Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I bought Snow Leopard just to have a disc that wasn't bound to my Macbook for $29. Those expectations were met. :rolleyes:

There's also a Universal Binary of Disk Inventory X. I don't think it has been updated in ages but it works just fine even if it's stuck in a perpetual beta.
 
I used WinXP for a long time, through SP1, SP2, SP3. I never noticed any major enhancements to the UI. Or noticeable speed improvements.

http://en.wikipedia.org/wiki/XP_Service_Pack_2#Service_packs


So what you're saying is that if there are new features and speed improvements it's a service pack. So what does it take for it not to be a service pack?

No, I'm saying that a service pack is a bundle of minor enhancements and features, as well as fixes.

Most reviews of 10.6 have been muted, such as Cupertino's home town daily's headline calling it a "small step".

Others:

John Murrell, SJMN -- "Snow Leopard’s improvements aren’t easily spotted."

Walt Mossberg, Wall Street Journal — “But I don’t consider Snow Leopard a must-have upgrade for average consumers. It’s more of a nice-to-have upgrade. If you’re happy with Leopard, there’s no reason to rush out and get Snow Leopard.”

Brian Lam, Gizmodo — “The changes here are modest, and the performance gains look promising but beyond the built in apps, just a promise. If you’re looking for more bells and whistles, you can hold off on this upgrade for at least awhile."

Michael DeAgonia, Computerworld — “Unless users know where to look, they won’t see much difference between Leopard and Snow Leopard."

Of course there's a continuum from "patch" to "service pack" to "major release", and no clear criteria exist to categorize them. 10.6 is in the "big service pack" to "minor release" region to many people.
 
Means machines with 64bit EFIs. You can check whether your EFI is 32 or 64 with the command ioreg -l -p IODeviceTree | grep firmware-abi

It's not as simple as that.

MacBookPro3,1... Core 2 Duo, EFI64, but no dice on the kernel. There is either another/more specific hardware requirement or a check for MBP4,1 or newer (for MBPs, of course).
 
Of course there's a continuum from "patch" to "service pack" to "major release", and no clear criteria exist to categorize them. 10.6 is in the "big service pack" to "minor release" region to many people.

Well, if people think it's more than 29/129 along this continuum then it's a good deal.
 
Code:
#include <stdio.h>
#include <stdlib.h>

int main()
{
unsigned int c,e;
unsigned long long d;
d=0;
for (c=30; c>0; c--) {
	e=1<<c;
	while(malloc(e)){
	d+=e;
	}
printf("Allocated %llu total, just finished %lu chunk size.\n",d,e);
}
printf("Total allocated is %llu\n",d);
return 0;
}

Results:

Now, there may be issues with malloc that is preventing this from achieving the 'true' maximum. But running this on G5 w/4GB of RAM, the maximum allocation works out to about 3.43 GiB.

If OSX didn't get in the way of the user process, I would expect it would probably be able to do a lot closer to 4GB.

On the plus side, the limit is obviously better than 2GB, and better even than Win32's 'special' 3GB mode. Unless anyone cares to make a better test case, I'd say the maximum memory of a 32-bit OSX app is 3.43GiB.

Strange: When I compiled with -m64 and ran on G5, I allocated about 40.21TiB, before I Ctrl-C'd the program. Swap usage was close to nil. What is up with that? I don't have a 40TB hard drive, so how is OSX doing these allocations? Am I doing something wrong, by merely thinking I can use that same code, compile -m64, and have it work? The programs internal counter of memory allocated and the Virtual Memory size in Activity Monitor are close enough that it doesn't appear there is a problem. What is this, Vapor-memory? (Noticed 32-bit version does the same.. No swap is 'used' by the untouched memory. Guess that is okay, and an program should be using API to determine amount of 'real' memory if that is what it really wants to know.)

Good job blowing your own theory. I've had Lightwave use more than 2GB of physical memory while rendering. In fact I have a scene that requires at least 3-4 GB of available memory just TO render. (Oh and LW on the Mac is still only 32bit...any day NewTek... any day is good.... :D )

-mark
 
Excellent review as always. As I said last night, John Siracusa's reviews are the ones I look forward to most with each new OS X release. Glad that MacRumors took my suggestion and posted it on the front page, too. ;) Though it would have been nice if it had the top spot for a little while longer... :rolleyes:

After reading that, I have difficulty seeing how people can still call this a service pack. Oh well!


For the OS history buffs, the end of the last page is great: links to Siracusa's equally thorough reviews of all OS X releases going back to the Developer Preview 2 in 1999. Great trip back down memory lane!

John Siracusa said:
 
Good job blowing your own theory. I've had Lightwave use more than 2GB of physical memory while rendering. In fact I have a scene that requires at least 3-4 GB of available memory just TO render. (Oh and LW on the Mac is still only 32bit...any day NewTek... any day is good.... :D )

-mark

I don't think I blew my 'theory.' I said I was quite confident a 32-bit OSX app can not use all 4GB of the address space, because the kernel is using some of it. My test tends to support my statement. The only question is where the boundary is drawn.

My other statement is that 32-bit apps can not use over 4GB in a way similar to the Windows AWE API. That appears to be confirmed, depending on how you split that hair: mmap will let you bank switch in an indirect way(relying on the file system cache), but you have to have actual files on the disk that are as big as the memory area. mmap will not let you do it without using files on disk. Since the mmap is going to try to keep your memory activity synced with the disc, the performance is probably not going to be what you want. Unless you just need fast read access.
 
Code:
#include <stdio.h>
#include <stdlib.h>

int main()
{
unsigned int c,e;
unsigned long long d;
d=0;
for (c=30; c>0; c--) {
	e=1<<c;
	while(malloc(e)){
	d+=e;
	}
printf("Allocated %llu total, just finished %lu chunk size.\n",d,e);
}
printf("Total allocated is %llu\n",d);
return 0;
}

Results:

Now, there may be issues with malloc that is preventing this from achieving the 'true' maximum. But running this on G5 w/4GB of RAM, the maximum allocation works out to about 3.43 GiB.

If OSX didn't get in the way of the user process, I would expect it would probably be able to do a lot closer to 4GB.

On the plus side, the limit is obviously better than 2GB, and better even than Win32's 'special' 3GB mode. Unless anyone cares to make a better test case, I'd say the maximum memory of a 32-bit OSX app is 3.43GiB.

Strange: When I compiled with -m64 and ran on G5, I allocated about 40.21TiB, before I Ctrl-C'd the program. Swap usage was close to nil. What is up with that? I don't have a 40TB hard drive, so how is OSX doing these allocations? Am I doing something wrong, by merely thinking I can use that same code, compile -m64, and have it work? The programs internal counter of memory allocated and the Virtual Memory size in Activity Monitor are close enough that it doesn't appear there is a problem. What is this, Vapor-memory? (Noticed 32-bit version does the same.. No swap is 'used' by the untouched memory. Guess that is okay, and an program should be using API to determine amount of 'real' memory if that is what it really wants to know.)

It is quite common (Linux does the same) that large malloc allocations only reserve address space, but don't actually reserve even swap space, until you start accessing the memory. I personally dislike that strategy, because even when malloc () returns a non-null pointer, it doesn't mean you can use it.
 
May I ask why? I don't mean to be argumentative, just curious. Are you one of those people who needs something new every 12 months? Personally, I don't need nor want something new just for the sake of something new. If it makes my life easier/better, I'm for it, but just because I'm bored & it's something different isn't my cup of tea. Just my 2¢

A faster turnaround time for OSs means a quicker response to developments in what is, after all, a very fast moving industry. It could be argued that if Microsoft had opted for a more incremental approach to OS development, rather than the, what was it? 5 year time lag between XP and Vista, they would have produced an OS that could have adapted better to all the technological changes that had taken place since its initial conception.
 
I can't remember any service pack which frees up 6GB of space and has massive changes under the hood.

A whole 6GB! Any machine with a hard drive small enough to benefit is too old to run SL.

Windows 7 has plenty of changes both visible and invisible, yet the fanbois here are saying its just a Service Pack. I think Aiden is just returning the compliment.
 
0 new functions in Snow Leopard is only half true.
When it comes to Quicktime X there are actually minus 50 new functions.
Is this a joke ?!
Will the Pro version be discontinued like so many other professional features of Apple products the last 12 months ?
This is the most disappointing update ever. Will Quicktime turn into a video viewer for house wifes just like the new screens ?

It's a good point: QuickTime Player X really lost all the good Pro functions: editing, outputting with custom compression etc. The new trimming features are kiddie-level at best.

BUT if you read the review, you see that you can have QuickTime 7 Player installed too, it's placed in your Utilities folder. That app remains Pro, and keeps all the good old QT7Pro features your used to. You can use either player.

I'm pretty sure Apple will keep QT7 Player alive until they make QTX Player able to handle the same Pro functions. Yeah, they'll probably charge you for it still
 
http://en.wikipedia.org/wiki/XP_Service_Pack_2#Service_packs




No, I'm saying that a service pack is a bundle of minor enhancements and features, as well as fixes.

Most reviews of 10.6 have been muted, such as Cupertino's home town daily's headline calling it a "small step".

Others:

John Murrell, SJMN -- "Snow Leopard’s improvements aren’t easily spotted."

Walt Mossberg, Wall Street Journal — “But I don’t consider Snow Leopard a must-have upgrade for average consumers. It’s more of a nice-to-have upgrade. If you’re happy with Leopard, there’s no reason to rush out and get Snow Leopard.”

Brian Lam, Gizmodo — “The changes here are modest, and the performance gains look promising but beyond the built in apps, just a promise. If you’re looking for more bells and whistles, you can hold off on this upgrade for at least awhile."

Michael DeAgonia, Computerworld — “Unless users know where to look, they won’t see much difference between Leopard and Snow Leopard."

Of course there's a continuum from "patch" to "service pack" to "major release", and no clear criteria exist to categorize them. 10.6 is in the "big service pack" to "minor release" region to many people.

Its a "small step" in terms of end user changes, but you and I both know that is OS involved a ******** more than just a service pack would. They rewrote a major amount of the OS for christs sake!
 
A whole 6GB! Any machine with a hard drive small enough to benefit is too old to run SL.

Windows 7 has plenty of changes both visible and invisible, yet the fanbois here are saying its just a Service Pack. I think Aiden is just returning the compliment.

The difference is windows 7 is built directly off Vista, making it a service pack.

SL is an actually rewritten OS.
 
The difference is windows 7 is built directly off Vista, making it a service pack.

SL is an actually rewritten OS.

I hate to tell you, but Windows 7 is based on Windows Vista, and Snow Leopard is based on Leopard.

That's pretty much the norm for new versions of software (OS X 10.0 aside). Did you expect them to rewrite it from the ground up each time they release a new version? That, my friends, is what causes things like Netscape 6 (and, for that matter, OS X 10.0). :)
 
The worst upgrade experience....

It took a full day to upgrade. The installer refused, saying my boot disk is unbootable. I spent the afternoon, evening, and overnight and eventually deleted the startup volume, recreated it, restored, all in order to be able to upgrade. The absolute worst Mac upgrade ever, and I mean since System7!:mad:
 
ArsTechnica does the best reviews hands down!

So true!

A Windows review tends to read like a summer holiday brochure... full of promises of sunshine and clear beaches :rolleyes:

That review is more like an engineering paper... explaining a bunch of new concepts and how it's going to change your world!


Now if only the reseller near me gets Snow Leopard stock! (Not in London atm... on holiday back home.)
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.