Filippo Barbari

Hello again, sorry for the late reply. I am posting this here instead of opening a new issue because I think it may be related. I tried [this simple dot...

> you may be interested in an existing implementation of dot product in hwy/contrib/dot, which includes unrolling. Thank you very much, I never checked the contrib sections. > That can...

Hello again, sorry for the delay. > would you be able to gather some evidence about the speedup vs runtime init for example by running on Graviton3 (SVE_256)? Sure. I...

> Could you please share the name of the reference BLAS implementation you are using for these tests? Cmake tells me this: ``` ... -- Found SystemBLAS: BLAS_LIBRARIES ... ```...