Date: Fri, 27 Sep 1996 11:26:48 GMT+0100 From: "Kees Jan Koster" <Kees.Koster@nym.sc.philips.com> To: hardware@freebsd.org Subject: Re: lmbench results for AMD 5k86-P100 Message-ID: <5D36502BB7@NLNMG01.nym.sc.philips.com>
next in thread | raw e-mail | index | archive | help
> > > Maybe later. > > I'd appreciate it ... > Well, I'm afraid it will be a while. I'm currently in the process of moving to Canterbury (UK) to start studying for MSc. Distributed Systems. Maybe in a week or two... No need to reply to me on this mail, my account will be removed as soon as I leave Philips. (Others may still be interested in your comments.) > > > Could someone shed some light on the 'bad MHz' messages below? > [...] > > According to an article in c't magazine (currently the [...] > else (same as lmbench :) > Thanks. I'm glad it is not due to possible memory problems :-) > > > Both Unix and MS-DOS work, altough MS-DOS still feels kind'a fragile. > > I don't care for DOS, but it's a little surprising to hear, > that Unix runs better (more reliable) than DOS on some > hardware. But I could consider this fact as pro-AMD5K :) > I narrowed the problem down to one game (Discworld, by Psychnosis). They were unable to cook up an answer. Oh well, Stonekeep and DescentII run fine, those are the only things I use MS-DOS for anyway. > > Seems that the 5x86/133 (which is rated as equivalent to the > Pentium 75) is 60% of the 5k86-100 (which ought to be as fast > as a P100) in those tests, were 2.1.5 and -current are not > too different. (Ie.: ignore the 5x86's PIPE performance ... :) > I was actually considering to buy the AMD-133, but it turned out to be only about DFL. 100,= cheaper for the set I wanted. I paid about DFl. 900,= for the Exp8661, P100, 512Pb cache plus 32Mb EDO. That would have been about DFL. 700,= for the AMD-133, plus Mainboard with only ;) 256kb cache and 32Mb (non-EDO) ram. 40% extra perfomance for DLF. 200,= is less than I was willing to pay for that and I've got the advantage that this mainboard will hold bigger cpu's in the future. > > [ results snipped ] > > Except for the pipe code, the 5k86 is a factor of 5/3 as fast ... > Yes, except for the pipe code. You don't happen to be working on improving that? ;) > [ snip ] > > The 5k86 is special in that its memory read performance is much > better (by a factor of 2) than its write performance. Not sure why, > but this may also be caused by the motherboards characteristics. > (What chip-set is that, BTW ?) > I think (check www.dataexpert.com) that it's the VX chipset. I'm not sure, but it is the chipset that was repeatedly branded as the 'low-budget' version on this list. Well, it _was_ cheap. Anyway, I saw in the ctcm results for comparing mainboards that this board outperforms some 'high-end' chipset-type mainboard. Check the ctcm motherboard benchmark results for that. (From memory: www.u-net.<something>/~sysdoc/ ) > > LMBENCH for sure is not the best CPU benchmark, but it seems > the 5k86 is at least comparable to a P5-100, and seems to be > a cost effective alternative for a server that does not need > to offer peak performance values ... > Hmm. AMD claims that it offers +30% performance over a `real' P100. (Yeah, right, and santa claus uses AMD cpu's, too) Still, I'm very happy with it. Perhaps someone who has an Intel P100 could post his lmbench results, too. I'm curious. Once more, no need to reply to me, I won't be around to read it, but others might be interested. I'll read the mail-archives later ;) Groetjes, Kees Jan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5D36502BB7>