home *** CD-ROM | disk | FTP | other *** search
- Path: phoenix.owl.de!not-for-mail
- Newsgroups: comp.sys.amiga.programmer
- References: <38232981@kone.fipnet.fi> <4ico8k$2vau@columba.udac.uu.se>
- From: "Frank Wille" <frank@phoenix.owl.de>
- Date: Tue, 19 Mar 1996 18:38:57 +0200
- Distribution: world
- X-NewsReader: IntuiNews 1.3b Beta 3 (5.11.95)
- Subject: Re: CHIP RAM speed test results
- Message-ID: <42116651@phoenix.owl.de>
- Organization: Phantasm
-
- Daniel Widenfalk wrote on 15 Mar 1996 21:42:44 GMT
- about "Re: CHIP RAM speed test results" :
-
-
- DW> Could someone with a 040 (or 060) please do a little test with the
- DW> MOVE16 instruction?
- DW>
- DW> move.l #1023,d0
- DW> .loop:
- DW> move16 (a0)+,(a1)+
- DW> dbf d0,.loop
-
- I compared a copy loop with 16 MOVE16s against a loop with 64 MOVE.L
- instructions, moving a block of 128k in Fast RAM. For copying data
- in Chip RAM there's no advantage in using MOVE16.
-
- With my 68040/25MHz I'm getting the following results:
-
- move.l (a0)+,(a1)+ -> 3.99 MB/sec
- move16 (a0)+,(a1)+ -> 5.91 MB/sec
-
- So MOVE16 is 50% faster, although the A3640 doesn't support burst
- transfers.
-
-
- -- _
- _ // Frank EMail: frank@phoenix.owl.de
- \X/ IRC: Phx @ #amiga(ger)
-
-