Comprehensive Cross platform benchmarking

[FrgMstr]

http://www.jc-news.com/parse.cgi?pc/benchmarking/xplat/ppc-x86

dont forget however the X86 procs used are the OLD Slot loading .25 micron K7s and P3's nothing new like a thunderbird , Athlon XP or a P4, of course not the latest G4's either but then there isnt a huge change in architecture in the G4 over the same period of time besides the move to a higher FSB and intergrated level 2 cache.

Anyways have a look. Its pretty big.

 

[Daeyin]

These Benchmarks have some SERIOUS flaws.

(Apart from some of the links no longer existing)

All the ones from spec.org have the disclaimer
http://www.spec.org/spec/disclaimer.html

Basically saying these aren't official, and the testers can show biased results.

You'll notice that some of the test results are gathered from data that appeear in several tests
"Photoshop (PS5Bench/21 filters):
<lower is better>
400 MHz G4: 121.8
450 MHz G4: 117.6
500 MHz G4: 101.8
500 MHz PIII (Win98): 161.5
600 MHz Athlon (Win98): 138.7
800 MHz Athlon (NT4): 100.8
800 MHz PIII (Win2k): 95.0
1000 MHz PIII (Win98): 85.8
(macinfo.de)
(xlr8yourmac.com)
(geocities.com)
(tech-report.com)
(tech-report.com)
(xlr8yourmac.com)
"

Going to the sites that he sites will show you that many of them were done different system configurations that happened to have the same processor in them.

The ones from Blanos.com seem to be legit. However, even those only give limited system specs.

The tests from swox.com include the warning that "the numbers for the G4 are unfortunately either estimated
or based on vendor supplied data (according to website)"
That is not part of the scientific process.


The ones that attempt to use very similar systems usually get results like we expect.

http://www.barefeats.com/pentium.html
http://sfgate.com/cgi-bin/article.cgi?file=/c/a/2000/04/17/BU1016CH.DTL

with the G4 being faster than pentium of almost twice the clock speed on some items, and at about the same speed or slightly lower for others.

Now, most of these tests are old, and this is not to say that the current G4 is faster than the fastest P4 or AMDs newest chips. I by no means believe this mostly because (at least with intel)
1) P4 is updated more often (new speeds almost every handfull of weeks). The G4 gets a boost 2 or 3 times a year (and we are long overdue for one).
2) the G4 came out before the P4 and is DATED DATED DATED. Apple needs to release a new architecture to replace it.
3) Apple, unlike intel, does not release its fastests chips for marketing reasons (and probably quality control). I am sure they have 1, 1.2, 1.4+ Ghz G3s sitting around that they wont use.

AND
With Apple saying that the next Macworld Expo is "Big, even by our standards" this whole arguement may be moot in one week. We'll have to see.

 

[vanguard]

Disclaimer: Even though I find myself defending Wintel all the time I really do prefer OSX. I just think it's important to know where we stand. Paying too much attention to silly things (IMHO) like case color, one sided benchmarks, or whatever does not help Apple. According to me, it's just sticking your head in the sand.

Daeyin,

The benchmarks you site are silly. They make no attempt to hide their apple bias and they are neither complete nor up to date. There are many examples this but I'll pull out a few of the good ones:


The Pentium 4 was only 30% faster than the G4/500 even though it was running at a clock speed that is 180% faster.

It's well known that Intel made sacrafices to hit higher clock speeds. They thought that these clock speeds would make up for the losses suffered during context switching and that they would pay off even further as their design allowed them to increase the clock speed even more. They were right. It was faster then and they have been able to increase the clock speed even more. It's nice to see that you did make the admission that Apple simply can't keep up with current offerings. I'll make my own admission. Right now is a harsh time to judge Apple, they'll probably make some step towards catching up to the x86 vendors next week.

Still, running at half the clock speed of the competition is a significant handicap, and as I tested the performance of the new gigahertz PCs, I wasn't at all certain Apple's machines could keep up.

I wonder who he's rooting for?

It's worth noting that he only tests the photoshop "bake off". That's not exactly a complete test of a system. It's a valid test but far from complete. It's like saying that Carl Lewis is a better running back than Barry Sanders because he's faster. (Ok, maybe American sports references aren't a good thing in a global forum. One was a great foot player the other was a track star.)

FLASH! AS OF OCTOBER 2001, THERE HAS BEEN A BREAKTHROUGH in Macintosh Quake 3 performance. I've achieved 184 frames per second on a Dual G4/800 with GeForce3 and OS X 10.1 running Quake 3 Arena 1.30 beta for OS X. A Pentium IV 1.6GHz with GeForce3 achieved the identical frame rate.

Ok, so it's safe to say that the x86 crowd has pulled away here too? Pentiums are at 2.0 ghz now and the AMD 1900+ is even faster than that in most benchmarks.

Anyway, I hope that Apple/Motorola/IBM really does something noteworthy next week. The gap is getting wider all the time.

 

[Daeyin]

Never said the ones I site were any less biased, just more of what we expect (i.e. what Jobs has spoon fed us).

And I did say they were old, and that current P4 and AMD chips are much faster than the Dual 800.

BUT, power aside, Apple has gold with OS X if they can do it right (and they are doing fine so far, IMHO).

 

[vanguard]

Originally posted by Daeyin
BUT, power aside, Apple has gold with OS X if they can do it right (and they are doing fine so far, IMHO).

Aha! Now we agree. Actually, I think we agree on all your points. Let's hope that after the next mac world apple can win the priase of independent sites and not just pro mac sites. In other words, let's hope they add horsepower to their list of "we're better" line items.