I've been looking at the 15" 2.4 Macbook Pro and after doing some research and looking at benchmarks I've noticed that the 2760QM and 2860QM in the current Macbook Pros are only about 10% slower than the i7 2600 desktop processor in terms of rendering, video encoding and other processor intensive tasks. I'm just wondering if anybody has any hands on experience with both the desktop version and the mobile version and can confirm similar performance. Considering the mobile chip has less than half the power draw of the desktop version I was a bit surprised to see this. Also, are there any issues with running the current Macbook Pros heavily for extended periods of time? I was originally planning on getting a desktop and a less powerful laptop but now I'm thinking I can just get a Macbook Pro and have the best of both worlds in one machine. Any help or advice is appreciated. Thanks
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Intel Mobile CPU's vs Desktop CPU's???
- Thread starter jumpman25
- Start date
- Sort by reaction score
Macbook Pros are awesome. You can have the best of both worlds. There are no problems with running intensive tasks on the mobile CPUs for extended periods of time.
Note however:
CPUs get hot - with a laptop, you will have the heat below your fingertips and the fan noise right on the table. For some people this is inconvenient or uncomfortable. The newer MBPs do a very good job of keeping the palm rests and keyboard area cool, but it's still noticeable - in particular in the summer.
Note however:
CPUs get hot - with a laptop, you will have the heat below your fingertips and the fan noise right on the table. For some people this is inconvenient or uncomfortable. The newer MBPs do a very good job of keeping the palm rests and keyboard area cool, but it's still noticeable - in particular in the summer.
Yes the MacBook Pro quad core processors are very close the desktop processors from Intel. The laptop processors differ in that they have very aggressive power management to keep the heat down.
You'll notice that the i7-2600K is around $325 whereas the i7-2860QM is over $550. So you can get close to desktop like performance, only it will cost you in both money and a lot heat used (and therefore shortened battery life).
You'll notice that the i7-2600K is around $325 whereas the i7-2860QM is over $550. So you can get close to desktop like performance, only it will cost you in both money and a lot heat used (and therefore shortened battery life).
I think your calculations are a little off. 3.4 vs 2.4 Ghz is about 40% difference in speed.
If you run really short benchmarks (30s) where the turbo still has enough influence it is a small difference, but with serious encoding jobs (a few minutes) the default clocks matter and there is a rather big difference.
Unless of course you compare to the 2600S with just 2.8Ghz but that is a in between thing and not really a desktop vs mobile comparison.
Also overclocking on Desktops works very well 4+Ghz default is usually no problem on air. There is no way to overclock a mobile CPU.
Still compared to the difference in graphics performance mobile and desktop CPU performance is fairly close.
If you run really short benchmarks (30s) where the turbo still has enough influence it is a small difference, but with serious encoding jobs (a few minutes) the default clocks matter and there is a rather big difference.
Unless of course you compare to the 2600S with just 2.8Ghz but that is a in between thing and not really a desktop vs mobile comparison.
Also overclocking on Desktops works very well 4+Ghz default is usually no problem on air. There is no way to overclock a mobile CPU.
Still compared to the difference in graphics performance mobile and desktop CPU performance is fairly close.
Last edited:
Thanks for all the replies. I appreciate the advice. With regard to the turbo boost, I thought it kicked in when the processor was under load and didn't have a limit as long as temps were ok. Is there really a time limit on the turbo boost?
I think this is indeed how it works. The 2860qm can run on 3.3 GHz with four cores as long as the cooling is sufficient. This is the limiting factor on laptops. On the mac, once the discrete GPU is also fully loaded, cooling is barely sufficient for the normal frequency.
That is a misinterpretation. If thermal was all that matters a sufficiently cooled system like a gaming notebook or a desktop would never run anything but the Quad Core max Turbo frequency for any given time.
Temperature is just one of the limits that can keep it from dropping back to normal operating frequency sooner.
After that burst it goes back to default or hovers slightly above default speed. Like 200-300 with occasionally trying to speed up again a little.
If it was running above TDP for a long time it would destroy battery life.
The speed it can run for a longer time also depends on the workload. There is a lot of logic in a CPU and rarely all is busy. If a workload utilizes less excution units it will create less heat and require less current than some other workload and it clocks a little higher. Other stuff with lots of SSE and what not will not let it run any faster than the default clock.
Also you need to remember if it was only about heat or the true max TDP a 2675qm would in almost all cases yield the same performance as a 2860qm for a long encoding task. Since the latter needs more current, a bit more tdp a bit more power at the higher clocks. In the end they would both end up running at close to the same speeds say 2.6Ghz for all 4 cores. That is not the case though.
Also the CPU needs to go back down to its in built "TDP" rating which is most likely not 45W but something different for the 2675 than the 2860 and below TDP in any case is also lower than the max quad core turbo which is probably at around 55W or more.
Temperature is just one of the limits that can keep it from dropping back to normal operating frequency sooner.
There are some youtube videos that show how it actually behaves. The max turbo really only kicks in for a very short period of time. Enough to launch an app but it only last for some 10 seconds or a little more. If you run a very short benchmark it greatly misrepresents performance.
After that burst it goes back to default or hovers slightly above default speed. Like 200-300 with occasionally trying to speed up again a little.
If it was running above TDP for a long time it would destroy battery life.
The speed it can run for a longer time also depends on the workload. There is a lot of logic in a CPU and rarely all is busy. If a workload utilizes less excution units it will create less heat and require less current than some other workload and it clocks a little higher. Other stuff with lots of SSE and what not will not let it run any faster than the default clock.
Also you need to remember if it was only about heat or the true max TDP a 2675qm would in almost all cases yield the same performance as a 2860qm for a long encoding task. Since the latter needs more current, a bit more tdp a bit more power at the higher clocks. In the end they would both end up running at close to the same speeds say 2.6Ghz for all 4 cores. That is not the case though.
Also the CPU needs to go back down to its in built "TDP" rating which is most likely not 45W but something different for the 2675 than the 2860 and below TDP in any case is also lower than the max quad core turbo which is probably at around 55W or more.
What is he max frequency on one core? I thought that the max turbo boost was 3.3ghz on one core, not all four. I believe the max turbo on the processor with all four cores active is somewhere in the 2.8 or 2.9 ghz range.
On wikipedia, the 2860qm (2.5 GHz i7) is listed with a 8/8/10/11 turbo. This would give max turbo of 3.3/3.3/3.5/3.6 GHz for 4/3/2/1 cores.
Yes this is true. In winter when the ambients are low, I can get all 4 cores past 3ghz for a prolonged period of time but in summer I'll be lucky if they can run at 2.2ghz.
Turbo boost was designed to give short burst of speed for smaller tasks to seem faster, don't expect it to run continuously but it can if the conditions are right.