[lug] CPU tester, Disk tester

[D. Stimits]

Hugh Brown wrote:

> I recently had a machine exhibit similar behavior.  The memory was at 
> fault but memtest caught the problem.
>
> If you can, start swapping components.  I'd go for replacing the mobo/cpu.

Heat will make some of those problems happen more often as well. I've 
also had bad memory which memtest failed on, but which produced a memory 
error during kernel decompress on new installs, always at the same 
offset. Longer memtest86 (like 3 days) showed it, but only on warmer days.

In case it is memory, you might go into the BIOS and increase wait 
states. If you have adjustable cpu voltages, boost voltage by 0.05 volts 
or so.

If the problems occur using X11, be sure that hardware accelerated 
OpenGL is disabled and that screen savers are not OpenGL, direct 
hardware accel makes it harder to isolate hardware problems versus 
config problems or driver problems.

D. Stimits, stimits AT comcast DOT net

>
> Hugh
>
>
> >Hi,
> >
> >I've a Redhat 9 box that crashes roughly once an hour
> >when doing just about any trivial task.
> >And it crashes reliably every time I run makewhatis.
> >No trace is left in the log file.  It's a brand
> >new installation, off the net, so I doubt it's been hacked.
> >
> >My guess is hardware.  I've run memtest86 and
> >the memory appears OK.
> >
> >I suspect it's either the Intel Pentium, the VIA board,
> >or the disk.  Do you know of any utilities to stress test
> >Pentiums, the board, or disks?
> >
> >I've written a small disk tester program, but in userspace I can
> >only test my allocated files.  Much more useful would
> >be a bootable tester that tests EVERY block.
> >
> >Any ideas on Pentium or disk testers?
> >
> >Thanks,
> >
> >Steve
>
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> Join us on IRC: lug.boulder.co.us port=6667 channel=#colug
>