DO NOT REPLY. BORROWING ACCOUNT. REPLY TO fmgst+@pitt.edu. THANKS
Greetings. I'm new to Linux (in fact, today was the first time I
got the boot & root disks to work... WOW! :-) but have a possible
solution to some peoples' problems.
A few years back I got a 80486/25 with 8 megs of ram, a 200MB
IDE, and a ProDesigner II VGA board and monitor. All was dandy
except that MS-Windows 3.0 would *reboot* my machine once in a
while. Support had no idea what the heck was wrong - and they
had the pleasure of playing with my motherboard twice :-) and
they finally sent me a new motherboard (but put the old RAM from
the original: hint-hint).
Well, I just stayed away from Windows utill I got 3.1. Just
after the installation I got the message I was waiting for -
"Partity error. System Halted" At least I now knew what the
hell was going on. After fooling with the 8 SIMMs I found one
that would crash Windows:
My test consisted of running 16 tasks of "Shaker" which
comes with MicroSoft C SDK and "shakes" memory around
to see if your program uses malloc() or the new one
that you're supposed to use - basicly ABUSES all the
RAM in the system.
Well, it turned out to be one darn 1MB SIMM that was the
problem. Again, the problem ONLY and I mean ONLY showed up
in Windows - not even extended memory testers or other system
"burn-in" programs (like CPU24.ZIP on 135.252.135.4)
Is your Linux crashing NOW-AND-THEN? Is your MS-WINDOWS also
crashing now-and-then? CHECK THAT RAM!!!
By the way, after pulling the "offending" SIMM, I noticed that
it's the "old" 9 chip design. I'm not sure it it's just shoddy
assembly (probably, since the contacts are only tinned not gold-
plated) but some of the chips were DARN close to each other - the
spacing was not even. Some leads ALMOST touched at room-temp and
I can only wonder what they did after being "shaken".... ;-)
So, next time you're upgrading RAM, get the new 3-chip SIMMs and
(with the prices being what they are) get faster ones.
--
________________________________________________________________________________
________________________________________________________________________________
_______________________karol@unix.cis.pitt.edu__________________________________
________________________________________________________________________________| Chuck Ebbert | Why do so many machines need "noapic"? |
| Greg Kroah-Hartman | [PATCH 004/196] Chinese: add translation of SubmittingPatches |
| Bart Van Assche | Integration of SCST in the mainstream Linux kernel |
| Jeremy Allison | Re: [RFC] Heads up on sys_fallocate() |
git: | |
| Corey Minyard | [PATCH 3/3] Convert the UDP hash lock to RCU |
| David Miller | [GIT]: Networking |
| Denys Fedoryshchenko | packetloss, on e1000e worse than r8169? |
| Gerrit Renker | [PATCH 27/37] dccp: Integration of dynamic feature activation - part 2 (server side) |
