#error: "CUDA atomics are only supported for sm_60 and up on *nix and sm_70 and up on Windows" #425
-
I'm getting this error on a Windows machine with device having compute capability 8.9 (4070Ti). I also get this error on a Linux machine with device having compute capability 6.0 (P100). The error is produced by line 12 of atomic_cuda.h. Any thoughts on this? |
Beta Was this translation helpful? Give feedback.
Answered by
PointKernel
Jan 19, 2024
Replies: 1 comment 1 reply
-
Just FYI, building on Windows (#195) is backlogged and won't be addressed in the short term. As for the build error itself, you may want to build your code by explicitly specifying a sm_70+ arch, e.g.: nvcc -std=c++17 -arch=sm_89 --expt-extended-lambda --expt-relaxed-constexpr mycode.cu |
Beta Was this translation helpful? Give feedback.
1 reply
Answer selected by
kevkrist
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Just FYI, building on Windows (#195) is backlogged and won't be addressed in the short term.
As for the build error itself, you may want to build your code by explicitly specifying a sm_70+ arch, e.g.: