Speedmark scores on the Mac Pro

[A Nonny Moose]

Macworld tested out the new Xeon powered Mac Pro's and compared them with the Power mac G5 Quad and the Dual 2.7 Ghz G5. So without further ado, let's look at some numbers:

 

Speedmark overall scores:

 

Mac Pro/2.66GHz dual-core Intel Xeon x2 299

 

Power Mac G5 Quad/2.5GHz dual-core G5 x 2) 262

 

Power Mac G5 Dual/2.7GHz single-core G5 x2 267

 

20-inch iMac/2GHz Intel Core Duo 210

 

So while the Mac Pro's are faster than their G5 brethren, they aren't stupidly faster, as many people screamed they would be. They are however, extremely faster when compared to the iMac Core Duo. There's only a thirty point difference between the top two systems, but the real issue lies with the Xeon system havng the current record in terms of Speedmark scoring. Let's go on:

 

Adobe Photoshop CS2 Suite

 

Mac Pro/2.66GHz dual-core Intel Xeon x2 1:25

 

Power Mac G5 Quad/2.5GHz dual-core G5 x 2) 0:45

 

Power Mac G5 Dual/2.7GHz single-core G5 x2 0:48

 

20-inch iMac/2GHz Intel Core Duo 2:31

 

So as expected, the G5 systems remain the kings of Photoshop, and those users will continue to wait until Adobe releases Universal apps before making the big switch to Intel. In working with universal applications, the gap seemingly narrows:

 

Cinema 4D XL 9.5.21 (render), Compressor 2.1 (MPEG 2 Encode), iMovie 6.0.1 (Apply Aged Effect)

 

Mac Pro/2.66GHz dual-core Intel Xeon x2 0:28 1:47 0:38

 

Power Mac G5 Quad/2.5GHz dual-core G5 x 2) 0:30 1:52 0:39

 

Power Mac G5 Dual/2.7GHz single-core G5 x2 0:52 2:17 0:43

 

20-inch iMac/2GHz Intel Core Duo 1:11 3:21 1:03

 

Not much difference in any multimedia application, not even in iMovie. This doesn't bode well for the "PowerPC is {censored}" camp, as there still isn't a heck of a lot of difference. In all of the tests, so far, the Mac Pro spanks the iMac Core Duo, though. Let's play with iTunes and an MP3 encode:

 

iTunes 6.0.4

 

Mac Pro/2.66GHz dual-core Intel Xeon x2 0:48

 

Power Mac G5 Quad/2.5GHz dual-core G5 x 2) 0:43

 

Power Mac G5 Dual/2.7GHz single-core G5 x2 0:46

 

20-inch iMac/2GHz Intel Core Duo 1:26

 

The G5 wins, but it isn't a decisive victory, much like the Intel victories weren't truly decisive. It's beginning to be such a pattern that it forces the Mac community to ask "Why even switch to Intel when I can get a G5 for cheaper and it will work just as fast?" Well, let's play a game and see some framerates:

 

Unreal Tournament 2004

 

Mac Pro/2.66GHz dual-core Intel Xeon x2 91.3

 

Power Mac G5 Quad/2.5GHz dual-core G5 x 2) 62.2

 

Power Mac G5 Dual/2.7GHz single-core G5 x2 44.3

 

20-inch iMac/2GHz Intel Core Duo 54.1

 

As explained by Aspyr, GHz is GHz regardless of processor architecture, so it was a contest the Xeon won with the G5 coming in second here. That was a no brainer and a decisive victory for GHz and graphics cards, as Macworld tends to keep things to a "bought it from the store" configuration.

 

Finder (ZIP archive)

 

Mac Pro/2.66GHz dual-core Intel Xeon x2 2:01

 

Power Mac G5 Quad/2.5GHz dual-core G5 x 2) 2:22

 

Power Mac G5 Dual/2.7GHz single-core G5 x2 2:15

 

20-inch iMac/2GHz Intel Core Duo 2:34

 

Another victory for the Xeon, but it isn't a massive one (yet again) when you compare it with the G5 systems. Again, it's the same pattern with the exception of the game test. There just isn't enough of a speed increase so far to justify switching from the PowerMac G5.

 

Now this is only one processor configuration for the Xeon Mac Pros, but it looks like users won't notice any immediate speed improvements unless they are playing games exclusively on the Mac Pro. For rendering and even average day to day use, there *really isn't much* of a difference between the Xeon and the G5 and it hardly justifies a major purchase. There is also a 2 Ghz Quad model and a 3 GHz quad model that have yet to be tested by Macworld. While I can forsee some speed improvements with the 3 GHz model, that low end 2 Ghz Mac might not be worth the silicon it's inscribed on when you compare it with the high end Power Mac G5.

[embries]

Couple of of differences of opinion here.

 

A Dual - Dual G5 Mac (Quad) is still $3299 from Apple, with only 512mb of RAM, as you can see here. In this comparison, I could not find a specification for how much RAM was in the G5, so we'll assume it's the base 512mb. You said, "Why even switch to Intel when I can get a G5 for cheaper and it will work just as fast?". But as we can see, there's at least $800 price advantage to the Intel system.

 

That being said, the Speedmark is not 100% unibin, and 30 points is a huge improvement. Let it go 100% uni, and that gap will only widen.

 

I think we're all a bit disappointed in the long wait for the Adobe universal binaries, we all know this is putting a slow down on the transition to Intel for several would-be switchers.

 

For the multimedia apps, I believe (in all sincerity) that the tests chosen are too short for a workstation-class machine. This presents a problem when you also want to include the Single Chipers (iMac) etc, but you can't really get a measure for what a quad system can do in terms of memory, disk, processor throughput for media applications in 30 seconds or less. Heck with the size of memory buffers, you may still be filling a controller card's memory buffer before the test is over. I don't know which way the test would go if it were done on a larger scale, but I think it would be a more valuable test. Particularly the Cinema 4D testing would benefit from a little more thorough testing.

 

91 fps on the 7300 is respectable, they didn't specify resolution or settings, so that could be high or low. I think it is important to note, however, that if everything was held constant in comparison to the other test systems, a 50% increase in fps in a cheaper system is a value that translates directly to an every day user.

 

Again I'll agree with you, Moose there may not be a lot of reason to switch if you already had a Quad G5, but there's no reason to say that the G5 is a better purchase if you were looking to buy new right now.

[bluedragon1971]

The G5 wins, but it isn't a decisive victory, much like the Intel victories weren't truly decisive. It's beginning to be such a pattern that it forces the Mac community to ask "Why even switch to Intel when I can get a G5 for cheaper and it will work just as fast?"

 

 

The reason for the switch won't be as apparent with these first-gen systems. IBM was basically running out of steam with PowerPC development, while Intel was just getting started with the new Core architecture. I suspect that we will begin to see a much faster ramp-up in speeds soon.

 

Oh, and has already been pointed out, the G5 system is actually more expensive than the Intel system.

[SenVa]

I don't think apple really expected G5 owners to switch to the Quad Xeon. If they did they would have made the Mac Pro sooner. The mac pro is just for high end users that are now ready to switch to apple because of the windows compatablity.

[benson]

It's also worth considering that the Mac Pro v PowerMac comparison possibly isn't the best for justifying or criticising the Intel switch. While the G5 may have be en a very capable processor, it required massive cooling, and as such could only perform to the best of its ability in the PowerMac - and as such one could expect a reasonably strong performance in this comparison.

 

However, the PowerPC chips Apple was using in other applications (PB, iBook, Mini) appear far inferior when compared to their respective Intel successors. Due to cooling or other restrictions, the G5 couldn't be shoehorned into certain devices, and IBM (etc) weren't providing any other competitive solutions. Given Apple's huge reliance upon maintaining/growing its share in the notebook market (12% of the US market right now?), they couldn't afford to lag further behind, and so it was perfectly understandable that they made the switch to Intel. Additionally, as stated above, Intel seems to be going somewhere, and are giving out regular updates to a variety of different lines - things will only get better.

[A Nonny Moose]

I do agree that Intel is going somewhere with their chips, especially in the fields Apple desires (faster, smaller, cooler). IBM and to a lesser extent Motorola just weren't doing that (Moto did have cooler chips, but not faster/smaller). The Mac Pro seems to be geared more for those wanting to go away from the older G4 towers.

 

I still think Macworld did the right thing in comparing them to the G5 towers, but I do wish they'd compare the Mac Pro's to the G4 MDD towers, because that's where I think Apple is aiming the Mac Pro toward.

 

As for G5 prices, I expect them to go down some on the reseller market (always check out PowerMax.com for used tower prices). With the Mac Pro's, there will be no more G5 towers. So there is no more supply with a reduced demand (because people are going to go with new Mac Pros). This in turn drives the price down. Now where the price goes is up to pure economics, but in a way, I expect them to wind up slightly (SLIGHTLY) lower than the Mac Pro prices in order for those used systems to remain competitive price wise.

 

BTW, the only Rosetta test on Speedmark is the Word test. Photoshop is done strictly as a Photoshop test and is no longer a part of the Speedmark score.