Re: [computer-go] Computer Go hardware

[Vincent Diepeveen <diep@xxxxxxxxx>]

At 10:39 19-10-2004 -0700, Peter Seibel wrote:
>"Frank de Groot" <frank@xxxxxxxxxxxxxxxxx> writes:
>
>>> One problem with using all possible resources is that improving and
>>> tuning program becomes more difficult (as it is slower.)
>>
>>
>> Not at all. RAM means "Random Access Memory". Which means that any
>> random location in RAM is accessable in approx. the same time.

Not at all. The bigger the memory buffer the slower the lookup.

I have free software to measure that for you if you like to.

At opteron it's about 91 ns to get a cache line from RAM with 400MB buffer.
at itanium2 it's about 280ns, at my old dual k7 chipset it's about 400 ns,
idemdito at dual p4-xeon machines.

However if you reduce that buffer then the lookup goes a lot faster. It
halves nearly.

If you keep within L2 cache of opteron that's fastest. Just 13 cycles.
P4 and prescott are 3 times slower there or so.

>> So if you have a program that needs to lookup 1000 patterns in a
>> 10,000-pattern database or a program that needs to lookup 1000
>> patterns in a 100,000,000-pattern database, this takes *exactly* the
>> same time.
>
>Only if you know, a priori, the exact address of the 1,000 patterns
>you want to look up. And if you know that you only need a 1,000
>pattern database, no?
>
>-Peter
>
>-- 
>Peter Seibel                                      peter@xxxxxxxxxxxxxxxxx
>
>         Lisp is the red pill. -- John Fraser, comp.lang.lisp
>_______________________________________________
>computer-go mailing list
>computer-go@xxxxxxxxxxxxxxxxx
>http://www.computer-go.org/mailman/listinfo/computer-go/
>
>
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/