Giters
cupy
/
cupy-release-tools
Tools for CuPy Release
Geek Repo:
Geek Repo
Github PK Tool:
Github PK Tool
Stargazers:
6
Watchers:
9
Issues:
30
Forks:
11
cupy/cupy-release-tools Issues
Migrate to manylinux2014
Closed
2 years ago
Consider using cibuildwheel
Updated
2 years ago
Comments count
2
Re-enable verification with CUDA 11.6
Closed
2 years ago
Comments count
1
Remove Nsight from builder base docker image
Updated
2 years ago
Propagate signals to docker command launched in `dist.py`
Updated
2 years ago
Use BuildKit so that Docker images can be reused from remote Docker registry
Closed
2 years ago
Use manylinux as a base builder docker image
Closed
2 years ago
Comments count
1
Improve `WHEEL_LINUX_CONFIGS` structure in `dist_config.py`
Updated
2 years ago
Remove ROCm 4.0 in `master` branch
Closed
2 years ago
Use Visual Studio 2017
Closed
2 years ago
Comments count
1
Test wheel build nightly for stables branch as well
Updated
2 years ago
Fix centos-vault is used for CentOS 7 in builder
Closed
3 years ago
Record file hash and commit ID used for each release
Updated
3 years ago
Detect failure on `pyenv install`
Updated
3 years ago
Add FlexCI config to test sdist build
Updated
3 years ago
Do `twine check` to check `long_description` is valid
Updated
3 years ago
Bump Python versions
Closed
3 years ago
Add an option to only `pyenv install` Python version needed for the build
Updated
3 years ago
Use newer CUDA for sdist build
Closed
3 years ago
Windows: preload is not tested
Closed
3 years ago
Dedicated description for ROCm packages
Closed
3 years ago
Remove `cudnn.h` inclusion after v8 release
Closed
3 years ago
Install cuTENSOR and NCCL using `cupyx.tools.install_library`
Closed
3 years ago
Use `cupyx.tools.install_library` instead of `install_cudnn.ps1` on Windows
Closed
3 years ago
Build with cuTENSOR enabled
Closed
3 years ago
Comments count
1
Merge `download_nccl.sh` into dist tool
Closed
3 years ago
Update README
Closed
3 years ago
cuTENSOR must be available for sdist build
Closed
3 years ago
Drop Python 2 support
Closed
3 years ago
Add CUDA 11.0 and drop 8.0/9.1
Closed
3 years ago