good article on P4/Atlhon FPU

G

Guest

Guest
read this for some insights on P4 / Athlon FPU performance and compiler optimizations. Very good article; though Im not too sure what to make of it..but it seems like the P4 does have some potential thanks to brilliant compiler development by intel.

<A HREF="http:///www.aceshardware.com/Spades/read.php?article_id=40000189" target="_new">http:///www.aceshardware.com/Spades/read.php?article_id=40000189</A>

---- Owner of the only Dell computer with a AMD chip
 
G

Guest

Guest
Yup, the P4 fpu crushed amd 2nd rate fpu. Well guess what the P4 is future and the Athlon is ok of date.

! Member of the Intel Triple Threat w/ Fugger and AMDmeltdown !
 
G

Guest

Guest
Thats exactly what the article is NOT saying. I guess its too technical or difficult for you. Here is small quote:

"it is safe to say that the Athlon x87 FPU is in reality far superior to any x87 FPU on the market".

Thats not to say this necessarely translates into performance, thats what the article is about.

---- Owner of the only Dell computer with a AMD chip
 
G

Guest

Guest
Yea, that was a pretty interesting article.

I wish I could have gotten the beta release of the Intel compilers for Linux working while I had access to that P4. Because, without that compiler, the P4 FPU sucks completely.

It wouldn't make to much difference for my app though. Even if my program got the biggest improvement seen in this article (about 33% wrt the 1.2GHz Athlon), it's price/performance ratio would still be about 12% higher then the Athlon systems I will likely go with. Also note that the P4 still did considerably worse on several of the benchmarks (arguably, the less important ones, unless your program happens to behave similarly).

Of course, if you app depends on maximum performance from a single box (mine is distributed), then the P4 could be a good choice.

The smart thing for Intel to do at this point, would be to give the compiler away, including the source. What they lose in revenue from compiler sales would likely be made up many times over in faster adoption of the P4.


<i>Cognite Tute</i>
(Think for Yourself)
 

74merc

Distinguished
Dec 31, 2007
631
0
18,980
I read that yesterday.
good article.
the P4 has potential, but in optimizing for it, you seem to artificially slow down the other processors performance on the same software.
double edged sword, sounds convenient for Intel. :)
but obviously, some programmers have noticed that as well...

----------------------
Independant thought is good.
It won't hurt for long.
 
G

Guest

Guest
"The smart thing for Intel to do at this point, would be to give the compiler away, including the source. "

Absolutely. However, I think the investment they made into this compiler is huge. Even for its current price, you could consider they are giving it away.. kinda..



---- Owner of the only Dell computer with a AMD chip
 

Kelledin

Distinguished
Mar 1, 2001
2,183
0
19,780
Lo freakin' L! You completely missed the entire point of the article!

Once again, in accordance with the <A HREF="http://forumz.tomshardware.com/community/modules.php?name=Forums&file=viewtopic&p=9642#9642" target="_new">general opinion of the community</A>, you are caned 50 licks for being too stupid to even read the article. The standard "reattach-your-ass-and-run" applies.

Kelledin

"/join #hackerz. See the Web. DoS interesting people."
 
G

Guest

Guest
lol
if you read the article you would now that an athlon still outperforms an P4 on the most benchmarks and the P4 is much better supported then the athlon and if you know that those results can't be compare in reallity and that the P4 with support will be worse then an athlon without support. and guys you P4 lovers won't be so happy when they compare an A4 with a P4; because an A4 has SSE and that helps the score about 6% higher then now theoretically so [-peep-] the P4 for now

computer