home *** CD-ROM | disk | FTP | other *** search
- Path: nntp.teleport.com!sschaem
- From: sschaem@teleport.com (Stephan Schaem)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: Free Cycles
- Date: 21 Feb 1996 18:48:23 GMT
- Organization: Teleport - Portland's Public Access (503) 220-1016
- Message-ID: <4gfpdn$elr@maureen.teleport.com>
- References: <1996Feb15.170809.6386@imada.ou.dk> <1139.6624T1309T1361@teclink.net>
- NNTP-Posting-Host: linda.teleport.com
- X-Newsreader: TIN [version 1.2 PL2]
-
- rad (rad@teclink.net) wrote:
-
- : Bjorn Reese <breese@imada.ou.dk> wrote:
- : >tim bscke (azure@people-x.people.de) wrote:
- : >> move.l (fast)+,(chip)+ ;18 cylces
- : >> clr.l (chip)+ ;24(!!!) cycles
- : >[...]
- : >> Why ?? This seems to be really strange to me.. esp. the clr.l..
-
- : >If my memory haven't betrayed me, clr reads the destination before
- : >clearing it. So a clr makes two chip accesses. This behaviour is
- : >also the reason why you should never use clr to clear a write-only
- : >register.
-
- : You are correct for the 68000. This is a good reason to stick with the 68010+
- : (if not the 68020+) where the clr instruction doesn't have this drawback.
- : Considering that the 68000's are only used in unaccelerated 500s, 600s and
- : 2000s I don't think that's really problem.
-
- clr.l (a0)+ take 6 cycle on a 030/020 , a move.l dx,(a0)+ take 4...
- dunno on the 040 and up... but using clr is not recomanded for speed.
- Its nice when you dont have a reqister free or set to zero, otherwise
- avoid.
-
- Stephan
-