Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!ucbvax!pasteur!helios.ee.lbl.gov!nosc!manta!dwi
From: dwi@manta.NOSC.MIL (Steve Stamper)
Newsgroups: comp.sys.amiga.tech
Subject: Re: 2620, how fast is it REALLY????
Message-ID: <829@manta.NOSC.MIL>
Date: 2 Jun 89 15:01:38 GMT
References: <25@iisat.UUCP>
Reply-To: dwi@manta.nosc.mil.UUCP (Steve Stamper)
Distribution: na
Organization: Naval Ocean Systems Center, San Diego
Lines: 23

I have gotten the following using Lattice C 5.02 and Drystone benchmark

Compiling lc -v -w -cf -L -O -m0 -rr drystone using Amiga 2000 /68000 w/
expansion RAM : 
Registers - 1923 drystones/sec
No Regs - 1666 drystones/sec
I did a changetaskpri 100 before I ran it and had to modify the linker file
to also include lcs.lib since time() don't work with only lcsr.lib

Using the same compilation but -m2 and the A2620 board enabled I get

Registers - 5000 drystones/sec
No regs - 4166 drystones/sec

Some other benchmarks I got are (with A2620 enabled)

fib 52 secs
qsort 11 secs
savage (25000) 4 secs
100 sieves 5 secs
Whetstones - 769231 Whetstones/sec

-Roger Uzun