-
-
Notifications
You must be signed in to change notification settings - Fork 4.9k
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.
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
Adding cublasMp recipes #26580
Merged
Merged
Adding cublasMp recipes #26580
Changes from 8 commits
Commits
Show all changes
15 commits
Select commit
Hold shift + click to select a range
4d2829f
Adding cublasMp recipes
adibbley afc5d79
Fix spacing
adibbley a998642
Remove extra requirements
adibbley c96057e
Apply suggestions from code review
adibbley 6fe4aed
Apply suggestions from code review
adibbley 56b6ec8
Apply suggestions from code review
adibbley 243a450
Update recipes/libcublasmp/conda_build_config.yaml
adibbley be6c603
Update recipes/libcal/meta.yaml
adibbley dafab1d
Merge remote-tracking branch 'upstream/main' into cublasmp
carterbox f5450fa
REF: Our channel pins cuda_compiler_version not cuda_version
carterbox e8c99bb
Drop share and src dirs
adibbley 1c7db1a
Add run_exports
adibbley 103b8c7
Add compiler('cuda')
adibbley 76d55bc
Add conda-build usage note
adibbley c807d53
Merge branch 'conda-forge:main' into cublasmp
adibbley File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
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
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,38 @@ | ||
{% set version = "0.2.1" %} | ||
|
||
package: | ||
name: cublasmp | ||
version: {{ version }} | ||
|
||
source: | ||
url: https://developer.download.nvidia.com/compute/cublasmp/redist/libcublasmp/LICENSE.txt | ||
sha256: c2a95a3d181879325e56333d7a592bb6ba5ab3a44bfde9574513ccc2f87f9c68 | ||
|
||
build: | ||
number: 0 | ||
noarch: generic | ||
carterbox marked this conversation as resolved.
Show resolved
Hide resolved
|
||
|
||
requirements: | ||
run: | ||
- libcal-dev 0.4.3.36 | ||
- libcublasmp-dev 0.2.1.427 | ||
|
||
test: | ||
commands: | ||
- exit 0 | ||
|
||
about: | ||
home: https://developer.nvidia.com/cuda-toolkit | ||
license_file: LICENSE.txt | ||
license: LicenseRef-NVIDIA-End-User-License-Agreement | ||
license_url: https://docs.nvidia.com/cuda/eula/index.html | ||
summary: Meta-package containing all NVIDIA cublasMp packages. | ||
description: | | ||
NVIDIA cublasMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
cuBLASMp is compatible with 2D block-cyclic data layout and provides PBLAS-like C APIs. | ||
A companion library, CAL, contains utilities to manage communicators and to synchronize processes in a safe way. | ||
doc_url: https://docs.nvidia.com/cuda/index.html | ||
|
||
extra: | ||
recipe-maintainers: | ||
- conda-forge/cuda |
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
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,2 @@ | ||
cuda_version: | ||
- 11 |
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
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,115 @@ | ||
{% set version = "0.4.3.36" %} | ||
{% set lib_version = version.split(".")[0]+"."+version.split(".")[1]+"."+version.split(".")[2]+".0" %} | ||
{% set platform = "linux-x86_64" %} # [linux64] | ||
{% set platform = "linux-sbsa" %} # [aarch64] | ||
{% set extension = "tar.xz" %} | ||
|
||
package: | ||
name: libcal | ||
version: {{ version }} | ||
|
||
source: | ||
url: https://developer.download.nvidia.com/compute/cublasmp/redist/libcal/{{ platform }}/libcal-{{ platform }}-{{ version }}_cuda{{ cuda_version }}-archive.{{ extension }} | ||
sha256: 271d9951ee5b5e4e6e9fcb7079ce3c36f5dd581b3a0f7d1937f81e0c0e058fb8 # [linux64 and cuda_version == 11] | ||
sha256: eac07cc1524b37c693944e3e91c4857ed66bc2bb2579407ce9795916c9a113a4 # [linux64 and cuda_version == 12] | ||
sha256: d0016b62d42916e37cd065279eb8fa4947e3423220812c696c09ea417477c770 # [aarch64 and cuda_version == 11] | ||
sha256: 3500c18236fc449cf0395ee422c5a1771ce6349dc547951eaa6562ffc4fc762c # [aarch64 and cuda_version == 12] | ||
|
||
build: | ||
number: 0 | ||
skip: true # [not (linux64 or aarch64)] | ||
missing_dso_whitelist: | ||
- "*/libcuda.so.*" | ||
script: | ||
- cp -rv include $PREFIX | ||
- cp -rv lib $PREFIX | ||
- mkdir -p $PREFIX/share/libcal | ||
- cp -v share/* $PREFIX/share/libcal/ | ||
- cp -rv src $PREFIX | ||
adibbley marked this conversation as resolved.
Show resolved
Hide resolved
adibbley marked this conversation as resolved.
Show resolved
Hide resolved
|
||
|
||
test: | ||
commands: | ||
- test -L $PREFIX/lib/libcal.so.0 | ||
- test -f $PREFIX/lib/libcal.so.{{ lib_version }} | ||
|
||
outputs: | ||
- name: libcal | ||
files: | ||
- lib/libcal.so.* | ||
adibbley marked this conversation as resolved.
Show resolved
Hide resolved
|
||
requirements: | ||
build: | ||
- {{ compiler('c') }} | ||
- {{ compiler('cxx') }} | ||
- arm-variant * {{ arm_variant_type }} # [aarch64] | ||
- {{ stdlib("c") }} | ||
host: | ||
- cuda-version {{ cuda_version }} | ||
- ucc | ||
run: | ||
- {{ pin_compatible("cuda-version", max_pin="x") }} | ||
- ucc | ||
run_constrained: | ||
- arm-variant * {{ arm_variant_type }} # [aarch64] | ||
# Tests are defined at the top level, due to package/output name conflicts. | ||
about: | ||
home: https://docs.nvidia.com/cuda/cublasmp/ | ||
license: LicenseRef-NVIDIA-End-User-License-Agreement | ||
license_file: LICENSE | ||
license_url: https://docs.nvidia.com/cuda/cublasmp/license/index.html | ||
summary: NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
description: | | ||
NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
cuBLASMp is compatible with 2D block-cyclic data layout and provides PBLAS-like C APIs. | ||
A companion library, CAL, contains utilities to manage communicators and to synchronize processes in a safe way. | ||
doc_url: https://docs.nvidia.com/cuda/cublas/ | ||
|
||
- name: libcal-dev | ||
build: | ||
run_exports: | ||
- {{ pin_subpackage("libcal", max_pin="x") }} | ||
files: | ||
- lib/libcal.so | ||
- include/cal.h | ||
- share/libcal/ucc.conf | ||
- src/libcal | ||
adibbley marked this conversation as resolved.
Show resolved
Hide resolved
|
||
requirements: | ||
host: | ||
- cuda-version {{ cuda_version }} | ||
run: | ||
- {{ pin_compatible("cuda-version", max_pin="x") }} | ||
- {{ pin_subpackage("libcal", exact=True) }} | ||
run_constrained: | ||
- arm-variant * {{ arm_variant_type }} # [aarch64] | ||
test: | ||
commands: | ||
- test -f $PREFIX/include/cal.h | ||
- test -L $PREFIX/lib/libcal.so | ||
- test -f $PREFIX/share/libcal/ucc.conf | ||
- test -d $PREFIX/src/libcal | ||
adibbley marked this conversation as resolved.
Show resolved
Hide resolved
|
||
about: | ||
home: https://docs.nvidia.com/cuda/cublasmp/ | ||
license: LicenseRef-NVIDIA-End-User-License-Agreement | ||
license_file: LICENSE | ||
license_url: https://docs.nvidia.com/cuda/cublasmp/license/index.html | ||
summary: NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
description: | | ||
NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
cuBLASMp is compatible with 2D block-cyclic data layout and provides PBLAS-like C APIs. | ||
A companion library, CAL, contains utilities to manage communicators and to synchronize processes in a safe way. | ||
doc_url: https://docs.nvidia.com/cuda/cublas/ | ||
|
||
about: | ||
home: https://docs.nvidia.com/cuda/cublasmp/ | ||
license: LicenseRef-NVIDIA-End-User-License-Agreement | ||
license_file: LICENSE | ||
license_url: https://docs.nvidia.com/cuda/cublasmp/license/index.html | ||
summary: NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
description: | | ||
NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
cuBLASMp is compatible with 2D block-cyclic data layout and provides PBLAS-like C APIs. | ||
A companion library, CAL, contains utilities to manage communicators and to synchronize processes in a safe way. | ||
doc_url: https://docs.nvidia.com/cuda/cublas/ | ||
|
||
extra: | ||
recipe-maintainers: | ||
- conda-forge/cuda |
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
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,2 @@ | ||
cuda_version: | ||
- 11 |
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
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,110 @@ | ||
{% set version = "0.2.1.427" %} | ||
{% set lib_version = version.split(".")[0]+"."+version.split(".")[1]+"."+version.split(".")[2] %} | ||
{% set platform = "linux-x86_64" %} # [linux64] | ||
{% set platform = "linux-sbsa" %} # [aarch64] | ||
{% set extension = "tar.xz" %} | ||
|
||
package: | ||
name: libcublasmp | ||
version: {{ version }} | ||
|
||
source: | ||
url: https://developer.download.nvidia.com/compute/cublasmp/redist/libcublasmp/{{ platform }}/libcublasmp-{{ platform }}-{{ version }}_cuda{{ cuda_version }}-archive.{{ extension }} | ||
sha256: 891e86a28cfe08772f2485e4db7ec8d28292e6a1b0e78967d08404c8371925b5 # [linux64 and cuda_version == 11] | ||
sha256: d86648a6e1f1d35eb25af258e1828e483aa07c568b95941f349fc68316f56290 # [linux64 and cuda_version == 12] | ||
sha256: 5f5813814343e83eeb743ddb4b1df0b346cdbf2b2026b32a0bc4a349361ec8e4 # [aarch64 and cuda_version == 11] | ||
sha256: d6cacf073bf87987b0f20a7a5a168ec503264d295e3aebf9df177326e8fcb286 # [aarch64 and cuda_version == 12] | ||
|
||
build: | ||
number: 0 | ||
skip: true # [not (linux64 or aarch64)] | ||
script: | ||
- cp -rv include $PREFIX | ||
- cp -rv lib $PREFIX | ||
|
||
test: | ||
commands: | ||
- test -L $PREFIX/lib/libcublasmp.so.0 | ||
- test -f $PREFIX/lib/libcublasmp.so.{{ lib_version }} | ||
|
||
outputs: | ||
- name: libcublasmp | ||
files: | ||
- lib/libcublasmp.so.* | ||
adibbley marked this conversation as resolved.
Show resolved
Hide resolved
|
||
requirements: | ||
build: | ||
- {{ compiler('c') }} | ||
- {{ compiler('cxx') }} | ||
- arm-variant * {{ arm_variant_type }} # [aarch64] | ||
- {{ stdlib("c") }} | ||
host: | ||
- cuda-version {{ cuda_version }} | ||
- libcublas # [cuda_version == 12] | ||
adibbley marked this conversation as resolved.
Show resolved
Hide resolved
|
||
- cudatoolkit # [cuda_version == 11] | ||
- libcal 0.4.3.* | ||
run: | ||
- {{ pin_compatible("cuda-version", max_pin="x") }} | ||
- libcublas # [cuda_version == 12] | ||
- cudatoolkit # [cuda_version == 11] | ||
- {{ pin_compatible("libcal", exact=True) }} | ||
run_constrained: | ||
- arm-variant * {{ arm_variant_type }} # [aarch64] | ||
# Tests are defined at the top level, due to package/output name conflicts. | ||
about: | ||
home: https://docs.nvidia.com/cuda/cublasmp/ | ||
license: LicenseRef-NVIDIA-End-User-License-Agreement | ||
license_file: LICENSE | ||
license_url: https://docs.nvidia.com/cuda/cublasmp/license/index.html | ||
summary: NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
description: | | ||
NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
cuBLASMp is compatible with 2D block-cyclic data layout and provides PBLAS-like C APIs. | ||
A companion library, CAL, contains utilities to manage communicators and to synchronize processes in a safe way. | ||
doc_url: https://docs.nvidia.com/cuda/cublas/ | ||
|
||
- name: libcublasmp-dev | ||
build: | ||
run_exports: | ||
- {{ pin_subpackage("libcublasmp", max_pin="x") }} | ||
files: | ||
- lib/libcublasmp.so | ||
- include/cublasmp.h | ||
requirements: | ||
host: | ||
- cuda-version {{ cuda_version }} | ||
run: | ||
- {{ pin_compatible("cuda-version", max_pin="x") }} | ||
- {{ pin_subpackage("libcublasmp", exact=True) }} | ||
run_constrained: | ||
- arm-variant * {{ arm_variant_type }} # [aarch64] | ||
test: | ||
commands: | ||
- test -f $PREFIX/include/cublasmp.h | ||
- test -L $PREFIX/lib/libcublasmp.so | ||
about: | ||
home: https://docs.nvidia.com/cuda/cublasmp/ | ||
license: LicenseRef-NVIDIA-End-User-License-Agreement | ||
license_file: LICENSE | ||
license_url: https://docs.nvidia.com/cuda/cublasmp/license/index.html | ||
summary: NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
description: | | ||
NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
cuBLASMp is compatible with 2D block-cyclic data layout and provides PBLAS-like C APIs. | ||
A companion library, CAL, contains utilities to manage communicators and to synchronize processes in a safe way. | ||
doc_url: https://docs.nvidia.com/cuda/cublas/ | ||
|
||
about: | ||
home: https://docs.nvidia.com/cuda/cublasmp/ | ||
license: LicenseRef-NVIDIA-End-User-License-Agreement | ||
license_file: LICENSE | ||
license_url: https://docs.nvidia.com/cuda/cublasmp/license/index.html | ||
summary: NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
description: | | ||
NVIDIA cuBLASMp is a high performance, multi-process, GPU accelerated library for distributed basic dense linear algebra. | ||
cuBLASMp is compatible with 2D block-cyclic data layout and provides PBLAS-like C APIs. | ||
A companion library, CAL, contains utilities to manage communicators and to synchronize processes in a safe way. | ||
doc_url: https://docs.nvidia.com/cuda/cublas/ | ||
|
||
extra: | ||
recipe-maintainers: | ||
- conda-forge/cuda |
Oops, something went wrong.
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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Q: I wonder if it'd make things easier to make the
cublasmp
meta package an output of libcublasmp-feedstock? (No strong opinion, just curious if this has been considered.)There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Keeping them separate helps with how we setup releases internally. The recipes would be updated with a similar process to CUDA, just on a smaller scale.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do you not want run_exports for this package? Otherwise, what is the purpose of this package? No downstream packages should depend on this pacakge directly because they would get all of the devlopement files for libcal and libcublasmp but none of the run_exports.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Same question. What's the purpose of this package? Also, I don't see a compelling reason to keep this in a separate feedstock.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This package is for convenient installation of all components, similar to the cuda metapackages.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Please add that this is not to be used in a conda-build context in the summary.
Also, please move this to libcublasmp recipe if there's not reason to keep this in a separate feedstock.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Added a note now.
Having separate feedstocks helps us with maintenance by keeping the design consistent across products.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That doesn't make sense to me at all.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can you explain in more detail?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
All CUDA metapackages are in separate feedstocks, I'm not seeing a reason to change that for cublasmp. Versions of these packages are not technically coupled, they are given a release label as seen in https://developer.download.nvidia.com/compute/cublasmp/redist/redistrib_0.2.1.json