cudaCheckError() failed : invalid device function about tffrcnn

[Lemonlemons]

Hey Shang,

To fix this I needed to change lines 10 and 25 of lib/make.sh. Specifically I needed to change the -arch sm_52 flag to -arch sm_37 because I'm using a Telsa K80 to match the compute of my GPU architecture. You can find the compute of your GPU at this page: https://developer.nvidia.com/cuda-gpus . Remember to re-make in your lib directory after making the change.

[Kinantb]

So for those who are still lost. Here are a few clean steps to resolve the issue (you need to recompile your CUDA):

1. Got to the following page https://developer.nvidia.com/cuda-gpus and find your GPU
2. Find the number (the "Compute Capability") next to your GPU name, e.g. for 680 it is 3.0
3. Remove the dot from it so it becomes 30
4. In the make_cuda.sh file required for compiling, change the number after "arch" flag in the nvcc command to the one you found above. Example:
   nvcc -c -o corr_cuda_kernel.cu.o corr_cuda_kernel.cu -x cu -Xcompiler -fPIC -arch=sm_52
   nvcc -c -o corr_cuda_kernel.cu.o corr_cuda_kernel.cu -x cu -Xcompiler -fPIC -arch=sm_30
5. Delete the folder of built files if you have already compiled it before
6. Run the make_cuda.sh and continue as usual