Path: utzoo!attcan!uunet!lll-winken!ncis.llnl.gov!helios.ee.lbl.gov!nosc!ucsd!ucbvax!bloom-beacon!mit-eddie!uw-beaver!rice!sun-spots-request
From: barnett@steinmetz.ge.com (Bruce Barnett)
Newsgroups: comp.sys.sun
Subject: Sun 4/110 without FPA is slower than a 3/50 without MC68881
Message-ID: <6586@vdsvax.steinmetz.ge.com>
Date: 26 Jan 89 06:30:53 GMT
Sender: usenet@rice.edu
Organization: GE Corp. R & D, Schenectady, NY
Lines: 45
Approved: Sun-Spots@rice.edu
Original-Date: 19 Jan 89 15:08:51 GMT
X-Sun-Spots-Digest: Volume 7, Issue 122, message 1 of 11

I just did a benchmark on a Sun 4/110 without the FPA option.
The benchmark is:
__________

main() {
        int i = 100000;
        double f = 0.0;
        while (i--) f += 1.0;
        printf("f=%g\n",f);
}
__________

Compiled on a 4.0.1 system using cc -O4

Results: (using csh's time command:)

Sun 4/280 		0.2u 0.1s 0:00 70% 0+128k 0+0io 0pf+0w
Sun 3/50 -f68881	0.4u 0.1s 0:00 106% 0+72k 0+0io 0pf+0w
Sun 3/50 -fsoft		5.7u 0.2s 0:06 96% 0+72k 0+0io 1pf+0w
Sun 4/110 - No FPA	0.8u 45.0s 0:46 99% 0+136k 0+0io 0pf+0w

I assume this is due to:

	Compiler always assumes the FPA chip is *ALWAYS* there.
	FP calls are trapped by the kernal and the hardware is emulated

I also assume that Sun Marketing made a promise to engineering that Sun
4's will always have the FPA chip. And of course the engineers believed
them.

I also assume that they were soooooo embarrassed, that Marketing tried to
cover up the fact that a Sun 4/110 is *90 times slower* than a Sun 3/50
w/68881.

******FLAME ON TO SUN*******

You can't hide this sort of information, you twits.  I knew the machine
would be slow for floating point, but if I knew it would make a PC look
like a Cray in comparison, I would not have ordered this configuration.

Sheeesh!

--
	Bruce G. Barnett 	barnett@ge-crd.ARPA, barnett@steinmetz.ge.com
				uunet!steinmetz!barnett