[CUDA] Support CUDA 9 and test different CUDA versions at CI #3880
Conversation
StrikerRUS
requested review from
btrotta,
chivee,
guolinke,
henry0312,
huanzhang12,
jameslamb,
Laurae2 and
wxchan
as code owners
| curl -sL https://apt.kitware.com/keys/kitware-archive-latest.asc | apt-key add - | ||
| apt-add-repository "deb https://apt.kitware.com/ubuntu/ $(lsb_release -cs) main" -y | ||
| apt-get update | ||
| apt-get install --no-install-recommends -y \ | ||
| cmake |
There was a problem hiding this comment.
@jameslamb Unfortunately, now we cannot use default CMake here because old CUDA Dockers are based on Ubuntus where we cannot pass CMake >= 3.16 check. 🙁
https://gitlab.com/nvidia/container-images/cuda/blob/master/doc/supported-tags.md
https://packages.ubuntu.com/bionic-updates/cmake
https://packages.ubuntu.com/xenial-updates/cmake
There was a problem hiding this comment.
blegh ok, that's alright
| - method: pip | ||
| python_version: 3.7 | ||
| cuda_version: "10.0" |
There was a problem hiding this comment.
GitHub Actions treats values written as x.0 like ints, so we need explicitly set them to strings.
| set(CUDA_ARCHS "6.0" "6.1" "6.2" "7.0") | ||
| if(CUDA_VERSION VERSION_GREATER_EQUAL "10.0") | ||
| list(APPEND CUDA_ARCHS "7.5") | ||
| endif() | ||
| list(POP_BACK CUDA_ARCHS CUDA_LAST_SUPORTED_VERSION) | ||
| list(APPEND CUDA_ARCHS "${CUDA_LAST_SUPORTED_VERSION}+PTX") | ||
| CUDA_SELECT_NVCC_ARCH_FLAGS(CUDA_ARCH_FLAGS ${CUDA_ARCHS}) |
There was a problem hiding this comment.
cc @ChipKerchner
I hope this change will not break anything on POWER machines.
There was a problem hiding this comment.
@ChipKerchner I'm so sorry for my poor English skills, but could you please rephrase? Do you think that
(1) this change can potentially hurt compilation on POWER systems
OR
(2) this change should be OK for POWER systems
?
And if (1), how this change can be harmful to POWER machines?
There was a problem hiding this comment.
I think this change is OK for POWER systems.
There was a problem hiding this comment.
OK, got it! Thanks a lot for finding time to comment!
|
This pull request has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this. |
I think testing different CUDA versions will greatly help us to be confident in CUDA code.
Linking #3431.
CUDA < 10.0 fails to recognize
7.5architecture.Refer to https://github.com/Kitware/CMake/blob/36bb0e32d7e3976eed424078cf5cac459651f0f1/Modules/FindCUDA/select_compute_arch.cmake#L50-L60.