movntq
category: general [glöplog]
Hi, today I've instaled service packs and the proccessor pack to my VC6 (without it, movntq is not supported in the inline asm). I've tried for first time the movntq in an AthlonXP 2800+ for big arrays (wich doesn't enter in the caches) with some 333 or 400 mhz ddr ram (I'm not sure of the speed of my ram). Changing my movs by movntqs the speed improvements are just impressive. My question is, do you know if it is the same for Intel proccessors? And, how faster use to be changing from one to other opcode? Do you know where to find info about this (I mean, speed improvements in both athlon and intel processors)?
