If you're using GCC, you can use this drop-in lvm.c replacement for one or two meagre % of overall execution speed improvement (hey, it's free; no discernable net footprint increase), as benchmarked with fib.lua... I was looking into using GCC's get-address-of-label thing to embed the instruction's label's address into the Instruction itself, but I really really don't think it's going to be worth the bother, even if it does turn out to be a measurable win, which I'm not sure of at all. So we're left with this (and a couple of minor micro-optimisations), which at least is a no-brainer. Regards, --Adam -- Adam D. Moss . ,,^^ adam@gimp.org http://www.foxbox.org/ co:3 Never attribute to affection that which can be adequately explained by loneliness.
Attachment:
lvm.c.gz
Description: GNU Zip compressed data