delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp/1994/04/08/23:35:52

From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Subject: Re: memxxx(), Duff's Device, etc.
To: FIXER AT FAXCSL DOT DCRT DOT NIH DOT GOV (Chris Mr. Tangerine Man Tate)
Date: Fri, 8 Apr 1994 21:30:46 -0600 (CDT)
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu (DJGPP Mailing List)

> On a related note, I'm curious about the Intel architecture.  Specifically,
> I'd like to know:
> a) Does it have odd-address access restrictions?

No.

> b) Are accesses on longword (4 byte) boundaries faster than word bounds?

Word access on word boundary is fine.  Long access needs to be on long 
boundary.  Violating these rules adds 3 clocks for each operand access
(on a 486, from the manual).  Since each movs instruction takes 7 clocks
(on a 486 once again), it seems that the worse case of 7+3+3 would be
potentially 2 times faster than 4*7 changing to movsd from movsb.

I don't have any 386 or Pentium timings handy, or know if there are
other gotchas here.

- Raw text -


  webmaster     delorie software   privacy  
  Copyright 2019   by DJ Delorie     Updated Jul 2019