sci-libs/caffe2-2.0.1-r4 USE=cuda depends on <nvidia-cuda-toolkit-12, as upstream compatibility (pytorch with cuda12?) seems already stable.. sci-libs/caffe2 can (and maybe should?) use cuda-12 if available Reproducible: Always Actual Results: When changed dependency for sci-libs/caffe2-2.0.1-r4 for cuda USE to <=nvidia-cuda-toolkit-12, no additional dependencies for gcc11 and nvidia-cuda-toolkit-11 introduced in my system
upstream issue is already resolved, applicable for all caffe2-2.x see https://github.com/pytorch/pytorch/issues/91122
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=60b645821036cba6aecb772566881d9df6f8786e commit 60b645821036cba6aecb772566881d9df6f8786e Author: Alfredo Tupone <tupone@gentoo.org> AuthorDate: 2023-12-17 09:30:11 +0000 Commit: Alfredo Tupone <tupone@gentoo.org> CommitDate: 2023-12-17 09:30:11 +0000 sci-libs/caffe2: enable nvidia-cuda-toolkit-12 Closes: https://bugs.gentoo.org/917585 Signed-off-by: Alfredo Tupone <tupone@gentoo.org> sci-libs/caffe2/{caffe2-2.1.1-r2.ebuild => caffe2-2.1.1-r3.ebuild} | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
