Author: Christophe Theron
Date: 00:44:49 05/19/00
Go up one level in this thread
On May 18, 2000 at 22:47:39, Robert Hyatt wrote:
>On May 18, 2000 at 16:50:05, Christophe Theron wrote:
>
>>On May 18, 2000 at 00:16:29, Dave Gomboc wrote:
>>
>>>On May 17, 2000 at 23:26:52, Christophe Theron wrote:
>>>
>>>>On May 17, 2000 at 17:29:17, Dave Gomboc wrote:
>>>>
>>>>>On May 15, 2000 at 15:31:51, Christophe Theron wrote:
>>>>>
>>>>>>On May 15, 2000 at 05:19:55, Jason Williamson wrote:
>>>>>>
>>>>>>>On May 14, 2000 at 22:39:26, Dave Gomboc wrote:
>>>>>>>
>>>>>>>>On May 13, 2000 at 17:16:34, Christophe Theron wrote:
>>>>>>>>
>>>>>>>>>On May 13, 2000 at 06:23:15, Dave Gomboc wrote:
>>>>>>>>>
>>>>>>>>>>On May 13, 2000 at 05:24:37, Graham Laight wrote:
>>>>>>>>>>
>>>>>>>>>>>Let's take a look at the points in favour of Crafty in the year 2000:
>>>>>>>>>>>
>>>>>>>>>>>* Popular opinion is that Crafty has recently made big improvements
>>>>>>>>>>>
>>>>>>>>>>>* The author (Bob) is the world's most experienced parallel processor chess
>>>>>>>>>>>programmer
>>>>>>>>>>>
>>>>>>>>>>>* Only recently have people started writing parallel processor programs for PCs
>>>>>>>>>>>
>>>>>>>>>>>* Bob is now tuning his program to play at tournament time controls
>>>>>>>>>>>
>>>>>>>>>>>2000 could be Crafty's best chance ever...
>>>>>>>>>>
>>>>>>>>>>I think Crafty's best chance was last year, when Bob could have brought a big
>>>>>>>>>>Alpha before most people were doing SMP at all.
>>>>>>>>>>
>>>>>>>>>>Dave
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>Which is not a very kind remark for Crafty, if you think about what you are
>>>>>>>>>implying... :)
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Christophe
>>>>>>>>
>>>>>>>>I'm implying that Bob had the chance for a bigger hardware edge at WCCC 1999
>>>>>>>>than I think he will have in the next WCCC, which IMO would have increased the
>>>>>>>>odds of him winning. I'm not sure what you read into my statement, but
>>>>>>>>basically I'm just saying that I believe that a factor of ten in hardware is
>>>>>>>>still reasonably important.
>>>>>>>>
>>>>>>>>Dave
>>>>>>>Seems to me that Bob will reassert his hardware edge with the beuwolf cluster.
>>>>>>>:) Not to many of you guys will be able to put together a beast like that.
>>>>>>
>>>>>>
>>>>>>
>>>>>>Makes me wonder if we play the same game. I'm not sure anymore.
>>>>>>
>>>>>>
>>>>>> Christophe
>>>>>
>>>>><shrug> Bob's research work is in parallel and distributed algorithms.
>>>>>Computer chess is a high-performance application that is entertaining to program
>>>>>and serves as a good testbed for ideas. Other computer chess software
>>>>>developers (such as yourself) come from different perspectives, so I suppose
>>>>>it's natural for you to have that sort of question.
>>>>>
>>>>>Dave
>>>>
>>>>
>>>>
>>>>Right. I find it more interesting to fight in order to take the best possible
>>>>advantage of the hardware we currently have than trying to be the first one to
>>>>make the best use of the hardware we will (maybe) have in 5 years (or more, or
>>>>never).
>>>>
>>>>I'm a software developper. I fight with my software, not with my hardware.
>>>>
>>>>I guess Bob's approach is as useful as mine anyway, but it's clear we have
>>>>different philosophies about life.
>>>>
>>>> Christophe
>>>
>>>Okay, but the "hardware [you] will (maybe) have in 5 years" is the "hardware
>>>[Bob] currently [has]". ;-)
>>>
>>>Dave
>>
>>
>>
>>No. SMP computers will never be the mainstream. I'm only interested in the
>>mainstream. The kind of computer everybody has.
>>
>>The kind of computer that _some_ guy can afford to build in _some_ university
>>does not interest me.
>>
>>The people saying that SMP is the future for everyday computers do not
>>understand the kind of revolution the computer industry is going into.
>>
>>
>> Christophe
>
>
>SMP computers _are_ mainstream. In another 3-5 years, every microprocessor will
>likely have two cpus on one die. But for today, Asus, AMI, several others, are
>all selling dual cpu machines like hotcakes. Not as many as the basic single-
>cpu machine, of course. But tens of thousands are being built/sold daily by
>a couple of pretty good companies.
Very limited use. Most people don't want/don't need/will never own one.
>I have been polling my undergraduate classes for a couple of years. In _every_
>class I teach, when I ask the question "does anyone here have a multiple-cpu
>machine?" the smallest number of "yes" responses has been 2. Out of 20. The
>most was 7 out of 45. These are students. Albiet CS students...
That is only a small non-representative fraction of the people using computers.
Anyway, we both have very different opinions about what the typical computer is
(and will be)... We both know that already, isn't it?
SMP will be useful for some applications, for some people. I don't deny it. I
don't deny that improving SMP and SMP programming is useful.
I might have a SMP version of Tiger in the future.
I just deny SMP computers will be the mainstream computers in the foreseeable
future.
Christophe
This page took 0.02 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.