Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: minnow68 who wrote (29111)	2/21/2001 10:43:22 AM
From: jcholewa	Respond to of 275872

> > P4 might be a good machine for a few scientists who run their own compiled floating point apps > only true if they purchase the Intel C compiler in addition to Visual C++ (which also means that > they can't be running Linux or Unix). Our good scientist would also have to be able to compile his > app with the Intel compiler, and trust that the results are valid. > Having used the Intel C compiler (admittedly V4.0), I would _never_ trust that compiler. IMHO, > what Intel shipped to me should have never been released, let alone charged for. I should note that my resident scientist, Tim Wilkens (coder of Science Mark), has been having massive problems with the Intel C compilers. His tests, which generally stress force interactions between teensy-weensy particles and wavicles and suchlike and soforth, compile wonderfully on many different architectures and a few compilers on x86 (I think), but the Intel C compiler plugin is just not friendly with this standard code. Tim is getting some help from Intel's helping people (dunno what they're officially called), though the help is proceeding rather slowly, I hear. I'm sure the compiler plugin compiles loads of programs from many coders, including the scientific type. But I think that Scumbria's blanket statement of it being good for scientists who run their own compiled apps should be tempered with the amendment that this does not always work, and additionally an optimized scientific app in some cases may still not be immensely P4 friendly. For example, when compiled for the P4 with the latest Intel compiler plugin, as shown in jc-news.com; , the povray benchmark (I admit that this isn't a scientific benchmark as many/most describe it, but it's probably close enough to make a point) has a marginal performance edge to the K7 processors (the recompile essentially doubled performance from the old binary to the optimized binary, but the compiler wasn't smart enough to put in any SSE2 fp code -- this is a factor that the entire universe ignores when they talk about optimization for new processors). -JC PS: Incidentally, SI gave me trouble with the URL I posted, which ends in a close parenthesis. The way I have fixed this is by placing a double quote mark directly after that ending parenthesis. For some reason, SI appears to have eaten that quote mark while keeping everything before it. However, if the link does not work for you, copy it to your clipboard and make sure it ends with a close parenthesis and not a double quote. :)