Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The new Celeron-A

Author: Robert Hyatt

Date: 20:56:09 08/25/98

Go up one level in this thread


On August 25, 1998 at 22:21:15, Bela Andrew Evans wrote:

>Anybody who wants to get a powerful chip for their chess programs
>might want to consider Intel's new Celeron-A chip.  Their are two
>versions, 300 and 333 mhz, and reportedly it is quite easy to
>overclock the 300 mhz version to 450 mhz.  The clock multiplier on
>the 300 mhz chip is locked to 4.5, but if you put the chip into a
>100 mhz bus motherboard, you get 4.5 X 100 = 450 mhz.  Benchmarks
>show the speed of an overclocked Celeron-A 300 mhz comparable to
>a Pentium II 450 mhz, yet it costs about $500 less.  You have to
>wonder what Intel is thinking??  The competition in the chip
>industry is getting really wild....
>
>Bela Evans


without more information, overclocking the celeron-A chip might be very
dangerous.  The original celeron had no L2 cache, which has been the
bottleneck  for all the chips since the first P6 was released.  And this
is why the PII's have been running the cache at clock/2 speeds.  The orig
celeron had no L2, which meant that overclocking was quite likely to work.

But the -A celerons have a L2 cache now, at full core cpu speed (it is
only 128kb, but that is oh-so-much-better than 0k).  The question to be
determined is can this cache be overclocked.  That was the main killer on
overclocking a P6/200, because the cache became unreliable.  What is unknown
is if the celeron-A has the same cache design as the xeon, which also runs
the cache at core clock speed.  If so, overclocking should probably work.
If they cut corners on the cache, overclocking will lead to bizarre bugs...



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.