Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

MPB i7 2.3GHz VS iMac i7 3.4GHz: Why such a small difference in CPU performance?

B

Broc27

macrumors member
Original poster
Mar 23, 2010
50
0
Hey there!

I've been wondering if anyone knows why a portable CPU can perform so well compared to its desktop counterpart even with such an important clock speed difference? I mean...

1. Geekbench result for 2.3 GHz Sandy Bridge i7 (2011 MacBookPro 17"): 10 383
2. Geekbench result for 3.4 GHz Sandy Bridge i7 (2011 iMac): 11 602

Any explanation?
 
I've seen a video of a guy doing the i7 3,4Ghz benchmarks on youtube and he got 12,8k out of it. Maybe you used the 32bit version.

Anyway i have to admit those mobile chips are amazing performers.
 
Broc27 said:
Hey there!

I've been wondering if anyone knows why a portable CPU can perform so well compared to its desktop counterpart even with such an important clock speed difference? I mean...

1. Geekbench result for 2.3 GHz Sandy Bridge i7 (2011 MacBookPro 17"): 10 383
2. Geekbench result for 3.4 GHz Sandy Bridge i7 (2011 iMac): 11 602

Any explanation?
Click to expand...

2.3ghz turbo mode 3.4/3.3/3.1 GHz
3.4ghz turbo mode 3.8 GHz

the 2.3 has a more aggressive turbo mode.
 
DucktapeFanboy said:
I've seen a video of a guy doing the i7 3,4Ghz benchmarks on youtube and he got 12,8k out of it. Maybe you used the 32bit version.

Anyway i have to admit those mobile chips are amazing performers.
Click to expand...

Yes, these are both 32-bit results since the geekbench early results for the iMac are limited right now.

I can believe a result of 12 800 out of the iMac but even then... Shouldn't we be talking about a 50% increase for the iMac, based on clock speed alone? I would have hoped for even more considering we are comparing laptop and desktop CPUs.
 
CPU core performance may not be the bottleneck. In fact, for many real world apps the CPU spends most of its time idle waiting for the data to arrive. So the memory and storage systems are the bottlenecks.
 
JeepGuy said:
2.3ghz turbo mode 3.4/3.3/3.1 GHz
3.4ghz turbo mode 3.8 GHz

the 2.3 has a more aggressive turbo mode.
Click to expand...

Isn't turbo mode activated only when not all cores are active? Geekbench is supposed to use all cores so Turbo Mode shouldn't factor that heavily in the equation, no?
 
Broc27 said:
Isn't turbo mode activated only when not all cores are active? Geekbench is supposed to use all cores so Turbo Mode shouldn't factor that heavily in the equation, no?
Click to expand...

I think that's not entirely true with the current version of turbo boost. But i can't provide technical details.
 
Broc27 said:
Isn't turbo mode activated only when not all cores are active? Geekbench is supposed to use all cores so Turbo Mode shouldn't factor that heavily in the equation, no?
Click to expand...

Turbo goes to a higher clock with fewer cores running, but still gets to 3.1GHz for the 2.3 quad with all cores. However, you will likely see a bigger performance increase than the benchmarks suggest, as the CPU is not tested under full load for long enough to heat up a lot. When it does heat up, the turbo boost frequency will drop closer to the base clock, or even down to the base frequency. The iMac has a more efficient cooling system, so turbo mode can be more active, plus at worst, it would drop to 3.4GHz, which is significantly faster than the 2.3GHz base for the MBP. If you did testing for transcoding a decently long video file, you would probably notice a bigger difference between the MBP and the iMac than benchmarks suggest.
 
I think the 20-25% improvement between last year's iMac and this year's is pretty good. We went from 2 cores/4 threads to four cores/8 threads in the MacBook Pro high-end and this is why GeekBench (and other testing) nearly doubled. Meanwhile, we are mainly increasing clock speed with a more efficient Sandy Bridge platform with the iMac.

In encoding tests, the iMac will be at least 10-15% faster than the high-end MacBook Pro (2.3GHz i7), while costing ~$550 less (17", 7200RPM drive, 2.3GHz) vs the 3.4GHz, 27", 7200RPM drive. The mobile processor is also nearly double the price of the i7-2600 in the iMac.

We'd likely need to go to a 6-core Sandy Bridge before seeing a more significant difference.
 
kevin2223 said:
I think the 20-25% improvement between last year's iMac and this year's is pretty good.
Click to expand...

I was not saying the iMac isn't an improvement or a good deal; it is. I was just wondering why it wasn't so much faster then the MBP. From what I read here they are pretty close for CPU-intensive jobs as long as those jobs aren't long winded; then heat becomes a problem for the laptop.

Is that right?
 
Last edited:
Yes. When utilizing Turbo Boost, the MacBook Pro processor gets near the base clock speed (3.4GHz) of the iMac (3.4/3.3/3.1 GHz at 1/2/4 cores).

What might make them even more similar (regardless of whether they are desktop or mobile CPUs) is the fact that they both have 8MB of L3 Cache, plus other similarities.

Right now, price and portability really separate the two, not really speed. With another 20% performance boost in Ivy Bridge; 8 cores at ~95 watts by Haswell, we should see the difference between the Macbook Pro and iMac become larger again.
 
Geek bench is a synthetic test and not always indicative of real world performance
 
theSeb said:
Geek bench is a synthetic test and not always indicative of real world performance
Click to expand...

I know, Geekbench only gives an idea of the relative maximum CPU performance of a computer compared to another. Real life differences are often even less important than what you see in Geekbench.

Considering that it is just even more surprising that the MBP is so close to the iMac.
 
The difference might even be smaller depending on the setup i've seen numbers of 12k+ being benched on some MBPs. I personally benched 11065 on my 2.2GHz model.

As for why this is...your guess is as good as mine.
 
Maybe the new imac is not fantastic compared to equivalent macbook pro

But the GPU is a lot better. Highest end mbp gpu is base model imac GPU, though half the memory
So anyday now i'd go with hi end imac and purchase basic sandy bridge mba later. That should fill your need perfectly, rather than spend all to macbook pro
 
Broc27 said:
I know, Geekbench only gives an idea of the relative maximum CPU performance of a computer compared to another. Real life differences are often even less important than what you see in Geekbench.

Considering that it is just even more surprising that the MBP is so close to the iMac.
Click to expand...

Not really, the way the core i processors work means in a short run CPU test like Geekbench what you're effectively seeing is the maximum possible CPU power. In the real world doing processor intensive tasks the mobile processor will drop down in speed as soon as the temperature starts to go up and the gap will rise. Frankly it's one reason I hate artificial tests like Geekbench, they just don't have much relevance to the actual performance the user will see.
 
Is it possible that geekbench would be updated soon to recognize 2011 iMac better? So iMac could score much much higher to their potential, maybe reach 14K - 15K?

Because I personally feel the score a bit too low for this kind of desktop processor. Even 12000 score is the custom i7 BTO .. The standard i5 3.1Ghz don´t even reach 10000. What the ...
 
I have seen iMac 27' 3.4 i7 ghz score 11k in geekbench.

But I seen other computers ( PC's ) score 17k in geekbench, so the desktop CPU can perform really good.
 
I just tested my MacBook Pro 15" Quad Core i7 2.2 GHz with 8 GB of 1333 Mhz RAM and I got 10099 with GeekBench 64 bit. Pretty good scrore. I have installed Windows 7 in BootCamp and I can run Crysis 2 and Battlefield Bad Company 2 without any problems with graphics set to high.

But, today I just ordered the 17" MacBook Pro with 2.3 GHz and high res anti reflective screen :) But I have to wait 2-4 weeks since they don't have it in stock in my local Apple store. But it'll be worth the wait :)
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back