I wonder that why does MASM32lib's memfill not use 'rep stos(?)' ? Which is fast? Thank you for reading and have a nice day! :-)
Posted on 2001-05-07 21:27:00 by nyam
nyam, The answer is simple, the unrolled loop in memfill is a lot faster than REP STOS(?). This will vary some from one processor to another but generally the older string instructions are not internally well optimised in later Intel processors. Regards, hutch@pbq.com.au
Posted on 2001-05-07 21:41:00 by hutch--
Thank you very much for your great reply! :) And, thank you for making memfill, Hutch. :cool:
Posted on 2001-05-09 11:39:00 by nyam