Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Comet B42 MMX P- 166 Mhz Vs CT 14.9 For Palm Ended 4.5-1.5 ?!

Author: Christophe Theron

Date: 07:15:18 04/18/02

Go up one level in this thread


On April 18, 2002 at 05:46:05, stuart taylor wrote:

>On April 17, 2002 at 19:13:38, Christophe Theron wrote:
>
>>On April 17, 2002 at 17:17:04, Ulrich Tuerke wrote:
>>
>>>On April 17, 2002 at 13:57:01, Roy Eassa wrote:
>>>
>>>>On April 17, 2002 at 13:54:13, Christophe Theron wrote:
>>>>
>>>>
>>>>>This is a very precious advice for your future hardware purchases: do not ever
>>>>>take MHz as a measure of speed. NEVER.
>>>>>
>>>>>The DragonBall inside the Palms need approx. 15 clock cycles (average) to
>>>>>complete one instruction.
>>>>>
>>>>>The Pentium needs approximately one clock cycle per instruction (actually I
>>>>>think it is 0.9 clock/instruction).
>>>>>
>>>>>The DragonBall has no L1 cache (not sure it needs one anyway).
>>>>>
>>>>>The reason for this difference is the number of transistors in each processor,
>>>>>and the power consumption.
>>>>>
>>>>>If you just count the MHz difference, you are going to be 15 times OFF the real
>>>>>number.
>>>>
>>>>
>>>>Doesn't that mean that a 16 MHz Palm is not much faster than a ONE MHz Pentium
>>>>would be??  And if so, how in heck can Chess Tiger be so strong on a 16 MHz
>>>>Palm??
>>>
>>>It's a fantastic program !
>>
>>
>>
>>Oh... Hem... (blush)  :-)
>>
>>
>>
>>    Christophe
>
>So even Pocket computer (ipac) and even ARM palm will only be like about 15
>Mhz.of a pentium? Does this follow in both of these?
>S.Taylor



Yes I think you are APPROXIMATELY right.

The iPac might be as fast as a 486dx-50, or a Pentium 25. I do not have the
exact numbers, but you got the idea.

Be very careful when you hear MHz. Without deeper knowledge of the processors,
do not compare MHz, it's a trap.

Some processors have passed the "MIPS" test. This test is a standard C program
that is executed for a fixed time. When the time is elapsed, one counts the
number of instructions executed and that gives a speed index.

Comparing MIPS numbers is much better than comparing raw MHz figures. However it
is still misleading.



    Christophe



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.