scripts for ROCm #27
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* add dependencies for rocm5.1 * install miopen dependencies using cmake * change lib path in build_rocm.sh for rocm5.1 * change amdgpu_version for rocm5.1 * remove 4.5.2 and add 5.1
Libdrm rocm fix
IFU 2022-05-16
[ROCm] add dependencies for rocm5.2
* change the MAYBE_LIB64 path for just for 5.2
* Changes to support ROCm 5.3 * Updated as per comments
- In ROCm 5.3 libtorch build are failing during magma build due to to missing python binary so added install statement
* Updating the condition for noRCCL build * Updated changes as per comments
IFU-main-2022-09-15
…rocm_fork Disable MLIR backend when building MIOpen
Use ROCm5.3 branch for MIOpen
…taging_branch Use staging branch of MIOpen for ROCm5.3
* Update to so patching for ROCm
Wildcard used in grep to grab the actual numbered so file referenced
in patchelf. This allows the removal of specifying the so number in
DEPS_LIST & DEPS_SONAME
This commit also adds the functionality for trimming so names to
build_libtorch.sh from build_common.sh
* Refactor to remove switch statement in build_rocm.sh
This commit refactors build_rocm.sh and brings in a few major updates:
- No longer required to specify the full .so name (with number) for ROCm libraries
- The .so versions are copied and the patching code will fix the links to point to this version
- No longer required to specify paths for ROCm libraries allowing the removal of the large switch
- Paths are acquired programmatically with find
- No longer required to specify both the path and filename for the OS specific libraries
- Programatically extract file name from the path
- Automatically extract Tensile/Kernels files for the architectures specified in PYTORCH_ROCM_ARCH
and any non-arch specific files e.g. TensileLibrary.dat
* Remove miopen custom build step * Bundle MIOpen db files in wheel * Correct path
* Use libtinfo.so.6 for Ubuntu 2004 * Fix to origname grep * Condition on ROCM_VERSION for libtinfo6
We require the same fix that was made on upstream pytorch pytorch/pytorch#91371 ROCm/pytorch@b72ec7c Without this change install_conda.sh stage fails ``` #21 6.254 CondaFileIOError: '/opt/conda/pkgs/envs/*/env.txt'. [Errno 2] No such file or directory: '/opt/conda/pkgs/envs/*/env.txt' #21 6.254 #21 ERROR: executor failed running [/bin/sh -c bash ./install_conda.sh && rm install_conda.sh]: exit code: 1 ------ > [conda 2/3] RUN bash ./install_conda.sh && rm install_conda.sh: ------ executor failed running [/bin/sh -c bash ./install_conda.sh && rm install_conda.sh]: exit code: 1 ``` Locally tested with the `/builder/libtorch/build_docker.sh`
Co-authored-by: Jack Taylor <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Added a draft PR for magma conda build to facilitate discussion. Presently I overwrite the magma dir with the changes, I can change this before considering merging.
I have tested building this package locally (the tars are too big to store here but logs are stored in magma-rocm5.3 and magma-rocm5.4)
Testing
make magma-rocm5.3ormake magma-rocm5.4./opt/rocm(as done in this PRscommon/install_magma.shmimicing the way CUDA handles this) - uploaded asrocm/pytorch-private:manylinux-rocm5.4-131-magma-wheel-testUnit Test
What's next
@pruthvistony @jithunnair-amd
This gives a base for us to now make any changes to get this ready. We need to decide on
install_magma.shto test a CI run.