Fully fused MLPs do not support GPU architectures of 70 or less.

NVlabs / tiny-cuda-nn

Lightning fast C++/CUDA neural network framework

Other

3.54k stars 437 forks source link

Fully fused MLPs do not support GPU architectures of 70 or less. #392

Open Orange-Ctrl opened 7 months ago

Orange-Ctrl commented 7 months ago

Hi, I use rtx 3090 and got this warning which is not supposed to appear. when I use tiny-cuda-nn on other project, I got warning "tinycudann was built for lower compute capability ({cc}) than the system's ({system_compute_capability}). Performance may be suboptimal." So wierd! Has somebody met this problem?

thanks in advance!

xiaochennnnnnnn commented 5 months ago

hello have you got the answer, i still have the same issue, thank you