Author: Chris Carson
Date: 08:45:52 06/18/02
Go up one level in this thread
On June 18, 2002 at 11:20:13, pavel wrote: > >>> >>> >>>why bother, when you can use a windows OS on these platforms? >>>besides it is clear that the curent fastest 64bit processor is not faster than >>>the current fastest 32 bit. >> >>This has been discussed in other threads. The 64 bit processor may be able to >>do more "useful work" in a single clock cycle than the current 32 bit processors >>can do, depending on the design of the processor. Clock rate is not a true >>indication of the amount of useful work that can be done in a given amount of >>time. In other words, a 64 bit processor, if properly utilized by the >>programmer, may be preferable over a 32 bit processor with a faster clock rate. >>We have different KINDS of processors to consider, too. >> > >I know that, but still the fastest available 64bit processor with the fastest >available proccessor speed (perhaps 1ghz) is not yet faster than the fastest >available 32bit processor at the fastest available clock speed (2.5ghz by intel >or 2100+ by AMD). > >That is also aplicable for chess programs. >so there is no point of using 64bit processor because it is not yet faster, let >alone the other complications. > > >>> >>>Even if they want to use 64bit, they can use windows, no need to get UNIX, or >>>port it on UNIX. >> >>Again, "discussed already in other threads." Whether or not there is a need >>depends on availability of the necessary hardware, including boards. The >>required availability is in doubt. It is not clear to me that the necessary >>hardware is available for a Windows implementation. > >Yes there is a windows2000 64bit version already available. > > >> >>Someone mentioned a specific number of nodes per second. [6,000,000 per sec?] > >that is the amount of speed Chessbase is expecting, there is no scientific basis >on this, since, as you mentioned, they havent yet decided what processor they >will be using. > Here is the link: http://www.brainsinbahrain.com/about/ I think you are right, it is only an estimate and it was for the Oct 2001 date, the new date is 02Oct, 2002 and the hw might be faster than for the original oct 2001 date. I hope they will give us a Fritzmark and node/sec count either before or after the match. > >>If that's true, then choice of the TYPE of processor implies a specific clock >>rate. In other words, the implied clock rate would depend on the processor. >>Choice of a "better" processor might offer some relief on required clock rate. >> >>[Forgive me for saying the same thing several different ways. That's just the >>way I think. Looking at several different "flavors" of the same thought.] >>> >>>pavs > > >personally I believe there is no way chessbase will use fritz under one of those >64bit processors. > >1) because they still not be fast enough >2) the amount of work needed to port it for that platform to take the advantage >of the processor. It is not clear if it is easy to port Fritz for 64bit >processor since AFAIK it is written purely on assembly, and i have not much >knowledge about that language. Also 32bit program wont do much on a 64bit >processor if you dont port your program for that type of processor to take >advantage of the architecture of the processor. > > >cheers, >pavs
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.