Closed htyu closed 5 months ago
@htyu Why do you need to add matrix_instr_nonkdim
to the autotuner? Is adding it here like waves_per_eu
not enough?
@htyu Why do you need to add
matrix_instr_nonkdim
to the autotuner? Is adding it here likewaves_per_eu
not enough?
Oh, I didn't know about waves_per_eu
. What does it do? We basically want to use the 16x16 mfma instruction instead of the default 32x32 one. Can it be achieved by tuning waves_per_eu
?
@htyu waves_per_eu
has nothing to do with matrix_instr_nonkdim
. I referred to it because you can use matrix_instr_nonkdim
(add it to kernel arg and tuning) in the same way as waves_per_eu
. So there is no need to add it to the autotuner.
@htyu
waves_per_eu
has nothing to do withmatrix_instr_nonkdim
. I referred to it because you can usematrix_instr_nonkdim
(add it to kernel arg and tuning) in the same way aswaves_per_eu
. So there is no need to add it to the autotuner.
I see. That also works. Thanks!
Adding
matrix_instr_nonkdim
to the autotune configs to enable tuning with different mfma shapes.