OSX 10.6 Snow Leopard Preview
Now that we’ve had a chance to really sit down and experience Apple’s newest cat, its time to share our initial reactions to OSX 10.6 - Snow Leopard. Right off the bat, its fast. Very fast. A clean installation took about 13 minutes from start to finish, which is a world of difference from the hour or so that a clean 10.5 Leopard install takes. This has been potentially attributed to the significant reduction in the size of core applications (Snow Leopard’s Mail application is under 100mb, whereas 10.5’s was nearly 300mb). Other theories involve a removal of PowerPC support freeing up space and speeding up the install time. Other than the actual time involved, there are very few differences between the 10.5 and 10.6 installation processes in the current build.
Once you’re up and running, it feels very similar to Leopard. I don’t know how much this will change through Snow Leopard’s development, but don’t expect a terribly different interface. The subtle changes to the current Aqua definitely look good though. The biggest changes are under the hood. Snow Leopard is fast. Very fast. Like, surprisingly fast. From boot times to general application usage, Snow Leopard was noticeably quicker than Leopard when using the same system. Apple and 3rd party applications alike, they all launched faster and performed smoother. I’m sure this can be attributed to the new 64-bit architecture, but its amazing how much of a difference it really is.
But I digress, as I’m sure you’re bored of all this text - lets go on to the good stuff! The pictures:
About This Mac:
Nothing to see here, just the standard “About This Mac” screen. Click for full screen.
Install Weight
This was very interesting, though probably evident by the quick install. By default, the System folder of Snow Leopard is nearly 1 gigabyte smaller than that of Leopard (10.5). Optimizations have clearly been made.
Now on to the software changes. Right off the bat, there were new (unreleased) versions of many of Apple’s “staple” softwares. Quicktime, Bootcamp, Mail, even the Address Book have seen updates. Address Book’s most notable feature is the Exchange Server integration so happily boasted by Apple at WWDC.
Software Update
By replacing the in-your-face popup window that was Leopard’s software update with the preferences panel, they took everything I hated and made it into a cleaner, friendlier interface. Bravo!
Boot Camp 3 Beta
This one was an “expected surprise”. I haven’t had a chance to play with the new drivers, but the current version 2 is already so polished that I doubt any major changes have been made.
32 Bit Mode
This is most likely a developer mode “feature” rather than a poor interface decision, but many of the system preference options require “restarting” into 32 bit mode. This is instantaneous (it simply closes system preferences 64 bit and opens 32 bit) but annoying.
Exchange Integration
Highly trumpeted, but nice to see anyways. Lacking a live Exchange Server, this feature went untested.
Quicktime 7.6
I was expecting to see the highly anticipated Quicktime X, but was instead greeted with the unreleased Quicktime 7.6
Thanks for looking! This post will be updated continually as our experience with Snow Leopard progresses. As of June 25th, we have experienced 0 crashes.
Tags: mac os x 10.6, mac os x snow leopard, mac osx snow leopard, mac osx snow leopard review, os x snow leopard review, osx 10.6, osx snow leopard Posted in
June 26th, 2008 at 5:31 am
Have you run any pro adobe/apple apps? is it stable? worth experimenting with?
June 26th, 2008 at 2:58 pm
Why wouldn’t you even bother using ANY pro apps with a Beta of 10.6
That spells trouble. Just wait until it is released to experiment with it.
June 26th, 2008 at 3:15 pm
I agree with Vito.
I think its only worth testing Snow Leopard if you have the intention of developing applications for the new OS, not if you’re just looking for a speed increase over 10.5 Leopard. I say this because while I encountered no crashes, I would be very wary of running critical applications (like Adobe’s Photoshop, etc) for fear of losing my work.
June 26th, 2008 at 6:03 pm
[...] has published a quick look at Mac OS X Snow Leopard. “The biggest changes are under the hood. Snow Leopard is fast. Very fast. Like, surprisingly [...]
June 26th, 2008 at 6:26 pm
Can you post some timed differences? I know you’ve stated it’s faster but comparison numbers would be appreciated!
June 26th, 2008 at 7:18 pm
Actual numbers is step 2. I need to get my hands on an accurate digital timer before I can record boot times. As of now, everything is perceived speed
June 26th, 2008 at 7:28 pm
“I’m sure this can be attributed to the new 64-bit architecture, but its amazing how much of a difference it really is.”
This is not just a hypothesis, but you are sure. 64 is more than 32, right? Why not produce 128bit processors? Or 256bit?
There is no reason why 64bit computing should be generally faster. Sure, some scientific apps will be benefited but not your iTunes or iWhatever.
June 26th, 2008 at 7:34 pm
AnonymousCoward, the name suits you well - (that’s actually what he put). You clearly have no understanding of what the 64 bit architecture provides.
When the entire operating system operates on a 64 bit architecture, the entire system speed is improved. This is because the components used to draw new windows, to create the graphic interface (dock, menu et al) can now fully utilize the 64 bit processors (the Core 2 Duo and Xeon processors in new model Intel Macs).
Geekbench testing reports a 6% increase with nothing changed between identical systems running 32 bit and 64 bit OS X. If that’s not a general increase in speed, I’m not sure what else to say.
June 26th, 2008 at 9:41 pm
OK, can you explain me what the 64bit architecture provides and what justifies a 6% increase in a typical workload?
June 26th, 2008 at 11:36 pm
Hey guys!
Remember that 64 bit architecture helps in a major way by allowing larger memory addressing, speedier execution as well in some cases. Only problem is that the binaries of 64 bit systems
are larger than the 32 bit counterparts.
June 26th, 2008 at 11:40 pm
I have to say AnonymousCoward is more correct in this instance. There is no reason you should see speed increases in 64-bit other than some multimedia (encoding), compressing and scientific apps. In fact, all other applications should see a slight DECREASE in speed due to the larger addresses.
There have been fully 64-bit operating systems for a long time now. For a modern case look at linux where every app is most definitely 64-bit and yet speed increases are only seen in applications that make use of the 64-bit registers such as your encoding apps and compression apps.
June 27th, 2008 at 12:59 am
IIRC, the only cause of the generalized 64-bit migration performance benefit has to do with internal kernel utilization of 64-bit values on 32-bit CPUs.
Meaning, the kernel was faking 64-bit integers which were used in some performance-critical areas due to necessity.
One prime example involves timing and thread scheduling. Any improvements with these areas will be noticeable ( via benchmarking, at least ).
Just my $0.02 ( which is only worth $0.0025 by 1998’s standards ).
–The loon
June 27th, 2008 at 1:21 am
The x64 architecture not only offers 64-bit addressing and registers, but it also extends the register set of x86 from 8 registers to 16 (*). This by itself helps quite a bit to reduce the number of accesses to memory and to increase the speed.
(*) okay, that is oversimplified and doesn’t count special purpose, floating point and SSE registers.
June 27th, 2008 at 1:26 am
Anyone heard anything more on SSE4 from the Intel side of things? I was curious to know how that compared to the old Altivec encoding from Freescale. I think that by itself makes a bigger difference than x86 vs. x64, but I cant really say that for sure, it would be interesting to see apple make use of the new encoding as they did with Altivec!
June 27th, 2008 at 4:07 am
[...] Full Preview [...]
June 27th, 2008 at 7:20 am
Apple has finally learned how to use shared libraries! Welcome to the Unix-like world, fellas.
I must agree that 64-bit processors themselves do not increase speed very much. For some processor-intensive operations, they increase speed. For some, they decrease speed - it evens out.
I’d suggest two factors that have really increased the speed in this article:
1. Doing a clean install of the operating system. Boot time is dependent mostly on disk speed, so if you defragment the operating system on disk, you get faster bootup and faster accessing of linked libraries.
2. Now you’re running 64-bit native, you have effectively an extra 700-800 megabytes of RAM available for non-kernel cache.
June 27th, 2008 at 8:32 am
64-bit architectures per-se don’t increase speed at all. However, in the specific case of x86_64, there are more registers usable in 64 bit mode than in 32 bit mode, and I believe there are some minor differences in the instruction set that as a side effect provide nice speed boosts.
Basically the designers of x86_64 cleaned up and added a few nice extras that causes the 6% performance increase.
June 27th, 2008 at 9:21 am
There is one reasons why 64 bit apps are faster, and there is only one and only existing on intel processors, it is simply the fact that AMD which designed the 64 bit mode modern intel processors have added a load of general purpose registers in that mode. A thing Intel simply ignored for decades. It is still less than the number of general purpose registers the ppc cpu has in 32 bit mode but the number is diminishing.
Probably the best thing to speed up the intel architecture for now is to increase the number of general purpose registers. The 32 bit mode had to my knowledge only 8 or so that 64 bit mode has 16-20.
Well you also get a speedup in high volumn calculations due to the bigger register sizes, but that does not add as much as the higher register number. Why it took Amd to finally add a load of registers and why Intel always has been doctoring on the numer of pipes etc… but never increased their measly number of registers (heck even the late 68000 incarnations had more than Intels 32 bit mode) is beyound me since every compiler builder knows that a a high number of registers reduces stack operations significantly.
June 27th, 2008 at 9:17 pm
why did amd increase only to 16 gpr from 8, rather than say 32 or 64 or even 128 gpr?
June 30th, 2008 at 11:46 am
[...] Preview: TestMac.net [...]
July 1st, 2008 at 7:42 am
i am a professional programmer and have been working with 64 bit servers for several years. let me help out with the confusion:
64 bit applications will consume more memory because variables are twice the size of 32 bit applications (this is a gross simplification).
the benefit is more available memory because of the larger adress space, but this is irrelevant for all use cases except scientific calculations, server applications and a few media editing apps.
the downside is *slower* execution speed because larger variables also means more work for the cpu. the overall execution speed however *increases* with 64 bit applications because modern cpus are optimized for 64 bit and actually have to do some extra work to run 32 bit code.
for end users, the speed gain should be barely noticeable. i guess the percieved snappyness can be attributed to other optimizations or maybe event just a fresh installation.
July 10th, 2008 at 10:36 pm
awesome! Is it hacked for PC yet?
July 13th, 2008 at 9:36 pm
nice article.