>You could do even more clever trick, something like
>
> if (!current->used_math) {
> unlazy_fpu(current);
> mmx_checksum();
> } else normal_checksum();
>
>is pretty easy to do, yet it should work _very_ well.
Hmm, it looks ok to me.
I would be still more convinced if the mmx routine would run some bit
faster. The numbers was something like 60 unit of time for mmx and 63
units of time for the current code + my patch. It would be interesting
also to implement a copy-user mmx (as suggested some seconds ago from
Artur) to see if using movq instead of movl would improve performances
some bit more.
Andrea Arcangeli
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/