Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Refuting nonsense about 64 bits servers/supercomputer chips

Author: Tom Kerrigan

Date: 15:36:44 09/03/01

Go up one level in this thread


On September 03, 2001 at 12:06:26, Vincent Diepeveen wrote:

>On September 02, 2001 at 17:26:31, Tom Kerrigan wrote:
>
>>On September 01, 2001 at 08:40:46, Vincent Diepeveen wrote:
>[snip]
>
>>>completely (especially price range), as well from the
>>>servers from alpha and especially from SUN and all other cpu
>>>brands when running 32 bits (computerchess) software,
>>
>>Seems like lots of computer chess programs are 64 bit now, and the Alpha can
>>more than hold its own against any PC processor running either 32 or 64 bit
>>code.
>
>This is what i mean, an alpha 2 processor is around $20000 in the
>store now. that's for a 833Mhz dual alpha.

In the post I was replying to, it sounded like you were referring to Intels and
possibly 64-bit CPUs in general. You didn't say anything about an 833MHz dual
Alpha specifically. Who cares what the price of a dual 833 is? Was that ever in
question?

>A dual 1.2Ghz K7 of course blows that away for DIEP.
>
>An alpha 633Mhz (21164) used to be for diep around the same
>speed like a PII at 380Mhz.

The 21164 sucked at any code that wasn't purely 64-bit. If you had reworked your
program to remove these bottlenecks, it would have been much faster. Nobody
cares about the 21164 anymore anyway.

>21264 has about the same number of registers as a K7,

Huh? I hope you're talking about rename registers or something, because there's
a factor of 4 difference in programmer visible registers (in the Alpha's favor).

>and the 21264 has huge costs for branch mispredictions.

What?? The 21264's pipeline is 7 stages. The Athlon's is 10. That means the
Alpha has a much smaller penalty for a mispredicted branch. Moreover, it's
widely accepted that the Alpha's BPU is better than the Athlon's. (The Athlon's
BPU is a pretty sore point--AMD fans have been hoping for a redesign for years.)

>The 4 instructions a clock it sometimes can do hardly make up
>for the lack of 400Mhz in speed.

The maximum number of instructions a chip can retire per clock doesn't have
anything to do with anything. The Athlon can retire 6 (or 8? counting FP) per
clock but that doesn't make it faster.

Maybe if you took some classes in computer architecture or at least read a few
hardware web sites, you wouldn't have to subject everybody to your seemingly
endless supply of misinformation.

-Tom



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.