https://blogs.gentoo.org/ago/2020/07/04/gentoo-tinderbox/ Issue: dev-util/nvidia-cuda-sdk-11.5.1 fails to compile. Discovered on: amd64 (internal ref: ci) NOTE: If you think this is a GCC-11 related issue, please block bug 732706.
Created attachment 761908 [details] build.log.xz build log and emerge --info (compressed because it exceeds attachment limit, use 'xzless' to read it)
Error(s) that match a know pattern in addition to what has been reported in the summary: Package 'glfw3', required by 'virtual:world', not found /usr/lib/gcc/x86_64-pc-linux-gnu/11.2.1/include/g++-v11/bits/std_function.h:435:145: error: parameter packs not expanded with ‘...’:
The same problem affects sci-chemistry/gromacs: usr/lib/gcc/x86_64-pc-linux-gnu/11.2.1/include/g++-v11/bits/std_function.h:530:146: error: parameter packs not expanded with ‘...’: I rebuilt sci-chemistry/gromacs (-2021.4, -2021.5, -2022_rc1) and its dependency dev-util/nvidia-cuda-toolkit (not -sdk) with gcc-10.3.0-r2 and gcc-11.2.1_p20220115 with same result, gromacs fails to build.
I've hit this too, it's just a GCC that is too new for CUDA. I've worked around this by compiling with C++14 rather than C++17, or by using an older version of GCC. The next CUDA release should have this fixed.
Even when adding -std=c++14 to CXXFLAGS it still won't compile with the same errors given. Using GCC 11.2.1.
Please switch to the latest CUDA (11.6.1) if you can. This is an upstream bug that has been fixed in the newer versions.
This isn't resolved, Gentoo's nvidia-cuda-sdk version is still 11.5.1
(In reply to Red from comment #7) > This isn't resolved, Gentoo's nvidia-cuda-sdk version is still 11.5.1 Use the toolkit instead?
(In reply to Sam James from comment #8) > (In reply to Red from comment #7) > > This isn't resolved, Gentoo's nvidia-cuda-sdk version is still 11.5.1 > > Use the toolkit instead? Are they the same thing? I was only installing nvidia-cuda-sdk because it was an optional dependency of another package (I forget which).
