Author: Matt Taylor
Date: 13:35:04 02/22/03
Go up one level in this thread
On February 22, 2003 at 10:42:30, Charles Worthington wrote: >On February 22, 2003 at 03:21:27, Matt Taylor wrote: > >>On February 22, 2003 at 02:21:52, Charles Worthington wrote: >> >>>On February 22, 2003 at 01:08:15, Jeremiah Penery wrote: >>> >>>>On February 22, 2003 at 00:31:38, Charles Worthington wrote: >>>> >>>>>Oh and i also believe that if AMD had the technology or the research funds to >>>>>have invented hyperthreading technology...they would have. it doesnt take a >>>> >>>>You're making some statements out of absolute ignorance here. >>>> >>>>First of all, Intel didn't invent "hyperthreading" technology. Nor do they own >>>>exclusive rights to such technology - there's no legal issue preventing AMD from >>>>implementing something like it in their processors. Second, for the past >>>>several years, AMD has produced more patents than Intel has, despite a R&D >>>>budget several times lower than Intel's. What does that say about Intel's >>>>"technology or research funds"? >>>> >>>>The P4 has notoriously bad IPC compared to other recent x86 processors. You >>>>could say that a big reason for adding hyperthreading is to ameliorate this >>>>condition. I.e., the addition of hyperthreading brings the IPC up to a more >>>>respectable level, though it's still somewhat lower than that of the Athlon. >>>>Without hyperthreading, all the P4 has going for it is a high clockrate. The P4 >>>>_needs_ hyperthreading to keep its performance advantage. >>>> >>>>>rocket scientist to see that two threads are better than one for multiple >>>>>applications. >>>> >>>>And 4 threads are better than 2...What's your point? >>> >>>Actually the point is really simple...hyperthreading produces a boost in speed >>>and an extra thread to run background applications faster. If AMD has the >>>know-how to incorporate hyperthreading technology into their processors they >>>would be foolish not to economically and performance-wise....The truth of the >>>matter is....they cant do it and maintain a bargain basement price for their >>>products...and in all reality...I'm not 100% sure they have the know how or else >>>they could be marketing a high end hyperthreading chip to compete with the >>>xeon...or maybe they just don't want that extra money? >> >>The truth of the matter is doesn't cost much to add HT to a processor. Bob >>corrected me some time ago when I tried to claim the same thing. >> >>-Matt > > >Then I stand corrected Matt. There are a few people here I do actually listen >to...You, Bob, Mike Burne and Eugene. I do not claim to know all about these >machines_but_economics I do know. But when it comes to computers and I hear >Aaron (22 Years old with no degree) trying to argue with a PHd that overclocking >a server is okay, It is more than a little amusing. I mean it's not like he can >have years of experience doing this.:-) I have known Aaron for a while. He does a lot of overclocking and reading on the subject. While I do disagree with him about overclocking, I turn to him whenever I have questions related to hardware (e.g. performance, stability, etc.) The big discontinuity between Bob and Aaron is, I think, both experience and perspective. If Aaron's machine crashes he lowers the clock speed and reboots. Bob has dealt with situations where if the machine crashes money is lost. Bob is more inclined to be extra careful in avoiding such a situation. Aaron has nothing to lose, so he goes for the extra speed. Aaron is right on one account, too. The engineers are conservative in clocking the processors. It is often safe to overclock, but it is always a gamble. I agree with Bob in that I don't care to gamble. I'd rather blame nVidia and Microsoft when things crash. -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.