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

Why would this happen? Processor Speed

W

wwchris

macrumors regular
Original poster
Nov 13, 2009
157
178
Atlanta, GA
My son and I both have 15-inch Macbook Pro's. Mine is an Early 2011 and his is a late 2011.

He has been complaining that his machine was slower than his friend's similar machine and I told him I assumed it was software. He said I was wrong.

I installed a fresh version of Mavericks on an external Thunderbolt drive and ran geek bench 64 bit on his and mine started from the same fresh install, external drive.

We both have 16 GB of RAM, but his is faster (1333 vs 1600). He has a faster Processor (2.2ghz vs 2.4ghz same gen) and he has a better GPU (6750 vs 6770 both with 1 gb).

He gets better single processor speed, but I am getting better multiprocessor speed. Seems like his machine is about 10% slower overall.

Anybody able to explain the these results?
 

Attachments

  • macbookcompare.jpg
    macbookcompare.jpg
    220.5 KB · Views: 360
Gav Mack said:
What's the exact cpu models on both. Check temperatures using istat running geekbench too may give us a clue too.
Click to expand...

If you click the images, you can see the processor types. They are both Quad core i7's of the same generation. Heat might be a part of it, but these machines were both cooled down before testing and it seems unlikely to me that heat would be a factor that fast (the test is less than 2 min and only gets the fans spinning for like 45 seconds). I'll see if I can get the heat info though.

Looking at geek bench's website, seems like maybe 5% of people with the same machine have similar performance. Could open the machine and use canned air on it and see if it is dust or something if it is hot. I dunno. Just seems weird to have a 10% performance difference only on the multicore.
 
maybe background processes caused the issue. The OS is doing things while these benchmarks run, even if all the other programs are closed. Like, one of the computers might have been indexing something for spotlight. To be more accurate, you'd need to run them both in safe mode, or somehow boot into the benchmark from a livecd or something.
 
Freyqq said:
maybe background processes caused the issue. The OS is doing things while these benchmarks run, even if all the other programs are closed. Like, one of the computers might have been indexing something for spotlight. To be more accurate, you'd need to run them both in safe mode, or somehow boot into the benchmark from a livecd or something.
Click to expand...

Well, I waited till spotlight was done indexing and let the machine run for 5 min to make sure all startup functions had completed. I ejected all the other drives so they wouldn't index and added them to the spotlight exceptions.
 
wwchris said:
Well, I waited till spotlight was done indexing and let the machine run for 5 min to make sure all startup functions had completed. I ejected all the other drives so they wouldn't index and added them to the spotlight exceptions.
Click to expand...

That was just one possible background process.

Another possibility is that one computer is throttling faster due to heat. Maybe the rooms that they were benchmarked in were vastly different temperatures, or some other reason. He is running faster clocked ram and has a higher clocked CPU, so that could run a bit hotter. It could even be different drivers installed. Did you repair disk permissions recently? Maybe there is some permissions issue that is eating clock cycles. There are plenty of possible reasons to account for.
 
I think I saw a post on the primatelabs forum once. Someone had the same problem (only with iMac vs Macbook) and the conclusion was, that geekbench, after an OS Update does behave differently on different machines. Did you do an Update or a Clean Install?
 
If you have both side by side you could do some heat-debugging on the side, open up terminal and enter

pmset -a thermlog
Click to expand...

and start the benchmark. At some point the OSX probably starts throttling the CPU, could be that the Late 2011 MBP gets throttled earlier, thus a lower Geekbench result...
worth a try at least.
 
matthewadams said:
If you have both side by side you could do some heat-debugging on the side, open up terminal and enter



and start the benchmark. At some point the OSX probably starts throttling the CPU, could be that the Late 2011 MBP gets throttled earlier, thus a lower Geekbench result...
worth a try at least.
Click to expand...

I have never seen geekbench cause a Mac's CPU to throttle, even on a lowly, little MBA.
 
Some of it solved!

Thanks for all the suggestions, very helpful! You guys are great!
So I was able to solve some of the issues.
I feel kinda stupid for not trying these first now. Rookie Move :)
So, I zapped the PRAM and reset the SMC.
Then, we were thinking about thermal paste and in a different thread, I found a guy who had pics of his fans filled with dust on the inside where you couldn't see.

https://discussions.apple.com/thread/2764570?start=840&tstart=0

Sure enough a bunch of the airflow was blocked with Dust bunnies. Removed the fans and cleaned inside and the vents behind the fans. (I do wish the guy who posted the photos had just taken out the screws and cleaned those :) )

My son's single CPU performance went up to 3000 and his multicore went up to 10000. Still slightly below mine, but almost a 10% improvement (which in this instance is close enough).

Again, thanks for the suggestions!
 
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