Date: Wed, 17 Dec 1997 13:18:44 -0800 From: Eric Hedstrom <erich@compecon.com> To: Dave Bodenstab <imdave@mcs.net> Cc: questions@freebsd.org Subject: Re: Cyrex floating point problem? Message-ID: <349841B4.156DB895@compecon.com> References: <199712172028.OAA18268@base486.home.org>
next in thread | previous in thread | raw e-mail | index | archive | help
once upon a time i had a similar problem that turned out to be because of the cpu overheating. everything would work except for povray (or presumably other FP-intensive programs), which would die after a minute or two. turned out the fan on the cpu heat sink had died. i replaced the fan, and it worked like a champ. this might not be the cause of your problem, but it's worth checking. eric hedstrom erich@compecon.com Dave Bodenstab wrote: > > I've got a new pc (built from used parts) and it has a Cyrix cpu. > Everything is working (dos/windows/fbsd 2.2.5) with one glaring > problem. When I run a floating point intensive program such as > povray, I get intermittent SIGFPE core dumps. The same input, > the same program, etc. sometimes works, and other times core > dumps. > > If I just run pov and don't touch the machine until it finishes, > my success rate increases. > > My guess is that there's some problem with context switching and > saving/restoring the FP state. Or else there's some problem > handling masked FP exceptions... > > Any ideas? > > Here are snippets from the boot probes and kernel config: > > /var/log/messages: > Dec 17 02:09:20 base586 /kernel.new: Copyright (c) 1992-1997 FreeBSD Inc. > Dec 17 02:09:20 base586 /kernel.new: Copyright (c) 1982, 1986, 1989, 1991, 1993 > Dec 17 02:09:20 base586 /kernel.new: The Regents of the University of California. All rights reserved. > Dec 17 02:09:20 base586 /kernel.new: > Dec 17 02:09:20 base586 /kernel.new: FreeBSD 2.2.5-RELEASE #4: Wed Dec 17 01:44:04 CST 1997 > Dec 17 02:09:20 base586 /kernel.new: bin@base586.home.org:/usr/src/sys/compile/base586 > Dec 17 02:09:20 base586 /kernel.new: CPU: Cyrix 6x86 (486-class CPU) > Dec 17 02:09:20 base586 /kernel.new: Origin = "CyrixInstead" DIR=0x1631 Stepping=1 Revision=6 > Dec 17 02:09:21 base586 /kernel.new: CPU cache: write-through mode > Dec 17 02:09:21 base586 /kernel.new: real memory = 33554432 (32768K bytes) > Dec 17 02:09:21 base586 /kernel.new: avail memory = 30474240 (29760K bytes) > Dec 17 02:09:21 base586 /kernel.new: Probing for devices on PCI bus 0: > Dec 17 02:09:21 base586 /kernel.new: chip0 <Intel 82437VX PCI cache memory controller> rev 2 on pci0:0 > Dec 17 02:09:21 base586 /kernel.new: chip1 <Intel 82371SB PCI-ISA bridge> rev 1 on pci0:7:0 > Dec 17 02:09:21 base586 /kernel.new: chip2 <Intel 82371SB IDE interface> rev 0 on pci0:7:1 > Dec 17 02:09:21 base586 /kernel.new: vga0 <VGA-compatible display device> rev 1 int a irq ?? on pci0:10 > Dec 17 02:09:21 base586 /kernel.new: Probing for devices on the ISA bus: > Dec 17 02:09:22 base586 /kernel.new: npx0 on motherboard > Dec 17 02:09:22 base586 /kernel.new: npx0: INT 16 interface > > kernel config: > machine "i386" > cpu "I486_CPU" > options "AUTO_EOI_1" > options BOUNCE_BUFFERS > device npx0 at isa? port "IO_NPX" iosiz 0x0 flags 0x0 irq 13 vector npxintr > > The comments in LINT about the npx flags imply that they do not apply > to me since I do not have "I586_CPU" as an option. > > Dave Bodenstab > imdave@mcs.net
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?349841B4.156DB895>