home *** CD-ROM | disk | FTP | other *** search
- Path: news.eunet.fi!newsmaster
- From: kp@direktor.voima.jkl.fi (Kari-Pekka Koljonen)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: CHIP RAM speed test results
- Date: 14 Mar 1996 11:12:54 GMT
- Message-ID: <1506.6645T16T1105@direktor.voima.jkl.fi>
- References: <1067.6635T1081T1567@direktor.voima.jkl.fi> <4hh5p0$8k6@serpens.rhein.de>
- NNTP-Posting-Host: kauhajoki.fi
- X-Newsreader: THOR 2.22 (Amiga;TCP/IP) *UNREGISTERED*
-
-
- >ChipRAM is the same speed, but chip bus bandwidth is not. Try the following
- >code:
-
- >loop:
- > movem.l (a0)+,d3-d7/a2-a5
- > movem.l d3-d7/a2-a5,(a1)
- > lea 32(a1),a1
- > dbf d0,loop
-
- >If that's not faster then your accelerator card has a lousy bus interface.
-
- > Michael van Elst
-
-
- My accelerator card does suck. Movem.l loop was SLOWER than move.l loop.
-
- Here're the busspeedtest results for everyone to laugh at:
-
-
- BusSpeedTest 0.07 (mlelstv) Buffer: 16384 Bytes
- ==================================================
- loop overhead: 7.9ns
- register move: 72.6ns
- memtype op cycle bandwidth
- fast readw 182.0ns 11.0MByte/s
- fast readl 216.5ns 18.5MByte/s
- fast readm 243.9ns 16.4MByte/s
- fast writew 216.1ns 9.3MByte/s
- fast writel 217.5ns 18.4MByte/s
- fast writem 198.6ns 20.1MByte/s
- chip readw 1065.0ns 1.9MByte/s
- chip readl 1066.6ns 3.8MByte/s
- chip readm 1097.3ns 3.6MByte/s
- chip writew 1061.8ns 1.9MByte/s
- chip writel 1064.1ns 3.8MByte/s
- chip writem 1064.1ns 3.8MByte/s
- rom readw 328.3ns 6.1MByte/s
- rom readl 511.0ns 7.8MByte/s
- rom readm 488.3ns 8.2MByte/s
-
-
-
-
- ***************************************************
- K-P Koljonen / Hippopotamus Design / iNSANE
- kp@direktor.voima.jkl.fi, kpkoljon@freenet.hut.fi
- http://direktor.voima.jkl.fi/~kp K-P@IRC
- ***************************************************
-
-
-