Skip to Content

Benchmarking results: Is Snow Leopard really any faster than Leopard?

Be sure to check all of our ongoing Snow Leopard coverage right here.

One of the biggest features of Snow Leopard isn't something apparent to the naked eye: software tweaks and refinements intended to make OS X a leaner, meaner OS for your fighting Apple machine. But is Snow Leopard really any faster? Now that I've successfully upgraded two Macs to Snow Leopard I've got some benchmarking results to share.

My Early 2008 MacBook Pro shipped with OS X Leopard 10.5.2 installed. I ran Geekbench on the stock OS X installation after upgrading the RAM to 4 GB to get a baseline for comparison of future performance. 18 months later I ran the same test immediately after updating to 10.6. Both tests were performed with Geekbench testing in 32-bit mode immediately after a restart, with no other programs open except the Finder, nothing loaded in Dashboard, and no Time Machine backup running.



Machine specs:

Intel Core 2 Duo @ 2.60 GHz w/ 4GB RAM

Average Overall Geekbench score for this model of MacBook Pro: 3304

Read on for the scores.Machine specs:

Intel Core 2 Duo @ 2.60 GHz w/ 4GB RAM
Average Overall Geekbench score for this model of MacBook Pro: 3304

Read on for the scores.

10.5.2 results:

Overall Score: 3364

Integer: 2965
Floating Point: 4785
Memory: 2412
Stream: 1692

10.6 results:

Overall Score: 3511

Integer: 3082
Floating Point: 5055
Memory: 2467
Stream: 1701

The benchmarking gains are small, particularly in memory and stream performance, where the gains are negligible and most likely just a statistical deviation. Overall performance has gone up by just over 4%, with the biggest performance gains seen in floating point operations, which increased by 5%.

My wife's late-2006 MacBook saw similar gains in performance, but with her machine I have three sets of values to compare: performance with a base installation of 10.5.2, with 10.5.8, and with a brand-new 10.6 install.

Machine specs:

Intel Core 2 @ 2.16 GHz w/ 2GB RAM
Average Overall Geekbench score for this model of MacBook: 2741

10.5.2 results:

Overall Score: 2773

Integer: 2451
Floating Point: 3894
Memory: 1975
Stream: 1577

10.5.8 results:

Overall Score: 2849

Integer: 2365
Floating Point: 4249
Memory: 1907
Stream: 1528

10.6 results:

Overall Score: 2977

Integer: 2444
Floating Point: 4484
Memory: 1975
Stream: 1577

Once again there's no gain in memory or stream performance, but measurable gains in overall scores and floating point performance. Between 10.5.2 and 10.6, there's a jump of 7.4% in overall score and a significant increase of 15% in floating point performance.

Joachim Bean's Geekbench scores on his Mac Mini showed a similar 3-5% gain in performance just by upgrading to Snow Leopard.

While the performance gains aren't huge, they do seem to reflect the optimizations present in Snow Leopard. And this is really just the tip of the iceberg when it comes to Snow Leopard's potential. Once developers start taking advantage of new technologies like Grand Central and OpenCL, applications in OS X will be screamingly fast.

Categories

OS TUAW Labs Snow Leopard

Be sure to check all of our ongoing Snow Leopard coverage right here. One of the biggest features of Snow Leopard isn't something apparent...
 

Add a Comment

*0 / 3000 Character Maximum

61 Comments

Filter by:
Gustav Hermann

I'm disappointed in Snow Leopard, all hype and no bite sad to say. I run both Windows and Mac machines, the improvements from Leopard to Snow Leopard as far as speed is not "that" noticeable, programs I run regularly on OSX take the same time to start up as before the upgrade ..... so I ask WHERE is this "major" speed improvement that Apple was bragging about? Sure it's faster on startup and shutdown but really that means squat in my life as I only start up once a day and only shut down once a day, so where's the huge benefit in that ??? I could care less that time machine is 80% faster, I don't use it anyway.

Being a user of both Mac and Windows OS's for many years I am more than qualified to say the following:

Microsoft has totally outdone Apple this year with their release of Windows 7, Win7 is the first OS I have seen that "actually" runs much faster on existing hardware, noticeably faster. I think Apple missed the boat this year concentrating more on producing a fancy looking product box. Every single Windows PC we have upgraded to Windows 7 runs faster and better on the same hardware.

October 15 2009 at 5:11 AM Report abuse rate up rate down Reply
Kevin

Using an old copy of Geekbench 4.0 (to avoid differences in versions between benchmarks I've done on earlier machines), my 2009 Mini's overall score was 228 in 10.5, 235 in 10.6. Not a lot of improvement, but enough to notice some "snap" to things.

September 09 2009 at 12:36 PM Report abuse rate up rate down Reply
Martin

Please remove this article.

a response to a few questions in your comments:

why is it slightly faster ?
well, probably because benchmarks aren't reliable on a multitasking computer, unless they are not interrupted at all.

time machine isn't the only factor, a benchmark can be slowed down by spotlight updates, or any other task, even an open window in the Finder.

the author should at least have closed the finder, and all other unnecessary processes, before testing the hardware, at least he would have had a valuable result: that the hardware runs at the same speed as before, what a surprise !

but, comparing OS's with these tests and publishing an article about it is completely misleading, to say the least.

September 03 2009 at 11:08 AM Report abuse rate up rate down Reply
Martin

Bullshit, you aren't testing the OS here, there should actually not be a difference at all, or very very marginal.

to test if snow leopard is faster than leopard you should test the OS, the file system, memory management (not just access to ram, but memory allocation)

loading of apps, finder file copies etc !

what a stupid article !

September 03 2009 at 10:55 AM Report abuse rate up rate down Reply
jill

http://www.electroniccigarettesinc.com

September 03 2009 at 10:30 AM Report abuse rate up rate down Reply
jill

It Better Be!
electronic cigarettes

September 03 2009 at 10:21 AM Report abuse rate up rate down Reply
slinkypomo

Oh wow, that is like WAY cool dude. I like it!

RT
www.anonymous-web.be.tc

September 03 2009 at 9:06 AM Report abuse rate up rate down Reply
Jeff Weitzman

I'd have paid the $29 upgrade just for the significant decrease in shutdown time. That's the kind of real-world improvement that doesn't make for sexy ad copy, but nearly every day I'm rushing to make the shuttle to the train station, and those seconds seem like hours!

September 03 2009 at 3:03 AM Report abuse rate up rate down Reply
1 reply to Jeff Weitzman's comment
mindfeck

There's no need to shut off your computer every time you leave for work. Even if you insist on it, you don't have to watch it shut down. Overall, a 5% change in speed is pointless too.

September 03 2009 at 4:38 PM Report abuse rate up rate down Reply
gimmeslack12

Use Graphs to compare!!! Please!

September 03 2009 at 2:04 AM Report abuse rate up rate down Reply
ojfl

I can say that in our PowerMac and our MacBook it seems faster. It seems much faster on the PowerMac where applications spring to life faster and functionality also seems faster.

September 02 2009 at 4:29 PM Report abuse rate up rate down Reply
2 replies to ojfl's comment
gshearman

Your PowerMac? PowerMacs are PPC only and therefore can't run Snow Leopard. Maybe you meant Mac Pro?

September 02 2009 at 5:46 PM Report abuse rate up rate down Reply
ojfl

Correct dagaz. MacPro. My PowerMac is still lonely running Leopard. But it still runs great!

September 02 2009 at 7:47 PM Report abuse rate up rate down Reply
Buy an ad here

Hot Apps on TUAW

Tweets

© 2012 AOL Inc. All Rights Reserved.