Data conversion when packing

I think we need a feature in which data conversion is done while packing. This way all architectures could get lower resolution kernels to work by converting either A or B or BOTH to a desired format and then using a kernel that uses that format.

It would be useful for SBGEMM and SHGEMM in some situations in which there is no fast hardware instructions with MADD in these formats.

Flow -

Current:

pack -> convert -> kernel

Proposed new feature:

pack_convert -> kernel

Yes, it would increase memory requirements. But for hardware that has slower conversions we are wasting NxKxM converts that could be done as NxK and MxK - which is a much smaller number.

Plus we could save the repetitive conversions that is done sometimes.

Inizia tracciando il flusso esistente pack → convert → kernel per SBGEMM e SHGEMM, quindi individua dove potrebbe inserirsi un percorso combinato pack_convert. La proposta è completa quando la conversione può avvenire durante il packing per A, B o entrambi, consentendo ai kernel a risoluzione inferiore di usare il formato desiderato ed evitando conversioni ripetitive.