Is my early 2006 iMac 32-bit or 64-bit?

See http://www.apple.com/macosx/technology/ for details on how Apple's selling SL's 64-bit support, not the culmination. Your machine doesn't qualify.

Glad we could help. Good computing.

64-bit firmware, as in Unified EFI 2.x wasn't finalized by the EFI Group until Intel, Apple, Microsoft, and other partners, reached a standard (of sorts).

Official support for UEFI implementation, booting, hinged on operating system support as well, and that seemed to be tied to 2008 updates from Apple and Microsoft. Official support for Windows 64 is tied directly to EFI64. (The iMac is the odd-dog out in that game.)

You can't have the OS without a preliminary hardware, and they each have their own tick-tock evolution. So the chicken-egg or cart-horse.

You can't have any that are 32-bit if you are booting the 64-bit kernel mode. And that part may be in better shape in the next year, and have to wait for big apps to be redone. Some apps don't and probably never will need to be written as 64-bit, and will always remain 32-bit.

I don't see that there really is anything to get out of it that you aren't, if you needed and were using 64-bitness, it would be maybe if you were running CS5 or later with 8-16GB RAM, maybe, because it will still work on 32-bit kernel as a 64-bit application.

Because the newest EFI specs reference partition table changes, and seeing how 10.6 Disk Utility does a "preparing disk" before install (good reason to boot from DVD instead of in-place) to those EFI partitions (move? change size? or just update?) I do backup, then format, then installed (and import settings afterwards) for best possible experience.

*(Unified Extensible Firmware Interface)* A standard programming interface for setting up and booting a computer from the UEFI Forum (www.uefi.org). It evolved from the EFI interface developed by Intel, which was first used in Intel's Itanium line. UEFI also provides runtime services, which are used to access the drivers that were loaded in the pre-boot environment. PC Encyclopedia

http://www.uefi.org/
http://en.wikipedia.org/wiki/UEFI

UEFI Support for Current and Future Windows Operating Systems
http://www.microsoft.com/whdc/system/platform/firmware/uefireg.mspx

Are you asking if graphics card has to support OpenCL 1.x, or OpenGL 3.x? or the drivers? I would assume any driver has to now be 64-bit for 64-bit mode.

I've already seen how older graphic cards perform worse now with Snow Leopard, sometimes they don't perform properly.

UEFI is new. SATA came out over 6 yrs ago. And it was a mess for years. SATA III (and USB3) who knows what problems they will bring next year. SATA originally was just a marvel chip interface to talk to an ATA mechanism. Native SATA made it interesting, because G5 came out supporting only what existed, 1.0, and didn't support new SATA drives.

Technology is fluid, moving, and nothing is standing still or idle.
So no, the change, whatever you meant, has not already happened. We've outgrown existing SATA II though.

I am trying to clear up whatever if any confusion remains, particularly any I contributed. Below, I am talking only about what can be done with extant Apple software, not by Hackintosh builders, etc. [Did this link get posted above?|http://support.apple.com/kb/HT3696] It indicates clearly which Intel Macs do and do not have 64-Bit processors. Only these have 64-bit processors:

Intel Core 2 Duo and
Intel Quad-Core Xeon.

Only machines with 64-bit processors will run 64-bit programs as 64-bit programs (i.e., via the 64-bit kernel). Also, all Intel Macs with 64-bit processors will run 64-bit programs as 64-bit programs. Nevertheless, as clarified below, not all Macs with 64-bit processors can be booted into the 64-bit kernel mode.

Macs can have 32-bit and 64-bit EFIs. If one has a 64-bit EFI, then it also has a 64-bit processor, and it will run 64-bit programs as 64-bit programs (i.e., via the 64-bit kernel). All machines of this type, except the original Mac Pro (discontinued January 8, 2008), can boot into 64-bit kernel mode.

I am not sure if there are any Macs that have 32-bit EFIs and 64-bit processors (or if such a thing is even possible). I doubt it. But, regardless, no machine with a 32-bit EFI can boot into the 64-bit kernel mode.

Then, there is the issue of whether 64-bit Macs (i.e., those with both 64-bit EFIs and 64-bit processors) lose anything performance wise from not being able to boot into 64-boot mode. [My test results|http://discussions.apple.com/thread.jspa?messageID=10316685&#10316685] suggest that:

1. 64-bit programs run faster if a 64-bit Mac is booted into 64-bit kernel mode than if booted into 32-bit kernel mode (i.e., in either case, they are run as 64-bit programs) and

3. 32-bit programs run faster if a 64-bit Mac is booted into 64-bit kernel mode than if booted into 32-bit kernel mode (i.e., in either case, they are run as 32-bit programs)

I believe that the [Ars Technica Review read|http://arstechnica.com/apple/reviews/2009/08/mac-os-x-10-6.ars] may or may not identify the reasons for these phenomena (which is something I have been trying to explain for quite awhile). I am not going to get into all the whys the read may not explain the phenomena except to mention that the tests in the benchmark programs I used may not be sufficiently challenging to bring what ARS Technica mentions into play. Here is what the read says:

"We've already discussed why, at least initially, you probably won't want to boot into K64. But as Snow Leopard adoption ramps up and 64-bit updates of existing kernel extensions become available, why might you actually want to use the 64-bit kernel?

The first reason has to do with RAM, and not in the way you might think. Though Leopard uses a 32-bit kernel, Macs running Leopard can contain and use far more RAM than the 4 GB limit the "32-bit" qualifier might seem to imply. But as RAM sizes increase, there's another concern: address space depletion—not for applications, but for the kernel itself.

As a 32-bit process, the kernel itself is limited to a 32-bit (i.e., 4GB) address space. That may not seem like a problem; after all, should the kernel really need more than 4GB of memory to do its job? But remember that part of the kernel's job is to track and manage system memory. The kernel uses a 64-byte structure to track the status of each 4KB page of RAM used on the system.

That's 64 bytes, not kilobytes. It hardly seems like a lot. But now consider a Mac in the not-too-distant future containing 96GB of RAM. (If this sounds ridiculous to you, think of how ridiculous the 8GB of RAM in the Mac I'm typing on right now would have sounded to you five years ago.) Tracking 96GB of RAM requires 1.5GB of kernel address space. Using more than a third of the kernel's address space just to track memory is a pretty uncomfortable situation.

A 64-bit kernel, on the other hand, has a virtually unlimited kernel address space (16 exabytes). K64 is an inevitable necessity, given the rapidly increasing size of system memory. Though you may not need it today on the desktop, it's already common for servers to have double-digit gigabytes of RAM installed.

The other thing K64 has going for it is speed. The x86 instruction set architecture has had a bit of a tortured history. When designing the x86-64 64-bit extension of the x86 architecture, AMD took the opportunity to leave behind some of the ugliness of the past and include more modern features: more registers, new addressing modes, non-stack-based floating point capabilities, etc. K64 reaps these benefits. Apple makes the following claims about its performance:

* 250% faster system call entry point
* 70% faster user/kernel memory copy"

The question remains as to whether the benchmark programs I used would be able to pick up the memory and speed benefits provided by being booted into the 64-bit kernel mode. Note that unless the programs simply do not provide accurate results, there is not much else that could explain the effects I found.

Message was edited by: donv (The Ghost)

I am not sure if there are any Macs that have 32-bit EFIs and 64-bit processors (or if such a thing is even possible). I doubt it. But, regardless, no machine with a 32-bit EFI can boot into the 64-bit kernel mode.

If you don't know that, then I wonder where you are coming from. Try Xeon Clovertown Mac Pro and Woodcrest. And that isn't addressed by the Apple article linked to.

Knowing where Apple fits into the EFI universe, they aren't alone or only player, might give wider perspective and appreciation I hoped.

PS: I am not a Hacker, I use Windows 7, even on my Mac Pro anymore, as I said above.

The hatter wrote:
If you don't know that, then I wonder where you are coming from. Try Xeon Clovertown Mac Pro and Woodcrest. And that isn't addressed by the Apple article linked to.

What is you first sentence meant imply? It doesn't seem very friendly. I think I made it very clear up front where I am coming from. Here is the way I started my post:

"I am trying to clear up whatever if any confusion remains, particularly any I contributed. Below, I am talking only about what can be done with extant Apple software, not by Hackintosh builders, etc."

I believe that I stayed exactly with my purpose. Just because I am uncertain about one unimportant to my purpose fact is no reason to dis me.

Knowing where Apple fits into the EFI universe, they aren't alone or only player, might give wider perspective and appreciation I hoped.

Sure, I'd be OK with knowing where Apple fits into the EFI universe. But, knowing this is not pertinent to my purpose stated above. Why don't you elaborate more, in addition to what you have said above, on this topic? I'd like to hear, but researching it right now is not what I am up to.

Jonathan Mortimer wrote:
I ask because Mactracker says that it's 32-bit, so does that mean that Snow Leopard would be a waste of time? I thought the whole point was that it's 64-bit, but if it runs in 32-bit mode only then I can't see the performance increase being anything to talk about.

I now know for sure that your Mac does not have a 64-bit processor. Thus, I conclude that Sno truly would be as waste of time (and a small amount of $) for you (and that includes Sno plus future point updates). The upcoming point updates are going to do nothing noticeable for performance except for those who are finding Sno to be incredibly slow (hopefully).

I assume that you are running Leo. To me Leo was very nice and, while I know that Sno and Leo differ in many ways, it's somewhat hard to tell them apart in everyday use. Consistent with what I said above, you would not get any noticeably significant performance difference from Sno. I certainly haven't, and, as far as I can tell, my installation is virtually perfect.

I have referenced some of my test findings above. [Other of my results|http://discussions.apple.com/thread.jspa?threadID=2134001&tstart=0] suggest that Sno is a better performer than Leo (which is consistent with most published studies). [However, I have some other findings that suggest the opposite.|http://discussions.apple.com/thread.jspa?messageID=10302521&#10302521 ] Below is what I did, my results, and the questions my tests raise.

"I have been doing a variety of performance tests. To date, they have implied the superiority of SL over L performance wise. But, now, I have some results that imply the opposite--by a seemingly large margin. I am trying to understand what might be wrong (or right) with my results because they are surprising.

My current results were produced using xBench. My test was carefully controlled in the sense that SL and L both are installed on large partitions of the same 7200 RPM external firewire drive. Also, both have identical apps, settings, and files. Both programs were installed using erase and install. Naturally, I was running 32-bit mode on SL since L is not amenable to starting in 64-bit kernel mode.

I did ten xBench passes on each installation. Having scrutinized them carefully, I am going to report my results only for the best pass for each installation. I assure you that it would not matter an iota which passes I compared. The differences I show below would exist in essence, for example, even considering the best SL pass vs the worst L pass (or the averages of the passes for each installation).

xBench Results: (Snow Leopard Given first below)
Overall: 156 177
CPU: 200 190
Thread: 345 343
Memory: 193 182
Quartz Graphics: 227 241
OpenGL: 96 185
User Interface: 375 413
Disk: 70 72

These results imply a slight superiority for SL from CPU through Memory. However, L is superior on the final 4 tests, including particulary unexpected results on both the OpenCL test and the user interface test. The OpenCL test result might even be characterized as disturbing. Actually, the User Interface results also are disturbing. Note that my results do not imply a software problem with respect to Disk.

Does anyone have any insights about the results I call "disturbing." I know, of course, that a benchmark program is not an ideal mirror of real-world performance. Still, my results were quite unexpected even in the context of a benchmark program."

Message was edited by: donv (The Ghost)

This is the most understandable explanation I have seen yet. Posted by "Kappy" in response to a similar question:

+"What you have is a computer with an EFI32 boot ROM which simply means you cannot boot the 64-bit kernel. It does not mean you cannot run 64-bit software. The 64-bit kernel only provides the ability to address much larger amounts or RAM. However, our first generation MPs do not have a wide enough memory bus to support even the maximum amount of RAM usable in the 32-bit kernel. There will be some difference in speed when running the 64-bit kernel as opposed to the 32-bit kernel, however that difference isn't necessarily faster. In some cases it may be slower. Also because most third-party hardware drivers are not now 64-bit capable they will not work with the 64-bit kernel. For compatibility with some of the graphic drivers Apple has made the 32-bit kernel the default at startup. The graphic drivers issue relates to those Macs using integrated GPUs such as those found in Mac notebooks and some iMacs. Although your MP cannot boot the 64-bit kernel it can run any 64-bit software, so for you there's really no issue of concern..."+

NO.

Don, there is a ton, a TON, of misinformation in your post. Please, if you don't understand it, don't try to answer the questions.

Snow Leopard boots a 32-bit kernel on all machines. Period.

It can (and does) run applications as 64-bit on ALL machines but the "Core Duo" machines that were introduced as the first Intel Macs at MWSF 2006.

You can see if you are running 64 bit applications by launching Activity Monitor. 64 bit applications get the "benefits" of 64 bit and you are getting the advantages of 64-bit.

Note, again, that no versions of Mac OS X boot with a 64 bit kernel. Only Mac OS X Server on the latest two models of Xserve boot the 64 bit kernel by default.

Wrong. EFI determines whether the kernel can run as 32-bit. It has nothing to do with the applications, which run as 64-bit and is what determines that the OS is 64-bit.

Nevermind. I didn't recognize the thread from page 2. What happened? 🙂

Maybe you don't understand. And, please don't suggest that I shouldn't post even if I am wrong. Most of us do it all the time. 🙂 I think that the best way for you to proceed is to tell me where I am wrong. Let's discuss it. That's what's going on here--trying to get some clarity.