Matrix-Matrix multiply source code

[tom (Thomas V)]

In the new thread I’m chasing things through the PyTorch code for bmm:

Matrix multiplication source code?

I’m interested in finding out some specific implementation details of matrix multiplication in PyTorch. Where would one find the source code (CPU implementation and CUDA kernel) for PyTorch’s implementation of matrix multiplication? Specifically, where would one find the code implementing torch.bmm()? I see that this question was already asked here, but not answered. Thanks!