How does your Mac fare against the duo?
After the initial sobbing in how fast the new mactels are and how pitiful our old G5 iMacs and towers are, reality has set in for us PPC G5 owners.
If you havnt seen the benchmarks by macworld, take a look:
http://www.macworld.com/2006/01/feat...est1/index.php
Though the new iMac duo fares faster, it is not 2X across the entire spectrum. How much of an affect the lack of altivec-ish like code is up for question.
xBench has already been compiled for the new mactels and this provides an excellent side by side comparison to the new macs. Interestingly, my dual 2.0 G5 scores 1.66X overall than the new iMac Duo 2.0 GHz
It looks like the bang for the buck is the same, 1.66*$1299 = $2156 about the same price of this G5 tower.
***Please note that when comparing scores you must compare results obtained with the same version of xBench***
If you havnt seen the benchmarks by macworld, take a look:
http://www.macworld.com/2006/01/feat...est1/index.php
Though the new iMac duo fares faster, it is not 2X across the entire spectrum. How much of an affect the lack of altivec-ish like code is up for question.
xBench has already been compiled for the new mactels and this provides an excellent side by side comparison to the new macs. Interestingly, my dual 2.0 G5 scores 1.66X overall than the new iMac Duo 2.0 GHz
It looks like the bang for the buck is the same, 1.66*$1299 = $2156 about the same price of this G5 tower.
***Please note that when comparing scores you must compare results obtained with the same version of xBench***
Comments
Originally posted by MajorMatt
Though the new iMac duo fares faster, it is not 2X across the entire spectrum. How much of an affect the lack of altivec-ish like code is up for question.
1: Jobs said the PROCESSOR is faster, not everything, infact he said something like "the HDD is still the bottleneck, we havnt changed that"
2: Intel has MMX and SSE which, when code is crafted for them, could probably smoke the Altvec engine
Originally posted by MajorMatt
After the initial sobbing in how fast the new mactels are and how pitiful our old G5 iMacs and towers are, reality has set in for us PPC G5 owners.
If you havnt seen the benchmarks by macworld, take a look:
http://www.macworld.com/2006/01/feat...est1/index.php
Though the new iMac duo fares faster, it is not 2X across the entire spectrum.
I would suggest to first check the comments in this thread. Since the applications tested are not multi-processor aware, the Macworld tests essentially proved that the Yonah chip is better per core than the G5 for general day-to-day use.
Also, from the other thread:
I have to say after a few hours with my 20" Intel iMac I love it. I just sold a 2.0Ghz 20" G5 and the Intel iMac feels much snappier. The biggest difference I've noted so far is in handbrake. H.264 encoding on the G5 was slow while the Intel iMac is basically encoding in real time. From a FPS perspective the Intel iMac is ~4-5X faster encoding @ an average of ~25 FPS while my G5's average was ~5-7 FPS.
Even video encoding is not anymore G5's strong point.
For other benchmarks, take a look in Craig Wood's page or in this Ars article. And have always in mind that software optimisations (minimal for the time being on Intel Macintosh) will make a big difference.
Originally posted by a_greer
1: Jobs said the PROCESSOR is faster, not everything, infact he said something like "the HDD is still the bottleneck, we havnt changed that"
Processor performance is a very complicated thing. The new processor is faster in some ways, the G5 is faster in others. For most things done on a Mac by most people the Intel processor is faster, but this does not hold for everything.
2: Intel has MMX and SSE which, when code is crafted for them, could probably smoke the Altvec engine
Unfortunately not. At anything approaching the same clock rate AltiVec will smoke SSE/MMX. Better cache and better memory latency/bandwidth will often be the larger factor, however, and outweight AltiVec's general supriority.
Originally posted by PB
I would suggest to first check the comments in this thread. Since the applications tested are not multi-processor aware, the Macworld tests essentially proved that the Yonah chip is better per core than the G5 for general day-to-day use.
Perhaps, but they also used some of the only currently availble universal apps, and the ones most likely to be used by Mac users - Apple's apps. If they're not multi-processor aware, perhaps they should be, in order to get the most out of this intel chip.
It seems to me that if you have a recent Mac, you're probably going to see slightly better performace with universal apps, and much worse on rosetta apps.
I'm going from a PowerBook 800 to a MacBook Pro, and I expect to see roughly equivalent performance on Rosetta apps but significantly better performance on universal apps. If you are switching from a current iMac, I doubt you'll be blown away by the speed very often.
Originally posted by BRussell
Perhaps, but they also used some of the only currently availble universal apps, and the ones most likely to be used by Mac users - Apple's apps. If they're not multi-processor aware, perhaps they should be, in order to get the most out of this intel chip.
The iMac Core Duo is the first dual core mainstream Macintosh. So, you can bet that Apple is working hard right now to get the most out of the new architecture (that is, not only Intel, but also dual CPU). When we will see the fruits of the effort is at this moment anyone's guess. Unless someone here knows and shed light.
It seems to me that if you have a recent Mac, you're probably going to see slightly better performace with universal apps, and much worse on rosetta apps.
This is true. In real world scenarios, where there is a mix of native and Rosetta code with, for the moment, the Rosetta overweighting the native for many people (especially video/image/audio pros), you will feel the performance hit. But it is always like that in such a transition. I expect that in a year or so, the situation will be smooth enough. It is a good thing that Apple pushes so hard the transition. It will make developers jump the train sooner rather than later, ultimately making life easier for early adopters.
I'm going from a PowerBook 800 to a MacBook Pro, and I expect to see roughly equivalent performance on Rosetta apps but significantly better performance on universal apps. If you are switching from a current iMac, I doubt you'll be blown away by the speed very often.
This too is true. In common OS operations, the new 20" iMac will feel like a 2.0-2.3 GHz Power Mac, more or less. And since this is a consumer machine with vendor software (iLife) that does not take advantage of two processors, expect this software to run like in the previous mono-processor (G5) model. However, as more multi-processor aware consumer software will become available (and it is strange that iLife still it is not), expect the dramatic improvement a dual core system will deliver. There is no magic in that, just two processors with the benefits that we know well for so long here.
I'd expect performance on intel Macs will only improve as more optimizations are made and more work is offloaded to the graphics card.
doesn't apple usually use more ram when testing them?
how much of a difference does it make having more ram in these tests?