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

rhb

macrumors newbie
Original poster
Oct 5, 2009
22
0
Hey all -

We have a few mid-2009 MBAirs lying around the office (C2D 2.13GHz SSD, 2Gb Ram, etc.), and have run a few tests on Lion vs SL that you might find interesting. These are developer machines, so we typically change out OS configs and app installations to see how the hardware performs under different environments.

Recently we did a simple check of Snow Leopard (10.6.8) vs Lion (GM Dev preview), and got a weird result. On a clean install of SL with nothing added to the box except Geekbench, the MBAir got scores in the 2600-2700 range, every time. But the same machines, running clean installs of Lion GM, got consistent scores of 1800-1900.

Now, we’re not hardware geeks around here, so I’m not sure I can give you an intelligent interpretation of those tests. The point here, though, is that we’ve read in many places that a 2Gb RAM MBAir runs Lion beautifully... but in our case, that’s clearly not happening. Or, it’s sort of happening; the OS runs, but the machine slows down, and we get lots of fan spinup and kernel-task process overloads.

So our (admittedly anecdotal) conclusion -- watch out for Lion on current MBAirs, or at least the 2.13 SSD revs. Maybe someone here with more hardware savvy can comment on what’s going on, or if our test results are just anomalous to the real truth...!:confused:
 

2IS

macrumors 68030
Jan 9, 2011
2,938
433
Little confused. Initially you thought it was the 2GB RAM causing the performance issue but you concluded with cautioning against the processor.
 

rhb

macrumors newbie
Original poster
Oct 5, 2009
22
0
Little confused. Initially you thought it was the 2GB RAM causing the performance issue but you concluded with cautioning against the processor.

Yep -- we're a little confused too. Not sure where the bottleneck is, so it's hard to say what's making Geekbench choke under Lion. Just wondered if anyone here has any insights that we don't have.
 

2IS

macrumors 68030
Jan 9, 2011
2,938
433
Could be both... Have you checked memory usage in activity monitor?
 

rhb

macrumors newbie
Original poster
Oct 5, 2009
22
0
Maybe Geekbench is not fully compatible with Lion?

You might be right -- but with the new Airs slated to drop in the next week or so (reportedly), I'd assume that the Geekbench guys anticipate that every reviewer alive is going to test Lion as we did. FWIW, we used the latest version of Geekbench --- off the top of my head, I don't know how 'current' it is.
 

rhb

macrumors newbie
Original poster
Oct 5, 2009
22
0
Could be both... Have you checked memory usage in activity monitor?

Yeah we did -- and it didn't show anything out of the ordinary. The CPU showed major chugging, though.
 

2IS

macrumors 68030
Jan 9, 2011
2,938
433
Interesting. I've got the 1.86GHz and 4GB of ram. Hope it runs Lion well as I really want to take advantage of some of the new features. My practical side says to wait until others try it first, but for $30 the impulsive side is probably going to triumph.
 

Duke15

macrumors 6502
May 18, 2011
332
0
Canada
Thats crazy, I too would like to know what the reason for this is, seems like lion is getting mixed reviews, should be interesting to see what its like once its released, and what its like on the MBAs
 

KnightWRX

macrumors Pentium
Jan 28, 2009
15,046
4
Quebec, Canada
Could be both... Have you checked memory usage in activity monitor?

Geekbench is a CPU synthetic benchmark. Numbers will vary depending on how many other processes are eating CPU ressources at the time the benchmark is run.

It has nothing to do with memory usage.

It's also a poor benchmark, unless you really don't know that your 2.13 ghz Sandy Bridge processor is slower than a 2.5 ghz Sandy Bridge processor.
 

Eidorian

macrumors Penryn
Mar 23, 2005
29,190
386
Indianapolis
Geekbench has exhibited some weirdness based on operating systems in my experience. It is supposed to be relatively agnostic test of performance but the scores have varied significantly on the same hardware based on what operating system you are running on it. That really bothers me.
 

2IS

macrumors 68030
Jan 9, 2011
2,938
433
Geekbench is a CPU synthetic benchmark. Numbers will vary depending on how many other processes are eating CPU ressources at the time the benchmark is run.

It has nothing to do with memory usage.

It's also a poor benchmark, unless you really don't know that your 2.13 ghz Sandy Bridge processor is slower than a 2.5 ghz Sandy Bridge processor.

I understand that, I'm just wondering what the memory usage is with the OS itself. (without Geekbench running) I'm also curious as whether or not the CPU chugging was only during the benchmark, which sounds reasonable, or if it was being taxed during normal use as well.

BTW, I think you mean 2.13GHz Core 2 Duo
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.