On 1/16/24 04:57, Daniel Henrique Barboza wrote:
I am now wondering if this is worth the trouble, and we should instead do:+ int vlmax = cpu->cfg.vlenb >> (s->sew - s->lmul); Like we're already doing in patch 9. Patch 12 would be a similar case.
This is more like what I expected to see. r~
