From owner-freebsd-questions Thu Jul 1 15: 7: 2 1999 Delivered-To: freebsd-questions@freebsd.org Received: from mailhub.scl.ameslab.gov (mailhub.scl.ameslab.gov [147.155.137.127]) by hub.freebsd.org (Postfix) with ESMTP id 57F9415723 for ; Thu, 1 Jul 1999 15:06:58 -0700 (PDT) (envelope-from ghelmer@scl.ameslab.gov) Received: from demios.ether.scl.ameslab.gov ([147.155.137.54]) by mailhub.scl.ameslab.gov with esmtp (Exim 1.90 #1) id 10zozl-00055N-00; Thu, 1 Jul 1999 17:07:49 -0500 Date: Thu, 1 Jul 1999 17:06:50 -0500 From: Guy Helmer To: "M. L. Dodson" Cc: Arcady Genkin , freebsd-questions@FreeBSD.ORG Subject: Re: FreeBSD is painfully slow on my 486 In-Reply-To: <199907012110.QAA86650@beowulf.utmb.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Thu, 1 Jul 1999, M. L. Dodson wrote: > Arcady Genkin writes: > > I'm quite desperate by now -- I dumped Linux for FreeBSD on an i486 > > that I used as a firewall, and FreeBSD is much slower. I mean *really* > > slow. The 486 is DX4 and works at 100MHz. It has 16M or RAM. Not > > exactly a screamer, but it is fast enough for a machine with no X > > installed. > > > > I compiled a custom kernel, disabling pretty much everything. > > > > I suspected that "Turbo" could have been turned off, but Linux kernel > > from a rescue disk reports 49 bogomips, which seems to be > > reasonable. I also know that the disk access is slower because fs's > > are mounted syncronously, but it shouldn't be *that* slower. > > > > I'll give you a couple of examples: kernel compilation takes 4 hours, > > whereas somebody on this list reported that his similar 486 takes 30 > > minutes to compile a kernel. Midnight commander takes 7 seconds to > > start, and I have to wait for 7-8 seconds for its file viewer to open > > a file. > > > > I hope somebody can help me determine whether FreeBSD doesn't support > > something on my computer (for example, the chipset is ALI1429 -- Linux > > had a special option for this chipset in kernel config). Perhaps I > > should throw in the towel. Or is there still hope for me? > > ... > Seriously, I had something similar happen to me. I would guess > it is your motherboard. My situation: 486 with 8MB and VLB SCSI > disk system (this was a long time ago). Measure speed with > bonnie. Add memory to bring to 16MB. Measure speed with bonnie. > Observe that bonnie showed the system to be slower than the 386SX > with 6MB and an IDE disk system, and the same version of FBSD, > sitting right next to the 486. The slowdown was easily > perceptible sitting at the machine. I posted all over; several > people said they had observed similar phenomena. I dumped it for > a Pentium. The motherboard was one of those that used some sort > of phase locked loop to generate the clock (no crystal on the > board). My speculation was that the additional load of the extra > (30 pin) SIMMS caused it to clock at some lower multiple of the > 25MHz it was supposed to be running at. (Don't flame me; I said > it was a speculation ;-) I noticed a severe slowdown on some 486's after adding more memory. In particular, one 486 with 16MB RAM running FreeBSD 2.x slowed to crawl when I added 4MB of RAM. I seem to remember a similar slowdown when going from 16MB to 32MB of RAM on a similar 486 running Netware 3.11. In both cases, I suspect that the level two cache implemented on the motherboard was only able to cache the lower 16MB of RAM. So, I tend to suspect the cache in your case. Is the 486 an Intel, or another brand? There are a couple of Cyrix kernel config options that have to do with cache (see the LINT config file). Guy Guy Helmer, Ph.D. Candidate, Iowa State University Dept. of Computer Science Research Assistant, Ames Laboratory --- ghelmer@scl.ameslab.gov Research Assistant, Dept. of Computer Science --- ghelmer@cs.iastate.edu http://www.cs.iastate.edu/~ghelmer To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message