Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Some Crafty 16.19 results on my XP 2.44GHz

Author: Matt Taylor

Date: 11:46:31 02/20/03

Go up one level in this thread


On February 20, 2003 at 11:50:31, Robert Hyatt wrote:

>On February 20, 2003 at 09:18:35, Jeremiah Penery wrote:
>
<snip>
>>It's not all that difficult to 'debug' any problems with overclocking.  Run a
>>program like Prime95 for a while (catches errors MUCH faster than anything else
>>I've seen).  If you get errors, clock down and see if they go away.  Obviously,
>>errors are a sign of an unsafe overclock.  It certainly doesn't mean that EVERY
>>overclock is unsafe.  Otherwise, you'd better send back those 2.8GHz Xeons right
>>away, because they're nothing more than overclocked 2.5GHz ones (in a manner of
>>speaking).
>
>The problem is that the 'error signs' are not like red flags.  It isn't easy to
>detect a single bit
>error in a floating point multiply, for example, if it only happens once every
>10^6 multiplies
>with specific bit patterns.  And that kind of failure has happened _many_
>times...

Yes. In two different cases I had bad ram (and did not realize it) because signs
were vague and intermittant. Though ram is not the same as a CPU, I think these
two examples clearly demonstrate how errors can go undetected. The ram was
running at rated speed, I think. I'm not sure because my motherboard's default
memory timings were too aggressive, and I had to go into the BIOS to change the
CAS latency. (There may have been other timings that were too aggressive as
well.)

In the first case, I ran several months with only a few warnings. The only
indicator I had was the java compiler crashing NT. After that, I didn't have any
problems for another month or two. All of a sudden the system became horribly
unstable. I removed the good DIMM and the other failed POST.

In the second case, I did not experience instability at all for several months.
I was on a coding spree at 4 AM. I recompiled my program, and the compiler
bombed out with an error. There were invalid characters in some runtime header
file. I thought that was strange, but I replaced the header file with a copy
from another machine and kept on working. About 30 minutes later, the compiler
started crashing. I replaced the compiler's executable and things went smoothly
again. The next day I put together my dual-CPU system and used the same ram, and
everything ran fine. I concluded that it must have been a motherboard or CPU
problem (the dual wasn't crashing or doing anything weird). About a month later
I ran the AMI diagnostics program, and it turned out it really was the ram.
After replacing the ram, I've used that same system for months as a router at
home. In a twist of irony, I loaned it recently to a friend whose 1.8 GHz
Pentium 4 broke. (Or so we thought.)

-Matt



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.