Author: Vincent Diepeveen
Date: 11:08:54 01/29/05
Go up one level in this thread
On January 29, 2005 at 11:00:24, gerold daniels wrote: >On January 29, 2005 at 10:33:04, Vincent Diepeveen wrote: > >>On January 29, 2005 at 09:48:07, gerold daniels wrote: >> >>>On January 29, 2005 at 09:09:30, Vincent Diepeveen wrote: >>> >>>>On January 29, 2005 at 08:31:27, Jason Kent wrote: >>>> >>>>>On January 29, 2005 at 08:28:11, Vincent Diepeveen wrote: >>>>> >>>>>>On January 29, 2005 at 08:20:07, Jason Kent wrote: >>>>>> >>>>>>Your machine will look like a dual machine. The OS doesn't know better than that >>>>>>it is 2 processor machine. >>>>>> >>>>>>>It looks like by the third quarter of this year, both intel and amd will be >>>>>>>selling dual cores. Are they basically handled as two processors under task >>>>>>>manager, and software? I'm guessing this is going mean that to get the most out >>>>>>>of your cpu, you will have to buy all the Deep versions. Maybe that is why SMK >>>>>>>decided to seperate the programs? >>>>>>>Jason >>>>>> >>>>>>I won't comment on SMK, nor on chessbase, but it's obvious that if the intel CEO >>>>>>says that intel will produce ONLY dual core cpu's within a year and nearly >>>>>>nothing else, that only parallel software will work for you. >>>>>> >>>>>>Paying extra for something that just uses a normal single cpu, is a very bad >>>>>>thing. It means basically you have normally spoken a crippled software program, >>>>>>as > 90% of all cpu's will be dual core. >>>>>> >>>>>>Vincent >>>>> >>>>>It sounds like there will be plenty of programming needed to fix all of the >>>>>current software to make it SMP compatible. Since I will be coming out of >>>>>school in another year or so, hopefully it will be easy to find work. :D >>>> >>>>Majority of applications that can eat any amount of cpu cycles are already >>>>parallel. >>>> >>>>Please note that search is one of the hardest to parallellize problems, because >>>>nowadays important is that your parallel search maintains the same branching >>>>factor like a single cpu search does. For example in databases parallel search >>>>is pretty trivial to make. >>> >>>good morning are the dual core better than the single processor and what about >>>the new cell chip. >> >>For chess the dual cores will be 1.8 - 2.0 better in scaling. For diep simply >>2.0 scaling nearly. >> >>The cell processor, not much is known about it. >> >>See what i wrote in aceshardware about that. >> >>The only info i could find of someone speculating what it would be is that there >>will be different versions. One cheapo version which is put in cheapo game >>consoles and won't be able to impress much. >> >>The highend version i saw someone predict 4.6ghz for it in 0.065 which i frankly >>doubt they will achieve that so easily with it. >> >>Basically if i understand well it is a special processor which has 1 main core >>with 8 vector helpers. >> >>Depending upon how much those are vectorized and how many integer instruction >>units those helpers will have, it will be a fast or slow processor for chess. >> >>Then secondly real important, perhaps most important, will be the communication >>speed between the helper cores. >> >>On paper it should get 250 gflop, that's however at 4.6Ghz which is a speed i >>seriously doubt they will get for vector processors. >> >>Let's just sit and wait for it. >> >>Knowing IBM it won't be a serious processor for us chessprogrammers to consider, >>just like power5 is not serious for us either. >> >>>gerold. > >thanks for the reply vincent i read that the cell chip would have a much larger >cache. > >gerold. Please realize that a power5 cardridge costs $100k. That's just a cpu cardridge, it has inside 8 cpu's or something.
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.