Code works on one GPU machine, doesn't work on another

[rusty1s]

Code works on one GPU machine, doesn't work on another

[lookthatdog]

File "/home/xxx/anaconda3/lib/python3.7/site-packages/torch_cluster/knn.py", line 123, in knn_graph
row, col = knn(x, x, k if loop else k + 1, batch, batch, cosine=cosine)

File "/home/xxx/anaconda3/lib/python3.7/site-packages/torch_cluster/knn.py", line 60, in knn
return torch_cluster.knn_cuda.knn(x, y, k, batch_x, batch_y, cosine)

RuntimeError: CUDA error: no kernel image is available for execution on the device (launch_kernel at /opt/conda/conda-bld/pytorch_1573049306803/work/aten/src/ATen/native/cuda/Loops.cuh:102)
I also have this errors on Tesla_P100_PCIE_16GB

[justanhduc]

@lookthatdog i guess you ran your code on some device other than 0 right? And that GPU is of different type from GPU0?

[rusty1s]

Eventually, this will help.

[justanhduc]

@rusty1s yes, or in my case I specified the architectures for the gencode flag of nvcc.