Why is Intel iMac so much faster than MacBook Pro?

[verb]

On Apple's Universal page ( http://www.apple.com/universal/ ) the bars inidicating speed differences at the bottom of the page seem strange. They seem to imply that the Intel iMac is significantly faster than the MacBook Pro.

Like Doom 3 is 2.3x faster on an Intel iMac than an iMac G5, but only 2.2x faster on a MacBook Pro than on an Powerbook G4. Since the iMac G5 is much faster than the Powerbook G4 that means the iMac is much faster?

The same phenomenon is true with pages also.

The bar lengths seem to imply the results of some sort of benchmark as opposed to the "2.3x" metric since they show the baseline at the bottom.

What's the deal?

 

[iSee]

You can't draw that conclusion...

You can't really draw that conclusion based on these results without (at least) having the same benchmarks between the powerbook and the iMac G5. You're assuming the iMac G5 would have beaten the powerbook significantly, but we don't know that. For example, in the Doom test, perhaps the powerbook scored about as well as the iMac G5 (that benchmark is based more on GPU than CPU), so if the MacBook Pro is about as fast as the iMac Intel, then you would expect them to speed up by about the same factor.

It would be a lot better to simply have head to head benchmarks between the Intel iMac and Mac Book Pro. Then we could really tell what was going on...

 

[Anonymous Freak]

They don't say...

A few things.

They don't give the exact configuration of each computer. If you assume the 'stock' version of each, then the 1.83 GHz MacBook Pro would have 1 GB of RAM, while the 2.0 GHz iMac would only have 512 MB. And, again if we assume stock, the MBP would have 256 MB video, while the iMac would have 128 MB video. (Same basic chip, but the MBP has the 'mobile' version.)

Between those, the MBP should score pretty darned close to the iMac.

So then lets look at the older machines:

iMac G5 'stock' had 512 MB of DDR2, and Radeon X600XT/128 MB graphics. The PowerBook G4 had 512 MB of DDR2, and Radeon 9700/128 MB graphics. Hrm. Yeah. The iMac G5 should have crushed the PowerBook G4.

I guess we'll have to wait for Apple to release the actual raw benchmark numbers (they usually do,) or wait for MBP's to ship and see some real-world benchmark comparisons.

 

[verb]

iMac G5 'stock' had 512 MB of DDR2, and Radeon X600XT/128 MB graphics. The PowerBook G4 had 512 MB of DDR2, and Radeon 9700/128 MB graphics. Hrm. Yeah. The iMac G5 should have crushed the PowerBook G4.

Moreover the difference between the Powerbook's Radeon 9700 and the MacBook's X1600 is much larger than the difference between the iMac's X600XT and X1600XT. Therefore the difference in Doom 3 performance should have been much larger for the MacBook than the iMac.

Does that mean the MacBook is simply a very poor machine (comparitively) for gaming? Normally I'd say that it might be just something with Doom3, but I think Apple would want to pick the game with the most impressive performance increase ( from the small selection of Universal games ) or not include it at all.

While I would not consider myself a hardcore gamer, I do occasionally enjoy playing a computer game on my mac. It would be nice to be able to play newer games (even at reduced quality, resolution) for a few years.

 

[Anonymous Freak]

Does that mean the MacBook is simply a very poor machine (comparitively) for gaming? Normally I'd say that it might be just something with Doom3, but I think Apple would want to pick the game with the most impressive performance increase ( from the small selection of Universal games ) or not include it at all.

The other odd thing is that Doom 3 plays quite well on the iMac Intel, even using the PPC binary! (Because the vast majority of the speed comes from the GPU, no emulation is needed, it just passes the GPU instructions straight through.) I'd like to see a Doom 3: iMac G5/iMac Intel(Rosetta)/iMac Intel(Native) comparison.

 

[radiantm3]

Don't forget the difference in HD speeds as well.

 

[gekko513]

Moreover the difference between the Powerbook's Radeon 9700 and the MacBook's X1600 is much larger than the difference between the iMac's X600XT and X1600XT.

The Radeon 9700 is faster than the X600XT.

 

[Josh396]

Moreover the difference between the Powerbook's Radeon 9700 and the MacBook's X1600 is much larger than the difference between the iMac's X600XT and X1600XT. Therefore the difference in Doom 3 performance should have been much larger for the MacBook than the iMac.

Is the card in the new iMacs a X1600XT? I thought it was the mobile version used in the Macbooks.

 

[howesey]

Don't forget the difference in HD speeds as well.

Macbook = SODIMM DDR2 RAM
iMac = regular DDR2 RAM


The NB is different too, which can really make a difference. My dad has a laptop with a P4 3.0GHz EE, it is nowhere as fast as a desktop because if the NB and RAM.

 

[Hydra]

The Radeon 9700 is faster than the X600XT.

Radeon 9700 mobility is Radeon 9600 in reality and so is X600 too so they're all the same, just different clock speeds...

 

[powerbook911]

Macbook = SODIMM DDR2 RAM
iMac = regular DDR2 RAM

Nope, So-dimms in the iMac too. Yep, So-Dimms in the iMacs. Annoys me a bit actually.

 

[gekko513]

Radeon 9700 mobility is Radeon 9600 in reality and so is X600 too so they're all the same, just different clock speeds...

I think you've got it backwards. The Radeon 9700 is more like a slow clocked (low power?) Radeon 9800 than a Radeon 9600 according to this that I found just now and from other places I've read too.

The Radeon 9700 has twice as many pixel pipelines and twice as wide memory bus compared to the Radeon 9600, X600 and X1600.

 

[Anonymous Freak]

DDR2 is DDR2, and iMac uses same NB as MBP.

Macbook = SODIMM DDR2 RAM
iMac = regular DDR2 RAM


The NB is different too, which can really make a difference. My dad has a laptop with a P4 3.0GHz EE, it is nowhere as fast as a desktop because if the NB and RAM.

As PB911 says, the iMac has SO-DIMMs, too. And there is no difference between SO-DIMMs and regular DIMMs of the same technology and speed. It's just a different physical form factor. So DDR2 667 on a SO-DIMM is exactly the same as DDR2 667 on a DIMM.

And as the iMac teardown shows, the iMac uses the same chipset that the MBP will use. How do I know? Because at present, there is only one chipset that supports the Core Duo. Intel's 945xM. There are two main versions, one with integrated graphics (945GM,) one without (945PM.) That's it. And we know that the MacBook Pro will use ATI graphics, so there's no reason to use the integrated graphics model... (Which, when paired with an external video chipset, is just as fast as the discrete, so it doesn't really matter which Apple uses.)

As for your dad's laptop? Well, notebook manufacturers can use desktop chipsets in a laptop if they feel like it. I mean, the Extreme Pentiums are desktop processors, so why wouldn't the company use a desktop chipset? He just chose poorly.

 

[spencecb]

You can't really draw that conclusion based on these results without (at least) having the same benchmarks between the powerbook and the iMac G5. You're assuming the iMac G5 would have beaten the powerbook significantly, but we don't know that. For example, in the Doom test, perhaps the powerbook scored about as well as the iMac G5 (that benchmark is based more on GPU than CPU), so if the MacBook Pro is about as fast as the iMac Intel, then you would expect them to speed up by about the same factor.

It would be a lot better to simply have head to head benchmarks between the Intel iMac and Mac Book Pro. Then we could really tell what was going on...

I'm afraid we can assume that an iMac G5 would beat the pants off of a PowerBook. There really is no comparison.

 

[mongoos150]

Indeed, just as the Intel iMac will beat the pants off of the iMac G5

 

[verb]

I think you've got it backwards. The Radeon 9700 is more like a slow clocked (low power?) Radeon 9800 than a Radeon 9600 according to this that I found just now and from other places I've read too.

The Radeon 9700 has twice as many pixel pipelines and twice as wide memory bus compared to the Radeon 9600, X600 and X1600.

I've found these debates to be completely fruitless since nobody can remember precisely what the differences between the 9700, 9700 Pro, 9700XT and the X800, X800 SE, X800 GT, X800 Pro, X800 GT, X800 GTO, X800 XL, and X800 XT are much less all the deritivies of the 9600, 9800, X300, X600, X700, X800, X850, X1300, X1600 and X1800 GPUs. It gets even more complicated once you start comparing in NVidia's offerings.

Does anyone have a simple Quake 3 benchmark from a 1.67Ghz Powerbook G4 and a 2Ghz iMac G5 (or similar)? That would seem a more concrete way to full understand Apple's claims.

 

[gekko513]

I've found these debates to be completely fruitless since nobody can remember precisely what the differences between the 9700, 9700 Pro, 9700XT and the X800, X800 SE, X800 GT, X800 Pro, X800 GT, X800 GTO, X800 XL, and X800 XT are much less all the deritivies of the 9600, 9800, X300, X600, X700, X800, X850, X1300, X1600 and X1800 GPUs. It gets even more complicated once you start comparing in NVidia's offerings.

Does anyone have a simple Quake 3 benchmark from a 1.67Ghz Powerbook G4 and a 2Ghz iMac G5 (or similar)? That would seem a more concrete way to full understand Apple's claims.

Did you even look at the link I gave?

 

[Mord]

the desktop 9700 is an r300 core thus has 8 pixel pipes, the 9700 mobility is a higher clocked RV360 thus is like a mobile 9600XT, the x600 is a much higher clocked 9600 with a few architectural differences, the x600XT should have about a 20-30% speed advantage over the 9700 mobility.

ati confused the **** out of people with the mobility 9xxx line, the mobility 9800 was really an 8 pipe x800.

 

[ewinemiller]

the desktop 9700 is an r300 core thus has 8 pixel pipes, the 9700 mobility is a higher clocked RV360 thus is like a mobile 9600XT, the x600 is a much higher clocked 9600 with a few architectural differences, the x600XT should have about a 20-30% speed advantage over the 9700 mobility.

ati confused the **** out of people with the mobility 9xxx line, the mobility 9800 was really an 8 pipe x800.

You can compare the specs of various cards here

http://www.gpureview.com/show_cards.php

Some vendors will tweak clock speeds (ie. can't assume a 9700 in an Apple is clocked exactly the same as a 9700 in a Dell), but it's usually pretty close.

 

[gekko513]

the desktop 9700 is an r300 core thus has 8 pixel pipes, the 9700 mobility is a higher clocked RV360 thus is like a mobile 9600XT, the x600 is a much higher clocked 9600 with a few architectural differences, the x600XT should have about a 20-30% speed advantage over the 9700 mobility.

ati confused the **** out of people with the mobility 9xxx line, the mobility 9800 was really an 8 pipe x800.

I see, yeah, I'm confused too now.

You can compare the specs of various cards here

http://www.gpureview.com/show_cards.php

Some vendors will tweak clock speeds (ie. can't assume a 9700 in an Apple is clocked exactly the same as a 9700 in a Dell), but it's usually pretty close.

Excellent link. It seems that the Mobility Radeon 9700 has just 4 pipelines, like the 600 series, but has a 256 bit wide memory bus.

 

[howesey]

As PB911 says, the iMac has SO-DIMMs, too. And there is no difference between SO-DIMMs and regular DIMMs of the same technology and speed. It's just a different physical form factor. So DDR2 667 on a SO-DIMM is exactly the same as DDR2 667 on a DIMM.

And as the iMac teardown shows, the iMac uses the same chipset that the MBP will use. How do I know? Because at present, there is only one chipset that supports the Core Duo. Intel's 945xM. There are two main versions, one with integrated graphics (945GM,) one without (945PM.) That's it. And we know that the MacBook Pro will use ATI graphics, so there's no reason to use the integrated graphics model... (Which, when paired with an external video chipset, is just as fast as the discrete, so it doesn't really matter which Apple uses.)

As for your dad's laptop? Well, notebook manufacturers can use desktop chipsets in a laptop if they feel like it. I mean, the Extreme Pentiums are desktop processors, so why wouldn't the company use a desktop chipset? He just chose poorly.

You know, I was saying this for ages, but all people could do was flame me and say I was incorrect.

 

[verb]

Conclusion?

I've reread all the posts a few times, but I cannot gather a conclusion. Using the Video Card Comparison I see that the video card on the iMac ( X600 XT) is more powerful than the Mobility Radeon 9700.

The only thing that would account for the info on the Apple page is that the iMac is NOT using the Mobility version of the X1600 and that accounts for the large difference in the performance change.

 

[howesey]

X600 (RV380) is based on the Radeon 9600, which to tell you the truth, is awful. I bought a Radeon 9800 Pro a good two years ago, which is a lot (I mean a lot) faster than the 9600. Only thing that bugs me with the iMac G5 is it's rubbish GPU.

 

[gekko513]

I've reread all the posts a few times, but I cannot gather a conclusion. Using the Video Card Comparison I see that the video card on the iMac ( X600 XT) is more powerful than the Mobility Radeon 9700.

Not by much. The GPU might even be slower if you base it on these benchmarks and consider that the iMacs have faster processors.

Doom 3 Benchmarks
iMac 2.1 GHz - 29 fps
iMac 1.9 GHz - 25 fps
PowerBook 1.67GHz - 26 fps

Unreal Benchmarks
iMac 2.1 GHz - 66 fps
iMac 1.9 GHz - 56 fps
PowerBook 1.67GHz - 45 fps

I gathered the numbers from two different barefeats benchmarks. Here's one of them. It shows the recent PowerBook versus the older iMac G5. I found another benchmark for the newer iMac G5s.

 

[carlos700]

9700XT

There was never a 9700 XT.

On a different note both the ATI MR 9700 and ATI R X600 both have 4 Pixel Pipelines whereas the ATI R X1600 has 12 Pixel Pipelines. So there would be a BIG difference between the two.