Author: Robert Hyatt
Date: 14:19:56 10/02/02
Go up one level in this thread
On October 02, 2002 at 16:25:34, Vincent Diepeveen wrote: >On October 02, 2002 at 16:20:39, Robert Hyatt wrote: > >>On October 02, 2002 at 15:13:32, Vincent Diepeveen wrote: >> >>>On October 01, 2002 at 22:43:58, Robert Hyatt wrote: >>> >>>>On October 01, 2002 at 09:43:00, Vincent Diepeveen wrote: >>>> >>>>>On September 30, 2002 at 12:13:44, Robert Hyatt wrote: >>>>> >>>>>>On September 30, 2002 at 00:09:28, Jeremiah Penery wrote: >>>>>> >>>>>>>On September 29, 2002 at 23:31:36, Robert Hyatt wrote: >>>>>>> >>>>>>>>I don't know what this means. I have several dozen programs (Crafty >>>>>>>>is only one) that we have run using intel's compiler and gcc, and in >>>>>>>>_every_ case, Intel's compiler is faster. On P2's, on P3's and on >>>>>>>>P4's... Of course I wouldn't use intel's compiler for an AMD chip, >>>>>>>>why would they want to optimize for a competitor's chip??? >>>>>>> >>>>>>>They don't have to optimize specifically for the competitor's chip, as Intel >>>>>>>compiler still produces probably the fastest binaries for AMD machines. Any >>>>>>>general optimizations (P2, P3, and even P4 optimizations (excluding SSE2 stuff >>>>>>>or whatever)) are just as helpful for AMD processors as they are for Intel ones. >>>>>> >>>>>> >>>>>>Maybe or maybe not. AMD's pipeline is different, and there are subtle >>>>>>differences in instruction choices, that can make a difference in speed. I >>>>>>don't see why the Intel compiler guys would bother studying AMD at all... >>>>> >>>>>I bet 50% of their time goes into studying what is faster for P4 than for K7 :) >>>> >>>> >>>>I'd bet they don't. Optimizing for a specific processor family is tough. >>>> >>>>Trying to optimize for one while producing code that does worse on another >>>>processor is a _real_ can of worms. I don't think anyone would waste that >>>>kind of time. >>> >>>I am very sure they will. We talk about billions being at stake here. >>>It is completely naive to suppose they do not study the K7. >> >> >>What does it feel like to be "the world's foremost authority on everything?" >> >>I am sure Intel has studied AMD chips. I am also sure that the _compiler_ >>guys are _not_ paying any attention to it, because there is no point to doing >>so. Have you written a compiler? I have. Have you written an optimizer? I >>have. Do you know what you are talking about here? I do. >> >>Don't make things up. Get real answers. Just like the quad 1.6ghz machines >>that >>you claim do not exist but which dell is shipping. >> >>I explained that to you a couple of weeks ago. I can't find that quad 2.2 I saw >>the >>output from, as that is what I was looking for. But I found several quads in >>the 1.4-1.6 >>range. >> >> >> >>> It is very >>>good deal to pay a few guys fulltime in order to sell for a couple of >>>billions more. Because if YOUR compiler, which without question >>>is doing great at specint tests, is going to let their processor look >>>better then you sell a couple of billions less. >> >>Again, the compiler guys are more interested in making code run faster on their >>processor. Not in making it run _slower_ on an AMD processor. Nothing forces >>_anybody_ to use the intel compiler to produce SPEC numbers. If their compiler >>is >>worse than MSVC then everyone would use MSVC. Intel would have wasted hundreds >>of thousands of dollars on the compiler development, and gotten _nothing_ from >>it. In >>reality, they just try to make the code run as fast as possible on their chips >>and to heck >>with everyone else... >> >> >> >> >> >>> >>>Do you want to take the risk of a couple of billions? >>> >>>Say 50 billion dollar? >> >> >> >>make up whatever number you want. It doesn't change the basic facts I gave you >>above. Wave your hands all you want, if it makes you feel better, although it >>won't >>make your ramblings true. > >Don't make idiot statements here. Lesson 1 in economy is: > - what is a company? If you would follow your own advice and "not make idiot statements here" it would save a _lot_ of discussion. >answer: an institute that tries to make money > >Now question to you is: do you want to make money or not? >answer: yes > >Hardest way to do it >answer: improving your processor > >Hard but way easier way to do it >answer: improve compiler such that you get faster but not competition That would only be true if there was _one_ compiler. But that is not the case. There are _several_. Intel. Microsoft. Gcc. The portland compiler group. Probably others I have forgotten about or not ever tested. Does not good to make one processor look bad on _your_ compiler, when everyone can use _another_ compiler and make all that work turn into wasted effort and expense... > >Of course the answer is not: "improve your compiler such that possibly >competition profits even more from your compiler than you do". The answer is to make your compiler produce code as efficient as possible for your processor. _period_.
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.