PDA

View Full Version : Mac Pro and Power Mac G5 benchmarks


MacBytes
Aug 9, 2006, 05:13 PM
http://www.macbytes.com/images/bytessig.gif (http://www.macbytes.com)

Category: Benchmarks
Link: Mac Pro and Power Mac G5 benchmarks (http://www.macbytes.com/link.php?sid=20060809181346)
Description:: Geek Patrol has benchmarked a Mac Pro and a Quad Power Mac G5. The Mac Pro isn't that much faster than the Power Mac G5, but it is a lot more affordable.

Posted on MacBytes.com (http://www.macbytes.com)
Approved by Mudbug

Fredo Viola
Aug 10, 2006, 06:45 AM
what a huge bummer. I thought they said it would be 2x faster? Big difference between that and actually only 7%. Ugh.

Timepass
Aug 10, 2006, 07:50 AM
wow looks like apple got cought lying about their system speed yet again. How many times is that now? and why do people still believe those numbers apple puts out?

Queso
Aug 10, 2006, 07:56 AM
Not surprised at all here. Apple has been dissing the G5 at every chance, but it's still an incredibly quick processor and comparable to the current Intel offerings. The Quad G5 is one incredibly fast machine.

BUT you can't get a G5 in a laptop. And there's the most important factor.

mkrishnan
Aug 10, 2006, 08:09 AM
Apple making exaggerated claims about the prowess of a new computer? Nooooooooooooooooooooo.... Nooooooooooooooooooooo.......... I don't believe it.

:eek: ;) :D

But they did say real-world performance. And Geekbench doesn't really do a good job of taking into account factors outside the CPU, like the change in the bus architecture, does it? Since it does such heavily CPU loaded tasks?

And people are doing this Photoshop test but they are not using a Photoshop universal binary (presumably, since nonesuch is available).

So one needs to look at real world performance metrics in a universal application in order to validate or invalidate Apple's claim.

Also, so I'm not super familiar with geekbench, but since it has no platform-specific optimizations, and it's written in C++, how well threaded is it? Is it adequately able to make use of multiple CPUs?

Sly
Aug 10, 2006, 08:11 AM
From Apples web page: the Mac Pro runs up to 2x faster than the Power Mac G5 Quad (1). (1) Based on estimated SPECint_rate_base2000 results. Testing conducted by Apple in July 2006 using pre-production 3GHz Mac Pro units and shipping 2.5GHz Power Mac G5 Quad units. SPECŪ is a registered trademark of the Standard Performance Evaluation Corporation (SPEC); see www.spec.org for more information.

As these Geekbench results show the Mac Pro 3Ghz would be more than 2x as fast as the Powermac with integer calculations and scalar floating point calculations. Please bear in mind this benchmark was conducted on the mid range 2.66Ghz Mac Pro not the top end 3Ghz model and that the Geekbench "score" is an indication of overall speed, where as Apples tests, not surprisingly are a narrow window on the latest models best performance.

RacerX
Aug 10, 2006, 08:22 AM
wow looks like apple got cought lying about their system speed yet again. How many times is that now?And yet they have a better track record than you... How in the world can that be? :eek:


As for the benchmarks... they say this is a mid-range Mac Pro (the 2.66 GHz model rather than the 3.0 GHz model).

Apple has always made sure that the mid-range system are about equal to the previous top of the line system. When Apple released the first G5s, they made the mid-range G5 a single processor system at 1.8 GHz... which benchmarked about even with the 1.42 GHz Dual G4 that was the previous top of the line. A few months later Apple replaced the single 1.8 GHz system with a Dual 1.8 GHz... which was quite a bit faster than the 1.42 GHz Dual G4.

Apple has historically kept the mid-range system within reach of the previous top of the line to ease buyer's remorse.

This is obvious Apple doing the same thing they've done for quite some time.

And looking closer at the benchmarks, the G5 did better on things PowerPC processors are better at and the Mac Pro did better on things that Intel has historically been better at.


No surprises here... must be a slow news day. ;)

spicyapple
Aug 10, 2006, 08:28 AM
The Xeon don't have PowerPC Altivec "Velocity Engine". Glad I still have my dual G5. :D

Timepass
Aug 10, 2006, 08:41 AM
Apple making exaggerated claims about the prowess of a new computer? Nooooooooooooooooooooo.... Nooooooooooooooooooooo.......... I don't believe it.

:eek: ;) :D

But they did say real-world performance. And Geekbench doesn't really do a good job of taking into account factors outside the CPU, like the change in the bus architecture, does it? Since it does such heavily CPU loaded tasks?

And people are doing this Photoshop test but they are not using a Photoshop universal binary (presumably, since nonesuch is available).

So one needs to look at real world performance metrics in a universal application in order to validate or invalidate Apple's claim.

Also, so I'm not super familiar with geekbench, but since it has no platform-specific optimizations, and it's written in C++, how well threaded is it? Is it adequately able to make use of multiple CPUs?

There is exaggeration and then there is just bold face lying. Apple chose the lying route. 10-15% over is exaggeration 20% is pushing it but this time apple was at a 1000% over. This is a new recored for apple in "exaggeration" The halo ones where at 200% and that was funny when apple got nailed on that one and people where unhappy about being lyed to. But at 1000% over board is a little insane. I honestly would love to see the media get a hold of this infomation and run with it. Apple deserves it and really needs to be taken down a few pegs and it high time they start using a bit more real world adverstiments that the BS that they been spitting out for years.

nospleen
Aug 10, 2006, 08:48 AM
I think we should wait to judge until we see real benchmarks. I prefer the tests that macworld puts out. These geek tests and xbench do not mean squat to me. Tell me how it does in FCE and iDVD, etc...

mkrishnan
Aug 10, 2006, 08:50 AM
There is exaggeration and then there is just bold face lying. Apple chose the lying route.

I personally think it's harmless posturing that no one believes anyways. Even SNL has made fun of Apple for doing this with the iPod Steve Jobs on the SNL News Broadcast sketch.

1) Switchers don't care about Mac stats vs. previous Macs. They care about how their computer will work on practical tasks vs. what they have now. So I don't see switchers buying droves of Mac Pros based on these fantastical pronouncements of improved speed.

2) As for the fanbase, well, they're getting Macs one way or another. And with all the hullabaloo, it usually ends up turning out that each generation of Macs is still better than the previous ones....

I view it as harmless verbal masturbation on the part of Apple (and, hello, pretty much the rest of the computer industry, although I think Apple is possibly the worst culprit). I think it's kind of comical. But it doesn't take much away for me from the the fact that they design nice computers. :)

MacsRgr8
Aug 10, 2006, 08:51 AM
It could get even worse....

Imagine that these Xeons are going to be benchmarked in the PC world, against older Intels like the Pentium D.
It wouldn't surprise me to find out that many non-SMP apps (like games, which are used alot in PC benchmarking) are not going to be faster on the newer Xeons, than the older higher clocked Pentium D's.

mkrishnan
Aug 10, 2006, 08:54 AM
It wouldn't surprise me to find out that many non-SMP apps (like games, which are used alot in PC benchmarking) are not going to be faster on the newer Xeons, than the older higher clocked Pentium D's.

Benchmarking a Xeon with a video game and calling it meaningful would be even lower than teh Steve's typical behaviour. :D

MacsRgr8
Aug 10, 2006, 08:56 AM
Benchmarking a Xeon with a video game and calling it meaningful would be even lower than teh Steve's typical behaviour. :D

LOL

Good point.
But I am pretty sure it will happen, and many PC gamers will start laughing at our "Quads".