Why is Intel iMac so much faster than MacBook Pro?

[Anonymous Freak]

Hrm...

Well, looking at the 'Universal' page, there are a few really odd results...

Comparing the 1.83GHz MBP to the 1.67GHz PBG4, the MBP is only 1.7x faster in their Pages test. But the 2.0GHz iMac/Intel is 1.9x faster than the 2.1GHz iMac G5. So the 5% slower clock speed iMac (slower than the PPC version,) has more of a speed boost than the 10% faster clock speed MBP (faster than the PBG4.) Not to mention the iMac actually saw a 5% REDUCTION in bus speed, while the MBP sees a 400% INCREASE in bus speed.

Comparatively, the MBP should have blown away the PowerBook by a much larger margin than the iMac comparison. I mean, look at the 'raw benchmark' comparisons that show the MBP 4-5x faster than the PBG4 while the iMac is only 2-3x faster. In real-world benchmarks, the MBP should at least have a bigger speed gain over the PBG4 than the iMac got.

The Doom comparison seems fine simply because of the graphics chipset improvement in both platforms. (Heck, even using the PPC version through Rosetta, the MBP should be faster than the PBG4.)

 

[jacobj]

Well, looking at the 'Universal' page, there are a few really odd results...

Comparing the 1.83GHz MBP to the 1.67GHz PBG4, the MBP is only 1.7x faster in their Pages test. But the 2.0GHz iMac/Intel is 1.9x faster than the 2.1GHz iMac G5. So the 5% slower clock speed iMac (slower than the PPC version,) has more of a speed boost than the 10% faster clock speed MBP (faster than the PBG4.) Not to mention the iMac actually saw a 5% REDUCTION in bus speed, while the MBP sees a 400% INCREASE in bus speed.

Comparatively, the MBP should have blown away the PowerBook by a much larger margin than the iMac comparison. I mean, look at the 'raw benchmark' comparisons that show the MBP 4-5x faster than the PBG4 while the iMac is only 2-3x faster. In real-world benchmarks, the MBP should at least have a bigger speed gain over the PBG4 than the iMac got.

The Doom comparison seems fine simply because of the graphics chipset improvement in both platforms. (Heck, even using the PPC version through Rosetta, the MBP should be faster than the PBG4.)

Could the Pages issue be related to the G5? There must be some things that the G4 would beat the G5 on when it comes to clock speed. I was in a mac shop yesterday benchmarking a Quad G5 with the Intel iMac 2.0 GHz. The test was simply running a 1080i HD trailer. Now they obvioulsy both managed to run at the full frame rate, but the iMac used 91% of the CPU whereas the G5 used 84%. So both were using single cores which means the comparitive clock speed for the iMac would be 72.8% running at 2.5.

I know that it is not quite as straight forward as that, but it shows that the Core Duo is a faster chip for some things than the G5. Different technology results in different strengths.

Also, Pages is notoriusly slow at most things and appears to be inefficient. It would therefore not be shocking to find that the code compiled for the intel is not as good as that for the G4. Perhaps Pages relies on some assets of the G4 that the Dup does not have?

Anyway, enough of rambling incoherently: the point is that different CPU architectures have different strengths and Pages is a cr*p app that should never be used full-stop, let alone in benchmarking tests...

 

[Anonymous Freak]

Also, Pages is notoriusly slow at most things and appears to be inefficient. It would therefore not be shocking to find that the code compiled for the intel is not as good as that for the G4. Perhaps Pages relies on some assets of the G4 that the Dup does not have?

Except I'm referring to the G4vsCD vs. G5vsCD. The Core Duos are the faster ones in both cases, it's just that the CD seems to have MORE of an improvement in Pages vs. the faster clocked G5 than the slower clocked G4. Implying that the G4 is actually better at Pages than the G5.

My point is that in the Pages test, the iMac Core Duo has a higher 'times faster than G5' rating than the MacBook Pro has over the PowerBook G4. This means that the Core Duo get more of an improvement over the G5 than the G4.

As the iMac Core Duo is 10% faster in raw clockspeed than the MacBook Pro; but the improvement in Pages is 11% better over the G5 than the MBP is over the G4, that would imply that a 1.67GHz PowerBook G4 is actually FASTER in Pages than a 1.9 GHz iMac G5 would be. (That's assuming that the Intel systems are normalized for clock speed, and that the MacBook Pro and the iMac/Intel would get the exact same 'raw scores' in every test if they were the same clock speed.)

I guess we'll only truly know for sure when someone benchmarks a 1.83 GHz MacBook Pro vs. a 1.83 GHz iMac Core Duo. If they score the same, then according to Apple's benchmarks, the 1.67 GHz PowerBook G4 is faster than the 2.1 GHz iMac G5.

 

[oingoboingo]

Don't overlook the obvious...

I think everyone is missing the obvious explanation for the benchmark discrepancy: Apple can't run an honest, solid set of benchmarks to save themselves. We might as well be arguing about whether Captain Kirk could beat Han Solo in a wet towel fight. Wait another week or two for MacBook Pros to arrive in the hands of the Mac-using community and let them run some credible benchmark comparisons.

 

[Mord]

I think everyone is missing the obvious explanation for the benchmark discrepancy: Apple can't run an honest, solid set of benchmarks to save themselves. We might as well be arguing about whether Captain Kirk could beat Han Solo in a wet towel fight. Wait another week or two for MacBook Pros to arrive in the hands of the Mac-using community and let them run some credible benchmark comparisons.

and then when they are about the same (maybe a bit higher, non beta software and all) people like you will quietly go away.....

 

[ewinemiller]

I think everyone is missing the obvious explanation for the benchmark discrepancy: Apple can't run an honest, solid set of benchmarks to save themselves. We might as well be arguing about whether Captain Kirk could beat Han Solo in a wet towel fight. Wait another week or two for MacBook Pros to arrive in the hands of the Mac-using community and let them run some credible benchmark comparisons.

Yeah, but really these days who cares, Lucas would just recut the scene so that Kirk snaps the towel first.

I agree on the Apple benchmarking. I think the quote I've seen that applies here "Don't assume conspiracy when incompetence will explain it just as well" or something like that.