In the past when writing extensions using THC one would use a macro THCUNN_assertSameGPU(state, X, Y) to verify if tensors X and Y reside on the same GPU.
Is there an equivalent to this if X and Y are given by torch::Tensor ?
In the past when writing extensions using THC one would use a macro THCUNN_assertSameGPU(state, X, Y) to verify if tensors X and Y reside on the same GPU.
Is there an equivalent to this if X and Y are given by torch::Tensor ?