Adjoint of Matmul

[Alireza Moradzadeh]

The current matmul and batched_matmul implementations break the gradient flow, as an example, in line

warp/warp/types.py

Line 3177 in 79a56a9

 ctypes.c_void_p(a.ptr), 

one should change a.ptr with ctypes.c_void_p(adj_a.ptr) and beta should set to 1. In the current implementation, old gradient of a is overwritten by matmul call, similar issue holds for b, c, and d.

[Miles Macklin]

Hi @moradza, thanks for the report - @daedalus5 can you take a look?

[Nicolas Capens]

Fixed by 9ae3877