Author: Matt Taylor
Date: 08:09:54 12/18/02
Go up one level in this thread
On December 17, 2002 at 22:51:47, Vincent Diepeveen wrote: >On December 17, 2002 at 14:14:53, Matt Taylor wrote: > >>On December 17, 2002 at 13:42:45, Robert Hyatt wrote: >> >>>On December 17, 2002 at 11:51:29, Vincent Diepeveen wrote: >>> >>>>On December 17, 2002 at 11:33:53, Robert Hyatt wrote: >>>> >>>>>On December 17, 2002 at 11:25:10, Vincent Diepeveen wrote: >>>>> >>>>>>On December 17, 2002 at 10:58:51, Bob Durrett wrote: >>>>>> >>>>>> >>>>>>Indeed you are correctly seeing that DIEP, which runs well on >>>>>>cc-NUMA machines as well, is a very good program from intels >>>>>>perspective, because even a 'second' processor on each physical >>>>>>processor which runs slower will still give it a speedboost, >>>>>>where others simply slow down a lot when you do such toying. >>>>>> >>>>>>So where many programs which will be way slower when running at >>>>>>4 processes/threads at a 2 processor Xeon, the software is the >>>>>>weak chain. >>>>> >>>>>What program fits this description? Not mine... >>>> >>>>*many* programs Bob. Crafty and DIEP aren't the only thing on the planet >>>>which gets used by most people who need a dual to perform multithreading >>>>for them. And majority of them uses NT4 server, 2000 server or XP pro/server. >>>> >>> >>>Yes, but you paint with a broad brush and say many will be way slower" and I >>>can't find a single one. I have run crafty. I have run several parallel >>>programs for things like molecular modeling, a simulation, and _none_ do worse >>>with HT on than with HT off. Not a single one so far. Some do _far_ better >>>with SMT than Crafty. Some do about the same. But I haven't found a one yet >>>that does worse unless you take the known exception for two processes running >>>on one physical cpu while the other is idle. That will be fixed shortly in >>>linux, and is already fixed in windows .net >>> >>>>For sure not some non existing OS that is seeing the clear difference >>>>between physical and split processors! >>> >>>Windows .net is not "non-existing". >>> >>>I don't follow the linux development kernels but wouldn't be surprised if they >>>already have fixes being tested... >> >>He is technically right about Windows .NET not existing. Microsoft isn't selling >>it yet, or at least I am under that impression as they usually test software >>-before- they release. I hope. >> >>>>>>In case of DIEP the bottleneck is the hardware clearly. Even >>>>>>something working great on cc-NUMA doesn't profit too much from >>>>>>the SMT/HT junk from intel. >>>>>> >>>>>>Though it is a great sales argument, the hard facts (11.4% >>>>>>speedboost) are not lying. >>>>>> >>>>>>So they need to press 2 cpu's which results in a cpu price >>>>>>2 times higher *at least* than an AMD cpu, the result >>>>>>is that you win 11.4% in speed. >>>>> >>>>>What are you talking about? SMT doesn't "press 2 cpus". >>>> >>>>the size of a P4 processor is a lot bigger than the AMD core, that's >>>>explaining for a big part why P4 is so much more expensive than a K7. >>> >>>What does that have to do with SMT? >> >>Let me rephrase your question. "The Tampa Bay Buccaneers are 11-3. What does >>that have to do with anything?" >> >>The size doesn't offset the cost by $600. Profit margins do. > >it depends how many channels get traversed. > >If i produce a product for $50 then the importer wants to sell it perhaps >for $200 and gets $150 . then the shop sells it for $200. > >So if a chip gets produced for $20 extra, then that's already like $100 >extra in the shops real quick... > >Also even a layman understands that millions more transistors on >a chip in this case means bigger chip size means also bigger number of >incorrectly pressed CPU's. > >So just the extra price for pressing the bigger cpu already means quite >some dollars extra which one has to pay. > >I do not know of course for what price these wafer machines >operate to produce 1 pallet full of cpu's. > >From 10 different hardware experts i heard 10 different prices >there, so i won't do a guess here, except than to claim that it >gotta be a lot more than an additional $1 you said in some other posting >here that the 2nd chip onto the P4 costs, it gotta be many >millions of transistors!!!!! > >Best regards, >Vincent > > > >>-Matt The last transistor count I saw looked something like this: Athlon - 37.6M transistors (AthlonXP 2400) P4 - 55M transistors (P4 2.53 GHz) I can't find HT data. I didn't check Intel's website because that usually results in a vain search when looking for any technical details. Still, HT didn't make sweeping changes. They just added a little multiprocessing logic. The major gap between P4/Athlon transistor counts is that AMD hasn't yet put a 512 KB L2 cache on Athlon. (That's been planned for over a year, and it supposedly will release early next year.) It is a fact that Intel uses larger wafers and thus gets lower yields. The die size on P4 used to be almost double the die size for Athlon, and it's still considerably larger, but the gaps will close when AMD puts a 512 KB cache on. Yes, it costs more to produce a P4. You want to know how to estimate their bottom-line? Go look at the low-end Northwood processors. Look at the price. The bottom-line is somewhere below that. I'm looking and I see it's below $100 USD. It costs them roughly the same to produce a P4 3.06 GHz. If the P4s that come off the line won't run 3.06 GHz, they sell them at 2.8 or 2.53 GHz. It doesn't take a hardware expert to figure this out; it only takes a little common sense. -Matt
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.