Hello to everybody,

I’m new at PyTorch and I have a question about moving the computations on GPU.

Suppose we have two different operations that involves different tensors, for example something like this:

z1 = x * y # first operation

z2 = x + y # second operation

where x and y are two tensors in the same device.

My question is: will the GPU calculate both z1 and z2 in parallel?

If not, how can I force the GPU to performs these two operations at the same time?

Thanks in advance!