iMac Core 2 Duo Benchmarks [Archive]

MacRumors

Sep 9, 2006, 02:36 AM

http://www.macrumors.com/images/macrumorsthreadlogo.gif (http://www.macrumors.com)

MacCentral posted (http://www.macworld.com/2006/09/firstlooks/imacbench/index.php) a first look at the new Core 2 Duo iMac along with some early benchmarks.

The new iMacs which were released on Wednesday (http://www.macrumors.com/pages/2006/09/20060906091309.shtml) incorporate the latest Core 2 Duo processor from Intel. The Core 2 Duo represents the continuation of the Core Duo line which first made its appearance in Apple computers in January. Intel has maintained that the new chips would provide roughly 20% improvement in performance.

MacCentral tested the new 17" 2GHz iMac and 20" 2.16GHz iMac and compared them to the previous 20" 2GHz Core Duo and the 2.66GHz Mac Pro.

The most direct comparison between the two processors comes between the 17" 2GHz Core 2 Duo iMac and the 20" 2GHz Core Duo iMac. The overall score was 10% better in the new model while the individual tests showed gains up to 20%.

spicyapple

Sep 9, 2006, 02:38 AM

Is 20% speed improvement a lot for a core 2 designation?

arn

Sep 9, 2006, 02:42 AM

Is 20% speed improvement a lot for a core 2 designation?

Well, they said 20%, and it appears to be true in some tests.

It looks pretty good since it's not any more expensive than the previous iMacs.

arn

EricNau

Sep 9, 2006, 02:42 AM

Not a huge difference, but every little bit helps.

But those Mac Pros... Wow!

Chundles

Sep 9, 2006, 02:44 AM

Is 20% speed improvement a lot for a core 2 designation?

The designation "Core 2" comes from the fact that it's a whole new architecture in the chip. The original "Core" processors aren't based on the Core architecture, they were based on the Pentium M. The Core 2 processors (Merom, Conroe and Woodcrest) are based on a 65nm dual-core 64 bit architecture - the desktop and workstation chips are based around the mobile Merom architecture.

Core 2 promised about 20% more performance at the same power requirement as the Core chip at the same clockspeed.

randyharris

Sep 9, 2006, 02:49 AM

20% is a decent improvement in speed. However, I won't be replacing my 20" Intel iMac any time soon. There will be a time when I can't resist the upgrade, and a 24" is very appealing. But it's not a necessity - it's a desire.

The way things progress, I might find myself looking at a 3ghz iMac Core 4 Duo's in a couple years - and when that come out I'll be looking more closely.

Randy at http://www.MacSeven.com

DMann

Sep 9, 2006, 02:55 AM

VicMacs

Sep 9, 2006, 02:57 AM

now is the Xeon processor faster than the core 2 duo?

Sol

Sep 9, 2006, 02:57 AM

Did a certain platform jumping red-cap wearing Italian plumber write the headline?

Mama mia! iMac Core 2 Duo Benchamarks!

swingerofbirch

Sep 9, 2006, 02:59 AM

I'm getting a 17" 2 ghz C2D iMac (that's a mouthful) soon.

Sounds exciting with all that speed! I don't know what to do with it! I guess my iTunes visualizations will be smoother, and I'll have the Dashboard ripple effect.

I also work in iMovie sometimes so the speed gain will be nice.

I am moving up from an eMac 1 gigahertz G4. So I'm sure it will seem very fast to me. Probably more than I need.

Chundles

Sep 9, 2006, 03:02 AM

now is the Xeon processor faster than the core 2 duo?

The Xeon processor used in the Mac Pro is the name for the Core 2 Duo workstation processor. It was originally codenamed "Woodcrest" and is the high-end version of the three Core 2 Duo processors.

Merom = Core 2 Duo
Conroe = Core 2 Duo
Woodcrest = Xeon

spriter

Sep 9, 2006, 03:08 AM

I am moving up from an eMac 1 gigahertz G4. So I'm sure it will seem very fast to me. Probably more than I need.

You're in for a treat. I went from a 1.2GHz G4 to MacBook (2HGz Yonah) and it's streets ahead in terms of performance. 4 times faster encoding a DVD with Handbrake is a godsend.

The Merom iMac's are a great spec for the price.

kresh

Sep 9, 2006, 03:12 AM

Spaceman Spiff

Sep 9, 2006, 03:16 AM

Benchamarks, eh?

As long as it's not beating the Mac Pro, I'm happy.

Lollypop

Sep 9, 2006, 03:20 AM

freiheit

Sep 9, 2006, 03:22 AM

But because not all applications and tasks take full advantage of the Mac multiprocessing capabilities,

And I'm thinking... why?! 10 years ago BeOS had this down pat. The whole system was multi-threaded and multi-processor aware from the kernel all the way up through the user interface including the system services used by all native applications. It was amazingly responsive and was reported (in major publications) to gain as much as 60-70% performance by having a second CPU. I realize MacOS X is based on some old NeXTStep code which was not made for multiple processors, but come on! This is the 21st century and Apple's been selling dual processor machines for about 5 years now.

Anyway, this is great news. I'd been drooling over the new iMacs since they were announced and wondering how much I might gain by upgrading from my 2GHz G5 PowerMac. It's very enticing.

Erasmus

Sep 9, 2006, 03:27 AM

Ah, less money for 4" more screen, a ~20% faster C2D CPU, a much much faster GPU, etc? Thankyou mister Jobs! Can't wait to see what Tuesday's iPods will be like! Not that I intend to get one, or maybe I do? Who knows? And here's to some sort of Mac with Kentsfield and R600 (More likely G80 :( ) at Macworld 2007, or maybe even earlier with a bit of luck!

It's good to know that one can get a 24" screen, 2.33 Core 2 Duo goodness with the Geforce 7600 256Mb and 2 gig RAM from 5 grand with enough cash spare to buy a new Macbook. (Well at least in edu pricing)

BTW, this is a bit irrelevant, but can anyone explain to me how one gets Xgrid, and what one needs to make it work? ie. Does it cost money, or can it be downloaded from the net, and do u need a FiberChannel card to make it work? Would be cool to have a Macbook Xgridded to a 24" iMac. Preferrably Merom Macbook and Conroe iMac. But might take a while to get that :o

Never mind. : (
Need 10.4 Server apparently, and at least 3 computers to make it worthwhile, it looks like.

bloodycape

Sep 9, 2006, 03:34 AM

Cant wait to see what the mbp can do with that chip. If only we could upgrade our core duo chips to the new core 2 duo easily.

DeSnousa

Sep 9, 2006, 03:36 AM

Any word on the difference between the C2D 2.16 and the 2.33? Is it worth the upgrade price?

iMikeT

Sep 9, 2006, 03:53 AM

I wonder how the 24" iMac equiped with a 2.33ghz Core 2 Duo would fare in that benchmark.

mdntcallr

Sep 9, 2006, 05:03 AM

i would think that the imac 24" 2.33 is looking mighty nice about now.

ezekielrage_99

Sep 9, 2006, 05:27 AM

That's pretty quick but I thought there may have been a little more speed from the Core 2 Duos.

miniConvert

Sep 9, 2006, 05:49 AM

I think we all knew that Merom would only bring modest performance gains. I'm surprised they're as high as they are. I'm still not sure why they're using the laptop line of processors in their mid range desktop but it's certainly a testament to the power of Intel's new chips. If it keeps the iMac's cool and efficient then it's all good.

Stridder44

Sep 9, 2006, 05:57 AM

I am moving up from an eMac 1 gigahertz G4. So I'm sure it will seem very fast to me. Probably more than I need.

Im still stuck on my 800 mhz G4 with it's crappy 2X Graphics card. 2X!! Oh how I wait for the day when...

Stridder44

Sep 9, 2006, 05:59 AM

And I'm thinking... why?! 10 years ago BeOS had this down pat. The whole system was multi-threaded and multi-processor aware from the kernel all the way up through the user interface including the system services used by all native applications. It was amazingly responsive and was reported (in major publications) to gain as much as 60-70% performance by having a second CPU. I realize MacOS X is based on some old NeXTStep code which was not made for multiple processors, but come on! This is the 21st century and Apple's been selling dual processor machines for about 5 years now.

Anyway, this is great news. I'd been drooling over the new iMacs since they were announced and wondering how much I might gain by upgrading from my 2GHz G5 PowerMac. It's very enticing.

I agree. This is ridiculous. It's like buying an HD tv and not getting HD channels...

Willis

Sep 9, 2006, 06:00 AM

64-bit likes rosetta it seems. Good speed gain for CS2. 27/36 seconds quicker aint bad. Im guessing the larger L2 cache has some play in that too.

I love that 24". Its a BEAST!

MacsRgr8

Sep 9, 2006, 06:08 AM

Too bad the old Quad G5 wasn't put in there.

foo10

Sep 9, 2006, 06:13 AM

I was thinkin of buying a C2D MBP 17" + 23" ACD, but no i've been thinking of going for a 24" iMac and later for the cheapest Macbook. It will be still some time thought, this G5 iMac is still running good for me. Maybe when Leopard is released.

wwworry

Sep 9, 2006, 07:25 AM

I am curious about the iMovie benchmarks. One might think the mac pro would be over twice as fast but it's not. Is that because of software limitations in iMovie?

I am about to make a purchase of either an iMac or a 2.0Ghz. MacPro for monthly workouts in Final Cut. Hard to decide.

BornAgainMac

Sep 9, 2006, 07:36 AM

It wouldn't have been worth it to wait for these unless you wanted the 24 inch display (and Firewire 800) with your iMac. The iMac really looks like a good switcher machine that gives Pro performance to the masses.

EagerDragon

Sep 9, 2006, 07:53 AM

EagerDragon

Sep 9, 2006, 08:03 AM

Manic Mouse

Sep 9, 2006, 08:13 AM

Until Leopard is out we wont see the true value of these babies. Also by then some of the apps will take advantage of the muti-cores and multi-cpus, and the changes to the OS will allow applications not written for more than 1 core to take some advantage also. So like I said Leopard will be the one showing the true potential of these babies. Can't wait!!!!!!!:cool:

The fact that the new iMacs can't address more than 3Gb of memory and are therefore operating on a 32bit logic-board makes me doubtful as to whether or not these systems are really 64-bit capable... It seems like some kind of hybrid 32/64bit system.

Will the C2D iMacs be able to run 64bit code, despite not having the 64bit address space (and being able to access over 4Gb or RAM)?

Azrel

Sep 9, 2006, 08:24 AM

Manic Mouse

Sep 9, 2006, 08:25 AM

Yes of course it can, you obviously don't understand what x86_64 is.

No need to be rude. I was just asking a question!

steve_hill4

Sep 9, 2006, 08:55 AM

My first question is if the Mac Pro offers less of a performance increase than it first appeared, (for the time being at least), would Apple use this as an argument against a Conroe Tower?

I'm still behind Apple increasing their product lines, (Conroe Tower, 13" MBP, 15 & 17" MBs), and they have effectively done just this with taking teir consumer iMac offering from 2 standard models to 4. Can we expect more of the same in the future?

Roll on September 12th and beyond.

steve_hill4

Sep 9, 2006, 09:06 AM

The fact that the new iMacs can't address more than 3Gb of memory and are therefore operating on a 32bit logic-board makes me doubtful as to whether or not these systems are really 64-bit capable... It seems like some kind of hybrid 32/64bit system.

Will the C2D iMacs be able to run 64bit code, despite not having the 64bit address space (and being able to access over 4Gb or RAM)?
Well, wasn't the iMac G5 restricted to 2GB, yet it was a 64-bit processor? A 32-bit computer can take up to 4GB, but due to the hardware Apple was/is using, they can't even take this.

What i find odd is that it appears to allow 1 or 2GB in either slot, but no more than 3GB in total. That is obviously the maximum the board can take, but it would have made a little more sense to allow 2GB in each. This will not really effect it's ability to run 64-bit software, just restricts how much memory can be used. Remember that you have been able to get AMD systems with 64-bit processors for some time now. They won't take more than 4GB, but will allow you to run 64-bit OSes and Apps.

I'm hoping by the time I'm after an iMac, it will take at least 4GB, have Blu-Ray as an option, (although I may opt for standard Superdrive if it is an option and buy a Mac compatible external later), include bigger hard drives and stick to a similar price point to now.

I'm tempted by the 20" now, but am not buying yet and would want about 320-400GB in there for the same price, perhaps even 2GB RAM. I've got time to wait however.

Manic Mouse

Sep 9, 2006, 09:18 AM

andiwm2003

Sep 9, 2006, 09:31 AM

Yes of course it can, you obviously don't understand what x86_64 is.

Manic Mouse obviously understands what 64 bit means. that is obvious if one is able to read the post.;)

i also wonder if the new systems are really future proof or if hybrid systems like this will in a few years be not compatible. also how will windows run on a system like this? what about games that usually expect a certain hardware setup?

Bibulous

Sep 9, 2006, 09:43 AM

QCassidy352

Sep 9, 2006, 09:44 AM

I guess I've got mind whip lash from the transition to Intel. It's still kinda hard to wrap the mind around these speed improvments. I'm still used to the very modest speed bumps from the PPC days.

How wonderfully refreshing it is to see these leaps in speed with each product update. I hope this pace keeps up. Some may disagree, but I think it's spectacular compared to what we used to get from Moto/Freescale/IBM.

Nonsense. Maybe at the very end, but we used to see updates like 800mhz-->1Ghz. That's a 25% increase in performance (if processor power is the limiting factor). Even the G5 went dual 2.0 --> dual 2.5 in one jump, which is a bigger increase than this.

This is NOT the promised 20% increase; it's a 10% increase. Some areas may show more, but that only means that some show less.

I'm just not sure why everyone is so impressed with these imacs. The 24" was a great addition (and the low-end 17" for $1k is nice too), but the 17" and 20" are the same as the old models except for a processor speed bump and 1 GB RAM standard. All that for less money is not bad, but it's hardly thrilling considering that it's the first update in 8 months.

epitaphic

Sep 9, 2006, 09:56 AM

This is interesting. So the frequency difference of 2.16 to 2.66 is 23% yet the MacPro is only 20-30% faster? So clock for clock, the iMac would be between 7% slower and 3% FASTER than the MacPro!

There is no chance in hell there'll be a Conroe system for some time. This is also the reason the Mac Pros are all quads, a duo would be just as fast for all but the 0.001% of people that do stuff that fully utilizes the quad.

zz5555

Sep 9, 2006, 09:57 AM

Well, wasn't the iMac G5 restricted to 2GB, yet it was a 64-bit processor? A 32-bit computer can take up to 4GB, but due to the hardware Apple was/is using, they can't even take this.

What i find odd is that it appears to allow 1 or 2GB in either slot, but no more than 3GB in total. That is obviously the maximum the board can take, but it would have made a little more sense to allow 2GB in each. This will not really effect it's ability to run 64-bit software, just restricts how much memory can be used. Remember that you have been able to get AMD systems with 64-bit processors for some time now. They won't take more than 4GB, but will allow you to run 64-bit OSes and Apps.

I suspect that it will take 2GB in each and you will get more than 3GB total, then. But HW addresses are going to eat up some of the 4GB total address space, so you're not going to get all 4GB. (Apple could restrict things to just 3GB regardless, although I can't see any reason for them to do so.) I suspect the 3GB is just a marketing move. After all, they wouldn't want to advertise something like "for an extra $500 we'll stick in a 2nd 2GB stick to give you an extra .1 GB of memory!"

I'd hoped for a full 64 bit system, though that probably wasn't realistic in the time frame. I'm not buying until January, so we'll see what comes out at MacWorld (which I think I'll actually go to this year) and then I'll decide. All in all, I think I'd be very happy with the 24" model (with a 2nd display).

Steve

zz5555

Sep 9, 2006, 10:00 AM

JRM PowerPod

Sep 9, 2006, 10:01 AM

JRM PowerPod

Sep 9, 2006, 10:03 AM

Now that they are 64bit all i need to do is find two 8 exactabyte DDR2 dimms and i'll be sweet

Chundles

Sep 9, 2006, 10:05 AM

Well, wasn't the iMac G5 restricted to 2GB, yet it was a 64-bit processor?

The last revision of the iMac G5 (the one with the iSight) had the option of 2.5GB of RAM. It had 512MB built-in and you could option a 2GB stick for the 1 open slot it had.

That 2GB of course cost an arm and a leg...

Mac Fly (film)

Sep 9, 2006, 10:20 AM

What about the 3.0Ghz Mac Pro?:D

AidenShaw

Sep 9, 2006, 11:01 AM

Good - now we won't have to wade through any arguments with fanbois who claim that the iMac is the "most powerful desktop on the planet"....

:D

Manic Mouse

Sep 9, 2006, 11:01 AM

With the decent graphics and these C2D's they make the iMac a formiddable machine. Alot of PowerMac's are going to be replaced by these new iMac's i feel. Probably Apple's most impressive, solid and reliable machine at the moment

Unless Leopard is designed to make full use of the extra threads/cores available on the quad-core Mac Pro. Like that OS someone mentioned earlier in the thread that saw 60-70% performance gains when the cores were doubled.

I think the Mac Pro is fairly safe as a workstation, but fewer people will use it as a simple desktop now that iMacs are so competitive.

AidenShaw

Sep 9, 2006, 11:03 AM

Probably Apple's most impressive, solid and reliable machine at the moment
I don't think that there's any data yet on failure rates and problems with the new Core 2 iMacs...

BRLawyer

Sep 9, 2006, 11:10 AM

Good - now we won't have to wade through any arguments with fanbois who claim that the iMac is the "most powerful desktop on the planet"....

:D

As previously confirmed, the iMac is the most powerful AIO desktop...the title you just mentioned belongs to the MacPro...sorry for the misunderstanding...:rolleyes:

How is Winblows going on your side, Aiden? Many BSODs today?

PowerBooks G5, oops, Mini Tower Macs next Tuesday!!!! :rolleyes:

BRLawyer

Sep 9, 2006, 11:12 AM

I don't think that there's any data yet on failure rates and problems with the new Core 2 iMacs...

He is talking about reliability data from at least the last 10 years, Aiden...data you must have no clue about, of course...after all, Macs just started doing Windows, right? :rolleyes:

Homy

Sep 9, 2006, 11:13 AM

I'm just not sure why everyone is so impressed with these imacs.
Faster processor, double the RAM, cheaper AND 21-37% better game performance:

New 17" C2D is 37% faster in UT 2004 than old 17" CD.
New 17" C2D is 21% faster in UT 2004 than old 20" CD.
New 20" C2D is 37.5% faster in UT 2004 than old 20" CD.

and I don't even play UT 2004 ;).

MattyMac

Sep 9, 2006, 11:17 AM

I want to see some unpacking pics of that 24inch model compared with the 20in. Soon enough I suppose.

ezekielrage_99

Sep 9, 2006, 11:19 AM

Faster processor, double the RAM, cheaper AND 21-37% better game performance:

New 17" C2D is 37% faster in UT 2004 than old 17" CD.
New 17" C2D is 21% faster in UT 2004 than old 20" CD.
New 20" C2D is 37.5% faster in UT 2004 than old 20" CD.

and I don't even play UT 2004 ;).

I think more than anything the facts that they are cheaper and faster the previous models are more impressive than anything else and with tempt a lot of people to join the "dark side" of Mac.

AidenShaw

Sep 9, 2006, 11:19 AM

Unless Leopard is designed to make full use of the extra threads/cores available on the quad-core Mac Pro.
The real problem isn't the OS as much as it is in applications.

A well-threaded O/S won't help make Photoshop or Avid run much faster, unless the application code is also able to use all of the cores that are present.

Some applications are inherently serial - you have to do step A, then step B (because step B depends on step A). It's not a matter of poor programming, it's that the task is serial. (Note that many Photoshop benchmarks quote "MP-aware" filters separately from actions that don't scale.)

For these "not well-threaded" applications, multiple cores will still be beneficial so that you can run multiple applications simultaneously - all at full speed.

There are some server-type applications (web or database) that run many (hundreds or thousands) threads simultaneously. (For a web server - each browser session is a natural thread.) For these applications, operating system efficiency is important. The reports that OSX is poor at threading (such as Mac OS X limits server performance (http://www.macnn.com/articles/05/06/15/os.x.server.review/)) aren't really that important for desktop apps that want to use all 4 cores (or soon 8).

http://www.anandtech.com/mac/showdoc.aspx?i=2436
The server performance of the Apple platform is, however, catastrophic.
...
Workstation apps will hardly mind, but the performance of server applications depends greatly on the threading, signalling and locking engine.

Mikael

Sep 9, 2006, 11:21 AM

The fact that the new iMacs can't address more than 3Gb of memory and are therefore operating on a 32bit logic-board makes me doubtful as to whether or not these systems are really 64-bit capable... It seems like some kind of hybrid 32/64bit system.

Will the C2D iMacs be able to run 64bit code, despite not having the 64bit address space (and being able to access over 4Gb or RAM)?
I've searched for address bus width for the Core 2 line, but haven't found anything. It is, however, not likely that Intel downgraded the width from the previous models, which would mean either a 36 or a 40 bit bus. Also note that AMD's K8 (Athlon 64, Sempron and Opteron) also feature a 40 bit address bus and not 64 as someone might think. I also think that the G5 uses a 40 or 42 bit address bus, so it's pretty much the same there too.

So, if Core 2 has a 40 bit bus (which is likely) you end up with 1TB of addressable memory space.

Also, Core 2 CPUs are fully capable of running 64 bit code. Whether the address bus is 64 bits wide or not hasn't got anything to do with the width of the internal data path and execution unit width.

Manic Mouse

Sep 9, 2006, 11:28 AM

The real problem isn't the OS as much as it is in applications.

A well-threaded O/S won't help make Photoshop or Avid run much faster, unless the application code is also able to use all of the cores that are present.

Some applications are inherently serial - you have to do step A, then step B (because step B depends on step A). It's not a matter of poor programming, it's that the task is serial. (Note that many Photoshop benchmarks quote "MP-aware" filters separately from actions that don't scale.)

For these "not well-threaded" applications, multiple cores will still be beneficial so that you can run multiple applications simultaneously - all at full speed.

There are some server-type applications (web or database) that run many (hundreds or thousands) threads simultaneously. (For a web server - each browser session is a natural thread.) For these applications, operating system efficiency is important. The reports that OSX is poor at threading (such as Mac OS X limits server performance (http://www.macnn.com/articles/05/06/15/os.x.server.review/)) aren't really that important for desktop apps that want to use all 4 cores (or soon 8).

http://www.anandtech.com/mac/showdoc.aspx?i=2436
The server performance of the Apple platform is, however, catastrophic.
...
Workstation apps will hardly mind, but the performance of server applications depends greatly on the threading, signalling and locking engine.

Ahh. Nice info Aiden, thanks for that!

Many BSODs today?

I must be one of the few windows users who has never seen a BSOD since Win 95...

I've been using the Vista Beta 2 exclusively since it's release and it's apparently a very unstable OS yet I haven't seen a single BSOD. Looking forward to upgrading to RC1 and all the performance improvements that will bring though! :D

What I'm wondering is how Leopard will change the performance of the iMacs and Mac Pros. Will having a full 64-bit operating system and applications mean they run faster, or will the end-user see little difference?

Evangelion

Sep 9, 2006, 11:29 AM

The fact that the new iMacs can't address more than 3Gb of memory and are therefore operating on a 32bit logic-board makes me doubtful as to whether or not these systems are really 64-bit capable... It seems like some kind of hybrid 32/64bit system.

Will the C2D iMacs be able to run 64bit code, despite not having the 64bit address space (and being able to access over 4Gb or RAM)?

Uh.... The CPU is 64bits, and it DOES have 64bits address-space. And it runs 64bit code just fine. It just happens that the rest of the system it's hooked up to does not support 4+GB of RAM. But that has NOTHING to do with the "bitness" of the CPU. C2D is a 64bit CPU, period.

Hell,there has been 64bit CPU's and machines for long time (SGI and Sun comes to mind), yet back in those days even 1GB of RAM would have cost an arm and a leg. Yet those machines were 64bit machines.

As to logic-board being 32bits... Uh, no. There might be various reasons why it doesn't support 4GB of RAM, and it isn't due to "bitness" of the logic-board. And pray-tell: what exactly is a "32bit logic-board"?

Manic Mouse

Sep 9, 2006, 11:30 AM

I've searched for address bus width for the Core 2 line, but haven't found anything. It is, however, not likely that Intel downgraded the width from the previous models, which would mean either a 36 or a 40 bit bus. Also note that AMD's K8 (Athlon 64, Sempron and Opteron) also feature a 40 bit address bus and not 64 as someone might think. I also think that the G5 uses a 40 or 42 bit address bus, so it's pretty much the same there too.

So, if Core 2 has a 40 bit bus (which is likely) you end up with 1TB of addressable memory space.

Also, Core 2 CPUs are fully capable of running 64 bit code. Whether the address bus is 64 bits wide or not hasn't got anything to do with the width of the internal data path and execution unit width.

Thanks :D

To be honest I'm not all that well versed in the differences between 64bit and 32bit computing...

AidenShaw

Sep 9, 2006, 11:33 AM

He is talking about reliability data from at least the last 10 years, Aiden...
You mean Powerbook hinges, iBook logic boards, MacBook Random Shutdown Syndrome (RSS) (http://blogs.zdnet.com/Apple/?p=274), eMac logic boards (http://www.macintouch.com/readerreports/emac/topic4116.html), Powerbook memory slot, White spots and lines on Powerbook screens, chipping paint, cracked cubes, iMac G5 video and power problems,Apple repair extensions (http://www.apple.com/support/),...

I see your point!

Evangelion

Sep 9, 2006, 11:34 AM

As previously confirmed, the iMac is the most powerful AIO desktop...the title you just mentioned belongs to the MacPro...sorry for the misunderstanding...:rolleyes:

You again with your ludicrous claims? What makes you think that Mac Pro is the fastest desktop there is? What is the secret ingredient that makes it faster than other machines, considering that it uses same components than others do? And since Mac Pro supports less RAM (16GB vs. 64GB) than Dell does (for example), how could you say that it's "the most powerful"? Compared to Dell, it will be dog-slow for tasks that require a lot of RAM.

How is Winblows going on your side, Aiden? Many BSODs today?

Now, I hate Windows and I use it at work because I have no choice. An there are plenty of bad things in it. But I don't get any BSODs. I really don't. Back when I used a desktop, I sometimes left the machine running for weeks and I had no problems.

When was the last time you used Windows? Back in Windows ME-days?

Seriously: I think you should take a chill-pill an dial-back that fanboyishness of yours.

steve_hill4

Sep 9, 2006, 11:35 AM

The last revision of the iMac G5 (the one with the iSight) had the option of 2.5GB of RAM. It had 512MB built-in and you could option a 2GB stick for the 1 open slot it had.

That 2GB of course cost an arm and a leg...
That's right actually, it was 512MB soldered and a single slot wasn't it? I also thought for a short time after the new C2D iMacs were announced and it said max. 3GB, 1GB would be soldered and there would be a single 2GB slot or two 1GB slots.

Some good points have been raised on the 64-bit OS front. Since Leopard will fully support 64-bit down to the kernal, I would hope this first 64-bit Intel iMac would benefit as much as the Mac Pro.

However, does anyone know for sure whether a) the OS X on the new iMacs is 64-bit and/or b) whether OS X on Intel has even been translated to 64-bit, (thinking Mac Pro here too)? It's something I have yet to bother looking into, but any answers here would be appreciated.

Manic Mouse

Sep 9, 2006, 11:36 AM

Uh.... The CPU is 64bits, and it DOES have 64bits address-space. And it runs 64bit code just fine. It just happens that the rest of the system it's hooked up to does not support 4+GB of RAM. But that has NOTHING to do with the "bitness" of the CPU. C2D is a 64bit CPU, period.

Hell,there has been 64bit CPU's and machines for long time (SGI and Sun comes to mind), yet back in those days even 1GB of RAM would have cost an arm and a leg. Yet those machines were 64bit machines.

As to logic-board being 32bits... Uh, no. There might be various reasons why it doesn't support 4GB of RAM, and it isn't due to "bitness" of the logic-board. And pray-tell: what exactly is a "32bit logic-board"?

Sorry, that post was just me airing my thoughts and seeing what other people had to say. I wasn't making any claims, I don't know a lot about 64-bit computing. Glad my ignorant thoughts turned out to be wrong though!

I just assumed that being 64-bit or 32-bit was a system wide principle, either or.

steve_hill4

Sep 9, 2006, 11:38 AM

Now, I hate Windows and I use it at work because I have no choice. An there are plenty of bad things in it. But I don't get any BSODs. I really don't. Back when I used a desktop, I sometimes left the machine running for weeks and I had no problems.

When was the last time you used Windows? Back in Windows ME-days?
I have had a few BSODs in XP, but they are rare. Mainly I used to leave my tower on 24/7 and the worst I would get is switching my monitor on first thing in the morning or when I got home from work and see it had rebooted itself and was telling me it was an invalid system disc.

Since Mac, never once has this happened, (even though I tend to power down on this more often, I still often leave running 24/7 if it is doing something that requires up time).

Strangely enough, I am one fo the few that never had that many problems with ME.

AidenShaw

Sep 9, 2006, 11:39 AM

As to logic-board being 32bits... Uh, no. There might be various reasons why it doesn't support 4GB of RAM, and it isn't due to "bitness" of the logic-board. And pray-tell: what exactly is a "32bit logic-board"?
The Napa chipset used with Yonah only supported 32 address lines.

A new Napa64 chipset is here that supports the additional address lines to allow > 4 GiB of physical memory.

http://www.cdrinfo.com/Forum/tm.asp?m=126194&mpage=1&key=&#126194

wizard

Sep 9, 2006, 11:48 AM

I think far to many persons here are dismissing Merom and are not thinking about this rationally. They are getting a 20% boost at the SAME clock speed as the old chips. That is similar to 400 MHz additional performance. Granted one can not scale the MHz without other issues coming into play but still it gives you an idea about how well the chips are doing. On top of that the 64 bit option isn't likely to even be in the picture at the moment nor the other chip improvements.

What I'm getting at is that Merom is demonstrating to be a significant step forward. The real surprise is that Intel actually delivered a nice chip set.

As to the laptop chip in the iMac come on everyone here knows about the thermal issues in that machine. It is no surprise at all. What it does demonstrate clearly is that multiprocessing can deliver very usable performance at very low powers. A number of people where dismissing multiprocessing as the wave of the future a couple of years ago (I remember because I was promoting it) this should cool off those concerns for the majority of the users out there. I'm with you on power usage more manufactures need to look at this issue seriously.

The thing with Merom is that I'm seeing big upside potential performance wise. I'm surprised that more people are not up on this. Intel can accomplish this by tweaking the current process or going to the next generation feature size. Merom is nice but it still needs more effort at power reduction to scale speed.

Thanks
Dave

I think we all knew that Merom would only bring modest performance gains. I'm surprised they're as high as they are. I'm still not sure why they're using the laptop line of processors in their mid range desktop but it's certainly a testament to the power of Intel's new chips. If it keeps the iMac's cool and efficient then it's all good.

Mikael

Sep 9, 2006, 11:50 AM

Thanks :D
No problem. :)

the title you just mentioned belongs to the MacPro...sorry for the misunderstanding...:rolleyes:
What a joke.

How is Winblows going on your side, Aiden? Many BSODs today?
Awww... Not this again. Windows doesn't blue screen without reason. If it did BSOD left and right, I would think that any of the three desktops at work should have gotten one in the 30,000 hours they've clocked now... I don't think I've ever seen a BSOD not being caused by bad or incorrectly configured hardware, overclocking or possibly bad drivers. As a matter of fact, the only BSODs I've had since 2002 (when I switched to XP) have been related to overclocked hardware. People need to check their damn hardware before going all ballistic over Microsoft's supposedly unstable operating system.

I always test the CPU/memory/mobo using Prime95 and Memtest86 right after I build a computer. I've caught some bad memory this way. Moreover, it has enabled me to stay BSOD-less on every machine I've built so far (that's ~15 machines in the past 5 years).

Mainly I used to leave my tower on 24/7 and the worst I would get is switching my monitor on first thing in the morning or when I got home from work and see it had rebooted itself and was telling me it was an invalid system disc.
That's almost 100% a hardware malfunction that causes Windows to restart after a serious failure. The default setting in Windows is to restart when it encounters a serious system failure and this can be disabled in the control panel to aid in seeking out the failing hardware. But I guess this is too late to fix now, since it sounds as if you sold the machine.

levitynyc

Sep 9, 2006, 11:52 AM

I have had a few BSODs in XP, but they are rare. Mainly I used to leave my tower on 24/7 and the worst I would get is switching my monitor on first thing in the morning or when I got home from work and see it had rebooted itself and was telling me it was an invalid system disc.

Since Mac, never once has this happened, (even though I tend to power down on this more often, I still often leave running 24/7 if it is doing something that requires up time).

Strangely enough, I am one fo the few that never had that many problems with ME.

Never had the BSOD on XP, but on ME ....WOW that was one shutty OS!

DrFrankTM

Sep 9, 2006, 11:53 AM

He is talking about reliability data from at least the last 10 years, Aiden...data you must have no clue about, of course...after all, Macs just started doing Windows, right? :rolleyes:

Am I the only one who's starting to find this petty crusade of yours annoying? Seriously, just drop it. The iMac is a great machine. I loved my G5 and I might well buy another iMac when I upgrade from the Mini I have now. A Mac Pro would be nice too, but we'll see. Anyways, the iMac is not the "most powerful desktop" in the world and no one gives a damn: it's an awesome machine regardless. So please, honestly, just drop it. What you said was false and still is. It's ok. Everyone makes mistakes. But drop it. It's starting to annoy me very seriously and I'd be surprised if I was the only one.

And I don't see what your personal attacks on some fine folks here are adding to the discussion. Along with opinions, AidenShaw - among others - brings a lot of technical expertise and hard facts to the table. I, for one, value that greatly, and I'm sure I'm not the only one. Indeed, the reason MacRumors is so awesome is that a fair share of the posters bring more than opinions to the forums: they also bring hard facts and educated guesses. I think that people can discuss respectfully, that people can argue respectfully, and that people can disagree respectfully. Most folks here do, and I'm sure you can do it too if you try even just a little. Thanks.

AidenShaw

Sep 9, 2006, 11:56 AM

I just assumed that being 64-bit or 32-bit was a system wide principle, either or.
A 64-bit operating system is one that provides 64-bit virtual addresses to its processes. It requires a CPU that supports 64-bit virtual addressing. The C2D is such a CPU, and runs 64-bit code and O/S.

These humongous 64-bit virtual addresses need to be translated to a physical address to reach the actual memory. The 64-bit CPU has a list of pages of physical memory, and tables to map a program's virtual address to a physical page. Once that mapping is done, the 64-bit virtual address can be used as a "synonym" for the actual physical address. The mapping is per process - two processes can use the same virtual address without interference - the same virtual address refers to different physical pages depending on the process which is using it.

In the case of the Napa(32) chipset, the chipset only has 32 address lines, and cannot physically handle more than 4 GiB of RAM. Some of those addresses are reserved for I/O purposes (such as the 256 MiB that is mapped to the VRAM of the video card).

If you plug 4 GiB of RAM into a Napa(32) system, you'll "lose" the memory that is over-mapped by I/O space. For example, right now I'm typing from a dual-Xeon (32-bit Netburst) with 4 GiB of RAM installed. Windows reports that I have 3520 MiB of memory. I've "lost" a half GiB due to these I/O space mappings. (My 4 GiB Yonah laptop reports 3.1 GiB available - PCIe systems seem to reserve a lot more memory for I/O than PCI-X systems.)

Apple is apparently saying that 3 GiB is the limit, so that they don't have to explain PCIe I/O bus mapping to people calling to complain that OSX isn't using all 4 GiB.
____________

This virtual-to-physical mapping has some other implications:

Windows & Linux 32-bit Operating Systems can use up to 64 GiB of RAM, when running on a logic board that supports the full 36-bit addressing of the P4
No 64-bit system has implemented 64 physical address lines. Some have implemented fewer than 32, but 36 (64 GiB) to 40 (1024 GiB) is typical (price a terabyte of RAM, and figure out how many slots you need ;) )
MacOS 9 and before didn't have per-process mapping tables - so if different processes used the same address, it would refer to the same physical memory. This meant a mistake by one application could make another application fail - or even the whole system.

Evangelion

Sep 9, 2006, 12:00 PM

The Napa chipset used with Yonah only supported 32 address lines.

Napa is the hardware-platform, composed of Yonah, Intel Mobile 945-chipset and Intel Pro Wireless. AKA third generation Centrino. And since the CPU used in that platform is 32bits, the platfom can be called a 32bit platfom. Note: this has nothing to do with the bitness of the logic-board. Napa64 (or rather: Santa Rosa) is Fourth generation Centrino that uses Merom and new chipset.

wizard

Sep 9, 2006, 12:04 PM

I see no reason why the new C2D iMacs can't run 64 bit code. Heck I've had a AMD 64 running Linux fro a couple of years now and it runs 64 bit code fine on 1 Gig or RAM. The 64 bit instructions are not part of the addressing scheme for the most part.

Now how well an application will run addressing more that 2 Gig of ram on these machines is another question. In part it will depend on how well virtual memory works. Performance wise it is always best to have all data in RAM, there is little doubt there, so you would not want to run large databases on a 64 bit machine with limited memory on a daily basis. It is simply a poor way to leverage the hardware. However not all 64 bit applications are data bound so one can still make serious use of the 64 bit capabilities.

The system is no more a hybrid than older machines of days pass that had 32 bit processors and could only address a small fraction of the available address space. All of the above being said though Apple is the one that writes the OS and they ultimately determine the capabilities on any one platform. I can't see them not enabling 64 bit when it is ready.

Dave

The fact that the new iMacs can't address more than 3Gb of memory and are therefore operating on a 32bit logic-board makes me doubtful as to whether or not these systems are really 64-bit capable... It seems like some kind of hybrid 32/64bit system.

Will the C2D iMacs be able to run 64bit code, despite not having the 64bit address space (and being able to access over 4Gb or RAM)?

Evangelion

Sep 9, 2006, 12:05 PM

In the case of the Napa(32) chipset

There is no "Napa chipset". Like I said, Napa is a hardware-platfom, composed to CPU (Yonah), chipset (Intel Express 945) and WLAN ()Intel PRO/Wireless). The amount of RAM might be limited due to timing-issues and the like.

flopticalcube

Sep 9, 2006, 12:06 PM

Just like the "good" old days all over again.

68000 32-bits inside, 24-bits (16MB) address outside.
8088 16-bits inside, 20-bits (1MB) address outside.

Software (excepting some parts of the OS) doesn't care. Merom is somewhat of a milestone in its 64-bit internals. Down the road there will be an OS release that will not support a 32-bit CPU like Yonah. Of course, 18 months later the next release probably won't support Merom (or Woodcrest) either. :rolleyes:

QCassidy352

Sep 9, 2006, 12:10 PM

With the decent graphics and these C2D's they make the iMac a formiddable machine. Alot of PowerMac's are going to be replaced by these new iMac's i feel. Probably Apple's most impressive, solid and reliable machine at the moment

No one who needs powerful graphics could go for anything except maybe the 24". The x1600 is pretty low-end for a mid-range desktop by now, and the nvidea 7600 is not bad but certainly not a powerhouse. And why would you say that the imacs are more "impressive, solid, and reliable" than the mac pros? Better values, maybe, but more impressive, solid, and reliable? :confused:

AidenShaw

Sep 9, 2006, 12:14 PM

Napa is the hardware-platform, composed of Yonah, Intel Mobile 945-chipset and Intel Pro Wireless. AKA third generation Centrino. And since the CPU used in that platform is 32bits, the platfom can be called a 32bit platfom. Note: this has nothing to do with the bitness of the logic-board. Napa64 (or rather: Santa Rosa) is Fourth generation Centrino that uses Merom and new chipset.

http://www.cdrinfo.com/Forum/tm.asp?m=126194&mpage=1&key=&#126194

"Santa Rosa is currently developed by Intel as successor of the current "Napa" platform. Napa will receive a 64-bit refresh in September of this year ("Napa64") to support the launch of the Merom processor. Napa64 will be replaced by Santa Rosa in the first or second quarter of 2007.

Compared to Napa and its Calistoga chipset, Santa Rosa and the Crestline chipset will not be just an evolutionary update, but bring several new features to the user."

There is no "Napa chipset". Like I said, Napa is a hardware-platfom...
My bad, thanks for the correction. Obviously I've been talking about the 945 (Calistoga) chip.

However, isn't Napa a "set of chips"? There's no "Napa chip", but there is a Napa chipset. Since Apple isn't using the Centrino brand, it's probably not a Napa platform per se.

steve_hill4

Sep 9, 2006, 12:15 PM

That's almost 100% a hardware malfunction that causes Windows to restart after a serious failure. The default setting in Windows is to restart when it encounters a serious system failure and this can be disabled in the control panel to aid in seeking out the failing hardware. But I guess this is too late to fix now, since it sounds as if you sold the machine.
Stevie doesn't sell machines, especially one's he can still use daily.

When I can be bothered, I will scrub both drives on it clean and reinstall XP and Fedora Core 5 on there. It still won't get used as much as my Mac though.

vd0t

Sep 9, 2006, 12:19 PM

Is one of the 512MB RAM on the iMac soldered on?? :confused:

levitynyc

Sep 9, 2006, 12:20 PM

No one who needs powerful graphics could go for anything except maybe the 24". The x1600 is pretty low-end for a mid-range desktop by now, and the nvidea 7600 is not bad but certainly not a powerhouse. And why would you say that the imacs are more "impressive, solid, and reliable" than the mac pros? Better values, maybe, but more impressive, solid, and reliable? :confused:

AMEN

Do you realize the Dell XPS 700 is capable of running 2 Nvidea GeForce - 1GB 7950 GX2 Quad SLI card for a total of 2 GB of Video

The fact that the new 24" Imac is only capable of handling a 256MB Video card is an embarassment as far as I am concered.

MacinDoc

Sep 9, 2006, 12:21 PM

You mean Powerbook hinges, iBook logic boards, MacBook Random Shutdown Syndrome (RSS) (http://blogs.zdnet.com/Apple/?p=274), eMac logic boards (http://www.macintouch.com/readerreports/emac/topic4116.html), Powerbook memory slot, White spots and lines on Powerbook screens, chipping paint, cracked cubes, iMac G5 video and power problems,Apple repair extensions (http://www.apple.com/support/),...

I see your point!
Umm, sorry, but according to Consumer Reports' (http://www.consumerreports.org/cro/index.htm)survey of 49,000 laptop users, Apple was tied for the third fewest number of laptop repairs since 2001, at 17% (compared to Sony and IBM at 16%, a negligible difference). And, according to their survey of 85,000 desktop users, all other desktop manufacturers had at least 36% more repairs than Apple, and Gateway had a whopping 91% more repairs than Apple.

Just because Apple offers extended service programs for computers beyond their warranty period does not mean that its computers are low quality, it only means that Apple is making an exceptional effort to maintain the loyalty of its customer base by doing more than the minimum required. Have you ever tried to get free service on a computer beyond its warranty period from another manufacturer?

speakerwizard

Sep 9, 2006, 12:22 PM

Dont know if anyone will bother reading this far in but these new imac core2duos still have the same 667mhz bus speed as before, we wont see new bus speeds till nearly a year from now i hear, that sucks, this system could probably warp ahead with a better bus speed, i wanna upgrade from my g5 dual 2.7 (1.35ghz bus speed if i remember right) but im not sure how much diffference it would really make to me, and going mac to an imac from a powermac is a bit odd for me (im a 3d animator) there are pros and cons but im still tempted, main think holding me back is an intel maya, although i could always bootcamp windows and use that version!

Evangelion

Sep 9, 2006, 12:23 PM

http://www.cdrinfo.com/Forum/tm.asp?m=126194&mpage=1&key=&#126194

[INDENT]"Santa Rosa is currently developed by Intel as successor of the current "Napa" platform. Napa will receive a 64-bit refresh in September of this year ("Napa64") to support the launch of the Merom processor. Napa64 will be replaced by Santa Rosa in the first or second quarter of 2007.

Wanna bet that Napa64 is 100% identical to ordinary Napa, apart from the fact that the CPU is Merom, instead of Yonah? Since Napa is a platform, just chaning the CPU to something else would mean that the platform has been refreshed.

MacinDoc

Sep 9, 2006, 12:23 PM

Is one of the 512MB RAM on the iMac soldered on?? :confused:
It was on a previous version of the iMac, not the current one.

MacinDoc

Sep 9, 2006, 12:28 PM

AMEN

Do you realize the Dell XPS 700 is capable of running 2 Nvidea GeForce - 1GB 7950 GX2 Quad SLI card for a total of 2 GB of Video

The fact that the new 24" Imac is only capable of handling a 256MB Video card is an embarassment as far as I am concered.
Sorry, but that's a ridiculous comparison. The only Mac you can reasonably compare the XPS 700 to is the Mac Pro, which has a lot more computing power for that kind of money.

BRLawyer

Sep 9, 2006, 12:32 PM

You again with your ludicrous claims? What makes you think that Mac Pro is the fastest desktop there is? What is the secret ingredient that makes it faster than other machines, considering that it uses same components than others do? And since Mac Pro supports less RAM (16GB vs. 64GB) than Dell does (for example), how could you say that it's "the most powerful"? Compared to Dell, it will be dog-slow for tasks that require a lot of RAM.

Now, I hate Windows and I use it at work because I have no choice. An there are plenty of bad things in it. But I don't get any BSODs. I really don't. Back when I used a desktop, I sometimes left the machine running for weeks and I had no problems.

When was the last time you used Windows? Back in Windows ME-days?

Seriously: I think you should take a chill-pill an dial-back that fanboyishness of yours.

Evangelion, I was picking on Aiden (as I always do) and we are normally able to exchange silly arguments with no fighting...

Could you just calm down and shut up for now? Really, go spend your posts on other people...no need to worry about my "fanboy" behavior...after all, this is a forum devoted to Mac fans and rumors...so please leave if you don't like it.

AidenShaw

Sep 9, 2006, 12:32 PM

http://www.tgdaily.com/2006/09/09/preview_kentsfield_processor/

Tom's Hardware benchmarks Intel's first quad-core "Kentsfield"

Culver City (CA) - Intel's first quad-core processor "Kentsfield" has found its way into the Tom's Hardware test lab. Several weeks before Intel will provide evaluation processors to the press, Tom's Hardware was able to obtain a qualification sample: The quad-core was sent through the entire test parcours and showed impressive performance.
...
Kentsfield, which industry sources refer to as "Core 2 Quadro," arrived as a 2.67 GHz version with a 266 MHz/1066 MHz FSB. The test engineers were able to adjust the FSB to 1333 MHz - which is still supported by the 975X chipset - and overclock the CPU by about 25%. The benchmarks were conducted with clock speeds ranging from 2.0 GHz to 3.33 GHz.

Kentsfield easily shattered previous benchmarks records and highlighted its horsepower especially in threaded applications such as audio and video processing.

Eidorian

Sep 9, 2006, 12:33 PM

Wanna bet that Napa64 is 100% identical to ordinary Napa, apart from the fact that the CPU is Merom, instead of Yonah? Since Napa is a platform, just chaning the CPU to something else would mean that the platform has been refreshed.I'm a little skeptical about Napa64 as well. I did read the previous links and articles on it. Why make Merom backwards compatible with Yonah?

http://www.tgdaily.com/2006/09/09/preview_kentsfield_processor/

Tom's Hardware benchmarks Intel's first quad-core "Kentsfield"

Culver City (CA) - Intel's first quad-core processor "Kentsfield" has found its way into the Tom's Hardware test lab. Several weeks before Intel will provide evaluation processors to the press, Tom's Hardware was able to obtain a qualification sample: The quad-core was sent through the entire test parcours and showed impressive performance.Nice information there. I wanted a little heads up on Kentsfield. Still, isn't it dual Conroe's with separate cache and then over the front side bus?

BRLawyer

Sep 9, 2006, 12:35 PM

Eidorian

Sep 9, 2006, 12:37 PM

No, I mean the lowest repair/return rate of the whole industry, and the highest praise in terms of support and reliability of the whole market...care to take a look around, perhaps?

We all know how public problems with Apple may get...1 PowerBook down is worth 1000 DOA Dells that never make it to the news...or perhaps you think the existence of iFixIt means all Macs are cracked...right? :rolleyes:Fixed!

I have to repair an iBook G3 this week. :rolleyes:

wizard

Sep 9, 2006, 12:38 PM

Sounds like a set of chips to me ;)

dave

There is no "Napa chipset". Like I said, Napa is a hardware-platfom, composed to CPU (Yonah), chipset (Intel Express 945) and WLAN ()Intel PRO/Wireless). The amount of RAM might be limited due to timing-issues and the like.

Evangelion

Sep 9, 2006, 12:38 PM

Could you just calm down and shut up for now? Really, go spend your posts on other people...no need to worry about my "fanboy" behavior...after all, this is a forum devoted to Mac fans and rumors...so please leave if you don't like it.

Well, since I have to suffer from your rabid fanboyishness as well (and it seems that I'm not the only one), then surely I can say something? I mean, it's not like your posts are invisible or something. Besides, if you post public messages on a public web-forum, do you REALLY have any grounds to complain if someone replies to your post?

Eidorian

Sep 9, 2006, 12:41 PM

Sounds like a set of chips to me ;)

daveNapa isn't a chipset. It's a grouping on Intel components (processor, northbridge/southbridge, and wireless) that make up the Napa platform.

Apple only uses the processors and northbridge/southbridge from Intel. Chipset normally ONLY refers to the north/southbridge.

Evangelion

Sep 9, 2006, 12:41 PM

Sounds like a set of chips to me ;)

dave

By "chipset" people usually refer to the southbridge/northbridge-combo. In this case that is the Intel Express 945.

MacinDoc

Sep 9, 2006, 12:42 PM

levitynyc

Sep 9, 2006, 12:43 PM

Sorry, but that's a ridiculous comparison. The only Mac you can reasonably compare the XPS 700 to is the Mac Pro, which has a lot more computing power for that kind of money.

My point is that with the new processors and RAM upgrades, the iMac is headed towards more powerful use that could potentially switch over a PC gamer. If you wanted to do some serious gaming on the 24" iMac you could....if not for the poor video card options.

Throw a dog a bone here and at least give us a 512MB option.

BRLawyer

Sep 9, 2006, 12:44 PM

Am I the only one who's starting to find this petty crusade of yours annoying? Seriously, just drop it. The iMac is a great machine. I loved my G5 and I might well buy another iMac when I upgrade from the Mini I have now. A Mac Pro would be nice too, but we'll see. Anyways, the iMac is not the "most powerful desktop" in the world and no one gives a damn: it's an awesome machine regardless. So please, honestly, just drop it. What you said was false and still is. It's ok. Everyone makes mistakes. But drop it. It's starting to annoy me very seriously and I'd be surprised if I was the only one.

And I don't see what your personal attacks on some fine folks here are adding to the discussion. Along with opinions, AidenShaw - among others - brings a lot of technical expertise and hard facts to the table. I, for one, value that greatly, and I'm sure I'm not the only one. Indeed, the reason MacRumors is so awesome is that a fair share of the posters bring more than opinions to the forums: they also bring hard facts and educated guesses. I think that people can discuss respectfully, that people can argue respectfully, and that people can disagree respectfully. Most folks here do, and I'm sure you can do it too if you try even just a little. Thanks.

Sorry, if you are feeling too stressed like a few others, you should know how to draw a difference between "personal attacks" as such, and generic/stupid/playful opinions on Windows and machines.

I love to joke and to make public my fondness for Macs. I participate in this forum for a long time, and have always had my arguments with Mr. Shaw, both in technical and playful terms...his posts are almost always directed at bashing Macs or praising its failures...and I don't care, honestly...I just follow on with the jokes.

I have been receiving personal attacks, instead, because people here can't tell such a difference...they get offended when MACHINES are bashed and then offend a PERSON back...

I don't do that, I bash machines and keep on playing on ad hominem arguments that don't have to be offensive...you may tell it by probably all my posts...so considering you seem to be a slightly more mature person, try to understand it better.

And please, getting "seriously annoyed" because I said the MP was the most powerful desktop? Really, get a grip and live a life outside...if you and others are that sensitive you need a reality check...this is just a damn computer forum, and we talk about damn computers...nothing else.

I have already defined "most powerful" as something more than just the speed of a chip...and I will keep saying it as much as I want, as Apple does, even if you show me a thousand SPECmarks...just leave me with my stupid jokes and don't answer, it's easier.

AidenShaw

Sep 9, 2006, 12:46 PM

Wanna bet that Napa64 is 100% identical to ordinary Napa, apart from the fact that the CPU is Merom, instead of Yonah? Since Napa is a platform, just chaning the CPU to something else would mean that the platform has been refreshed.
You may be right.

I couldn't find anything in the Intel technical documentation on the 945 to show a new revision or stepping of the chip - in fact most of the 945 docs at the Intel website don't mention the Core 2 at all....

BRLawyer

Sep 9, 2006, 12:48 PM

Well, since I have to suffer from your rabid fanboyishness as well (and it seems that I'm not the only one), then surely I can say something? I mean, it's not like your posts are invisible or something. Besides, if you post public messages on a public web-forum, do you REALLY have any grounds to complain if someone replies to your post?

Yes, I do, when such replies tell of personal offenses and nothing else...just read the rules of the forum. I was writing to Mr. Shaw, bashing PCs and praising Macs...nothing else.

The only remarks that may be close to "offenses" occur exactly when I have to reply to your stressed words that show a really angry and stupid degree of overreaction...relax, this is just a computer forum, and a MAC one for that matter.

Ah yeah, from the Apple Suisse website:

Le Mac le plus rapide jamais conçu.
Depuis l'annonce de la transition vers les nouveaux processeurs Intel, vous rêvez certainement d'un ordinateur de bureau professionnel offrant des performances hors pair.

MacinDoc

Sep 9, 2006, 12:48 PM

My point is that with the new processors and RAM upgrades, the iMac is headed towards more powerful use that could potentially switch over a PC gamer. If you wanted to do some serious gaming on the 24" iMac you could....if not for the poor video card options.

Throw a dog a bone here and at least give us a 512MB option.
I'm not a gaming expert, but from what I've read, 512MB is no faster than 256MB for most current gaming applications, it's the throughput of the GPU that counts. Do any gamers out there want to comment on this?

levitynyc

Sep 9, 2006, 12:53 PM

I'm not a gaming expert, but from what I've read, 512MB is no faster than 256MB for most current gaming applications, it's the throughput of the GPU that counts. Do any gamers out there want to comment on this?

I don't think that you could run Oblivion particularly well with only 256MB of Video Memory. Maybe you could, but you couldn't nearly max of the Graphics or else you would suffer some frame rate issues.

Eidorian

Sep 9, 2006, 12:53 PM

You may be right.

I couldn't find anything in the Intel technical documentation on the 945 to show a new revision or stepping of the chip - in fact most of the 945 docs at the Intel website don't mention the Core 2 at all....That's what tipped me off as well. I've had to hit those tech pages to glean information off of them.

wizard

Sep 9, 2006, 01:03 PM

I'm a little skeptical about Napa64 as well. I did read the previous links and articles on it. Why make Merom backwards compatible with Yonah?

What is to be skeptical about? Seriously Intel continuously improves its hardware, they have to or end up getting trounced on by the competition. In fact recent history with respect to AMD demonstrates what happens when they don't take a serious look at their hardware.

The issue with Merom and this iterations backward computability is that it gets INTEL 64 bit hardware to market fast as frankly they weren't even competing in that realm. For Intel 64 bit is serious issue as they are behind the eight ball or this one. It is an example of Intel being asleep at the wheel as they focused on who needs 64 bit instructions when a good part of the market demand was for addressable ram.

Nice information there. I wanted a little heads up on Kentsfield. Still, isn't it dual Conroe's with separate cache and then over the front side bus?

Frankly I haven't followed Kentsfield that much, more of an AMD man, but what is interesting to me with respect to this thread, is that MEROM the platform has a long way to go yet. That is the iMac just released is more or less a first generation implementation of Merom. I'm left with the impression that Apple just slapped the new processor in the old socket and gave us all a surprise this week. But that is what backward computability is all about. Mind you I know nothing about the current logic boards but I'm sure that information will float across the web soon.

Dave

Eidorian

Sep 9, 2006, 01:08 PM

What is to be skeptical about? Seriously Intel continuously improves its hardware, they have to or end up getting trounced on by the competition. In fact recent history with respect to AMD demonstrates what happens when they don't take a serious look at their hardware.

The issue with Merom and this iterations backward computability is that it gets INTEL 64 bit hardware to market fast as frankly they weren't even competing in that realm. For Intel 64 bit is serious issue as they are behind the eight ball or this one. It is an example of Intel being asleep at the wheel as they focused on who needs 64 bit instructions when a good part of the market demand was for addressable ram.I'm skeptical that Napa64 is a different chipset then the standard Intel 945 mobile series. Core 2 Duo works in the same socket as Yonah but somehow you need a Napa64 chipset to get full 64-bit addressing? They haven't changed a thing with the 945. Napa64 is just the 945 chipset with a Merom instead of a Yonah. We won't see any real change until we hit Santa Rosa.

Frankly I haven't followed Kentsfield that much, more of an AMD man, but what is interesting to me with respect to this thread, is that MEROM the platform has a long way to go yet.

DaveKentsfield is two Conroes on a single die. They don't share cache like the previous Pentium D chips. So they'll each have 4 MB of cache and then communicate over the front side bus.

Multimedia

Sep 9, 2006, 01:21 PM

Looks like MacCentral forgot to mention the fact that no matter how few cores an application can use - even if it's only ONE, the fact that more can be run at full speed SIMULTANEOUSLY is the whole reason for wanting-having-needing more cores - not wiether or not what you normally run can use 2, 3 or even all 4 cores at this time. OS X automatically delegates work to however many cores are vacant or underused so the user gets immediate benefit from 4 cores they will never get from 2. And I am 100% certain that tthe benefit is radically more than 20-30%.

It's an old think I always do only one thing at a time mentality that overlooks this otherwise obvious reason - a new way of working and a new way of thinking about how to do work - for going with more cores if you can afford it.

Eidorian

Sep 9, 2006, 01:23 PM

Looks like MacCentral forgot to mention the fact that no matter how few cores an application can use - even if it's only ONE, the fact that more can be run at full speed SIMULTANEOUSLY is the whole reason for wanting-having-needing more cores - not wiether or not what you normally run can use 2, 3 or even all 4 cores at this time. The OS delegates to however many cores are vacant or underused so the user gets immediate benefit from 4 cores they will never get from 2. And I am 100% certain that tthe benefit is radically more than 20-30%.

It's an old think I always do one thing at a time mentality that overlooks this otherwise obvious reason for going with more cores if you can afford it.Heh, that's pretty funny. I have quite a few applications that'll hit one core at 100%. (Q emulator is the best example) Luckily, even though it's not multi-threaded a have another core free to do my work while Q eats up 100% of one.

I run Windows 98 in Q for laughs. I liked Windows 98...

flopticalcube

Sep 9, 2006, 01:25 PM

AidenShaw

Sep 9, 2006, 01:27 PM

Kentsfield is two Conroes on a single die. They don't share cache like the previous Pentium D chips. So they'll each have 4 MB of cache and then communicate over the front side bus.
Minor terminology correction - the "die" is the silicon chip, the "package" is the carrier with the pins....

http://www.xbitlabs.com/web/2006-6-22.html

Kentsfield consists of two Conroe dies, each featuring two cores and 4MB of L2 cache.

See http://www.tomshardware.com/2005/12/04/top_secret_intel_processor_plans_uncovered/page4.html
(this shows a Pentium D image, but Kentsfield is doing the same trick.)

Eidorian

Sep 9, 2006, 01:27 PM

Maybe they should have run all their benchmarks at the same time!It also depends if you can run multiple instances of that application. A little help here Multimedia? I know you've used multiple instances of Toast. Care to enlighten us on what other applications we can do the same? Maybe we should make a guide on it...

http://www.xbitlabs.com/web/2006-6-22.html

Kentsfield consists of two Conroe dies, each featuring two cores and 4MB of L2 cache.
I thought so. We've beaten Core 2 Duo chips to death and their design.

Multimedia

Sep 9, 2006, 01:47 PM

We won't see any real change until we hit Santa Rosa.

Kentsfield is two Conroes on a single die. They don't share cache like the previous Pentium D chips. So they'll each have 4 MB of cache and then communicate over the front side bus.This is why I think Apple has got to be thinking about how they can put Kentsfield and then Tigerton - or perhaps beginning with Tigerton if it's a lot cooler - into the next generation of iMacs. If they stick with Merom, they will not get to four mobile cores for another 1-2 years at the soonest - as I understand the roadmap.

Do we know Kentsfield pricing yet? Probably not.

I think a redesign of the iMac's cooling system is imperative so that they can keep Kentsfield/Tigerton cool inside the new design - at least in the 23" model + a 30" model next year. I like the perforated edge approach as a best possible solution depicted here in this 30" aluminum iMac mock-up (I have no idea who created this - sorry). I'm thinking the bottom and top edges would want to be perforated as well. In this mock-up, it's not clear they are.

Body could still be plastic. But Aluminum is a great heat conductor so the whole body would be air cooling the insides like non-fan aluminum hard drive cases do today.

wizard

Sep 9, 2006, 01:49 PM

I'm skeptical that Napa64 is a different chipset then the standard Intel 945 mobile series. Core 2 Duo works in the same socket as Yonah but somehow you need a Napa64 chipset to get full 64-bit addressing? They haven't changed a thing with the 945. Napa64 is just the 945 chipset with a Merom instead of a Yonah. We won't see any real change until we hit Santa Rosa.

Well if it gives you 64 bit memory addressing then it certainly is a newer chip I'm not sure what you where expecting an new front side bus maybe? Maybe the chip set (945) is a modest upgrade but in the case of he IMac if it were implemented would have resulted in a larger address space for the PC. That is a real change. It is interesting that Apple apparently didn't implement Napa64 in the new iMac, I do wonder why as the release dates almost coincide.

As for Santa Rosa what there is so important to you that you want to wait? Just curious as I'm far from being in a position to purchase a new PC at the moment so this discussion doesn't really matter. It is more of a technical interest than anything else.

Kentsfield is two Conroes on a single die. They don't share cache like the previous Pentium D chips. So they'll each have 4 MB of cache and then communicate over the front side bus.
That sounds like a description for the old D model but you are saying a single die. Frankly it sounds like a dead end processor to me. Communications between the two subsections should be via a separate communications path. I have this feeling that the manufactures are rushing to quad core a little to fast.

Dave

bretm

Sep 9, 2006, 01:53 PM

And I'm thinking... why?! 10 years ago BeOS had this down pat. The whole system was multi-threaded and multi-processor aware from the kernel all the way up through the user interface including the system services used by all native applications. It was amazingly responsive and was reported (in major publications) to gain as much as 60-70% performance by having a second CPU. I realize MacOS X is based on some old NeXTStep code which was not made for multiple processors, but come on! This is the 21st century and Apple's been selling dual processor machines for about 5 years now.

Anyway, this is great news. I'd been drooling over the new iMacs since they were announced and wondering how much I might gain by upgrading from my 2GHz G5 PowerMac. It's very enticing.

Well they were selling them back in 1996 so you might want to add 5 years to your 5 year statement. You could buy dual 604e in the 9500 and the 9600 too I think.

Eidorian

Sep 9, 2006, 01:56 PM

Well if it gives you 64 bit memory addressing then it certainly is a newer chip I'm not sure what you where expecting an new front side bus maybe? Maybe the chip set (945) is a modest upgrade but in the case of he IMac if it were implemented would have resulted in a larger address space for the PC. That is a real change. It is interesting that Apple apparently didn't implement Napa64 in the new iMac, I do wonder why as the release dates almost coincide.

As for Santa Rosa what there is so important to you that you want to wait? Just curious as I'm far from being in a position to purchase a new PC at the moment so this discussion doesn't really matter. It is more of a technical interest than anything else.http://guides.macrumors.com/Merom

That sounds like a description for the old D model but you are saying a single die. Frankly it sounds like a dead end processor to me. Communications between the two subsections should be via a separate communications path. I have this feeling that the manufactures are rushing to quad core a little to fast.

DaveYou are correct it is a rushed quad core. At least we get more cores out a little faster. Though it's not the best implementation.

AidenShaw

Sep 9, 2006, 02:06 PM

http://guides.macrumors.com/Merom

You are correct it is a rushed quad core. At least we get more cores out a little faster.
The biggest advantage is that you get quad cores without having to pay for Xeon chipsets and memory.

It's also big for the Windows/Linux side of the world. Much of the software is licensed per socket.

- XP Home - 1 socket
- XP Pro - 2 sockets
- Win2k3 Server - 4 sockets

With a quad core, you can run an 8 CPU XP Pro system without forking over the bucks for Windows Server. Add to that per-socket licensing for many software packages, and it's a huge cost savings.

Though it's not the best implementation.
Careful here - it's almost as good as the current Mac Pro quad configuration. There you have two dies communicating over the FSB and Northbridge...

Eidorian

Sep 9, 2006, 02:09 PM

The biggest advantage is that you get quad cores without having to pay for Xeon chipsets and memory.

It's also big for the Windows/Linux side of the world. Much of the software is licensed per socket.

- XP Home - 1 socket
- XP Pro - 2 sockets
- Win2k3 Server - 4 sockets

With a quad core, you can run an 8 CPU XP Pro system without forking over the bucks for Windows Server. Add to that per-socket licensing for many software packages, and it's a huge cost savings.

Careful here - it's almost as good as the current Mac Pro quad configuration. There you have two dies communicating over the FSB and Northbridge...Oh yeah, I forgot about the Windows socket limitations. I know it'll be a great performer but a "better" chip will always come out later. Kentsfield appears to be an Extreme Edition chip until quad core trickles down to more normal desktops. Still, I can see some new Mac Pro running off a single Kentsfield.

Multimedia

Sep 9, 2006, 02:11 PM

It also depends if you can run multiple instances of that application. A little help here Multimedia? I know you've used multiple instances of Toast. Care to enlighten us on what other applications we can do the same? Maybe we should make a guide on it...Preemble clarification: I use Toast (http://www.roxio.com/enu/products/toast/titanium/overview.html) in a highly unorthodox way - nothing to do with writing DVDs or CDs. I use it most of the time to write DVD IMAGES that Handbrake (http://handbrake.m0k.org/) understands how to make priistine mp4 files from. I am able to reduce a 4.3GB original EyeTV HD broadcast recording down to 351MB using this method. The result is an excellent, albeit soft, version of the original that can go on an iPod or two on a CD and when played on an analog TV from the iPod looks just like a DVD. On a HD monitor it still looks great. Just a little soft. Sound quality is identical to the original.

I haven't explored what else we can run simultaneously beyond Toast (http://www.roxio.com/enu/products/toast/titanium/overview.html) and Handbrake (http://handbrake.m0k.org/). I can run as many instances of those as I like. But I run out of cores even just running both of them because they will each use more than two cores given the chance to run alone. Running them simultaneously even with a second Handbrake running third, still gets all the jobs done faster than waiting for two to run and then running the third. Handbrake will process up to about 150-160 fps when two copies are running while it will process only about 93-100 fps alone.

Handbrake FPS readings vary a lot between the analysis pass and the writing pass - much slower writing on the second pass than studying-planning the writing scheme on the first pass on both the Quad and the Mac Pro. On the Mac Pro, Toast will use almost all 4 cores given no competition. But so far I'm not convinced it is encoding EyeTV recordings for DVD images much faster than it does on teh Quad - yes 7.1 UB. I need to go back and exact time some encodes on the Mac Pro then compare that here on the Quad to tell.

Just tried to launch a second copy of EyeTV and it's a no go. Maybe if I have another liscense with another tuner like the new hybrid it will work with a second copy - don't know yet. Probably getting an EyeTV hybrid tuner (http://www.elgato.com/index.php?file=products_eyetvhybridna) soon so I can record two HD shows at once.

A Multi-Instance and Multi-Core Usage Guide would be a great help. Does someone with authorization want to start a thread on this subject? I am not authorized to create new threads. But I would be happy to contribute to it. If someone with new thread creation permission does it, please post a link to it here. Thank you.

Multimedia

Sep 9, 2006, 02:15 PM

Oh yeah, I forgot about the Windows socket limitations. I know it'll be a great performer but a "better" chip will always come out later. Kentsfield appears to be an Extreme Edition chip until quad core trickles down to more normal desktops. Still, I can see some new Mac Pro running off a single Kentsfield.Yeah that will be the Mac Pro Jr. while the rest of the Mac Pros will be running pairs of Clovertowns.

BRLawyer

Sep 9, 2006, 02:16 PM

Well they were selling them back in 1996 so you might want to add 5 years to your 5 year statement. You could buy dual 604e in the 9500 and the 9600 too I think.

You are right, but if I remember well the 9500 had 604s, not 604e...and actually the 9600 had the best-ever case for any Mac...you just had to pull down one of the panels to have full access to everything in the MOBO and drives...really beautiful...

http://www.everymac.com/systems/apple/powermac/media/easytower_movie.html

Evangelion

Sep 9, 2006, 02:18 PM

Well if it gives you 64 bit memory addressing then it certainly is a newer chip

64bit addressing arrives with the new cpu. so the point is that napa64 isn't really new, it just uses merom instead of yonah.

epitaphic

Sep 9, 2006, 02:25 PM

Handbrake FPS readings vary a lot between the analysis pass and the writing pass - much slower writing on the second pass than studying-planning the writing scheme on the first pass on both the Quad and the Mac Pro.
That's because the second pass only uses one core.

Eidorian

Sep 9, 2006, 02:26 PM

Preemble clarification: I use Toast in a highly unorthodox way - nothing to do with writing DVDs or CDs. I use it most of the time to write DVD IMAGES that Handbrake understands how to make priistine mp4 files from. I am able to reduce a 4.3GB original EyeTV HD broadcast recording down to 351MB using this method. The result is an excellent, albeit soft, version of the original that can go on an iPod or two on a CD and when played on an analog TV looks like a DVD. On a HD monitor it still looks great. Just a little soft.

I haven't explored what else we can run simultaneously beyond Toast and Handbrake. I can run as many instances of those as I like. But I run out of cores even just running both of them because they will each use more than two cores given the chance to run alone. Running them simultaneously even with a second Handbrake running third, still gets all the jobs done faster than waiting for two to run and then running the third. Handbrake will process up to about 150-160 fps when two copies are running while it will process only about 93-100 fps alone.

Handbrake FPS readings vary a lot between the analysis pass and the writing pass - much slower writing on the second pass than studying-planning the writing scheme on the first pass on both the Quad and the Mac Pro. On the Mac Pro, Toast will use almost all 4 cores given no competition. But so far I'm not convinced it is encoding EyeTV recordings for DVD images much faster than it does on teh Quad - yes 7.1 UB. I need to go back and exact time some encodes on the Mac Pro then compare that here on the Quad to tell.

Just tried to launch a second copy of EyeTV and it's a no go. Maybe if I have another liscense with another tuner like the new hybrid it will work with a second copy - don't know yet. Probably getting a hybrid tuner yet so I can record two shows at once.

A Multi-Instance and Multi-Core Usage Guide would be a great help. Does someone with authorization want to start a thread on this subject? I am not authorized to create new threads. But I would be happy to contribute to it. If someone with new thread creation permission does it, please post a link to it here. Thank you.Well anyone here can start a Guide on the wiki. Ask around if anyone else knows more on the subject. Otherwise I picked you from our previous ramblings on Core 2 Duo and quad-core machines.

Edit: Interesting usage of Toast/Handbrake there.

64bit addressing arrives with the new cpu. so the point is that napa64 isn't really new, it just uses merom instead of yonah.I guess that resolves the Napa32/64 argument. If there ever was one...

Yeah that will be the Mac Pro Jr. while the rest of the Mac Pros will be running pairs of Clovertowns.Cube? 24" iMac?

Multimedia

Sep 9, 2006, 02:30 PM

That's because the second pass only uses one core.No. it's still using more than one core. More likely because of the speed limitation of the hard drive writing the mp4 file.

Eidorian

Sep 9, 2006, 02:37 PM

No. it's still using more than one core. More likely because of the speed limitation of the hard drive writing the mp4 file.I know this sounds silly but how do you monitor processor usage from a process via Activity Monitor? (I have the %CPU column up but is there a way to get more detailed information?) I have the Developer Tools installed too. I'm not a developer but well...my work requires me to have them installed anyways.

BRLawyer

Sep 9, 2006, 02:42 PM

Cube? 24" iMac?

Actually it's gonna the 30th Ann. Mac. Apple is not doing a mass-produced headless Mac.

wizard

Sep 9, 2006, 02:42 PM

http://guides.macrumors.com/Merom

Yeah it is interesting but in the context of a desktop machine you are not getting a lot for the wait. A new front side bus and a Merom to go with it. AND 64 Bit support which can be very important for some.

I guess what I'[m saying is that if you are willing to wait for this upgrade then you really don't need a new computer even with this rather significant update to the iMac. Maybe that is where our paths diverge as I see this as a significant upgrade. Sure it is a stop gap measure for 64 bit support but it does offer significant performance advantages and should adapt well to Apples move to 64 bit.

You are correct it is a rushed quad core. At least we get more cores out a little faster. Though it's not the best implementation.

That is what I thought but like I said I don't follow Intel deeply. I do know that with Core 2 Intel has the potential for significant upside on clock rates. It looks like we could see both a core race and a clock rate race again. As to AMD I'm not 100% on their quad either but I think it is a single chip implementation. Maybe a few moths slower in coming but the impression is a solid offering.

What I'm wondering is where the optimal number of cores is for the average desktop user. I know that dual has some pretty amazing results on the desktop so how far do we go for core wise. 4, 8, 12 or more? Especially on i86, it is to bad the PPC guys never got their acts together.

Dave

Multimedia

Sep 9, 2006, 02:43 PM

I know this sounds silly but how do you monitor processor usage from a process via Activity Monitor? I have the Developer Tools installed too. I'm not a developer but well...my work requires me to have them installed anyways.Yes that's right. I always have Activity Monitor on so I can see exactly what's going on with my four cores. I have the sort on the percentage column on the left followed by the application name and then I stick it in the lower right corner of my two screens. By keeping it open I can make sure nothing has crashed.

Both Toast and Handbrake occasionally crash during an encode or even while Toast is writing the image after an encode. Occasionally it's due to a bad original file MPEG2 glitch that will keep causing Toast to crash repeatedly. But ususally I can relaunch and re-run the process and it works fine the second time.

Multimedia

Sep 9, 2006, 02:56 PM

Yeah it is interesting but in the context of a desktop machine you are not getting a lot for the wait. A new front side bus and a Merom to go with it. AND 64 Bit support which can be very important for some.

I guess what I'[m saying is that if you are willing to wait for this upgrade then you really don't need a new computer even with this rather significant update to the iMac. Maybe that is where our paths diverge as I see this as a significant upgrade. Sure it is a stop gap measure for 64 bit support but it does offer significant performance advantages and should adapt well to Apples move to 64 bit.

I do know that with Core 2 Intel has the potential for significant upside on clock rates. It looks like we could see both a core race and a clock rate race again.

What I'm wondering is where the optimal number of cores is for the average desktop user. I know that dual has some pretty amazing results on the desktop so how far do we go for core wise. 4, 8, 12 or more? Especially on i86, it is to bad the PPC guys never got their acts together.

DaveWell I am 100% certian four is not enough. But I have read there may be diminishing returns beyond 16 perhaps even beyond 8 - I have no idea. But Clovertown can't come soon enough for me. Would be great if the speed will rise rapidly as well next year. Hope you're right. With both up a lot, what I am doing could be done in a few minutes insead of hours. That's my dream.

Multimedia

Sep 9, 2006, 03:16 PM

Heh, that's pretty funny. I have quite a few applications that'll hit one core at 100%. (Q emulator is the best example) Luckily, even though it's not multi-threaded a have another core free to do my work while Q eats up 100% of one.

I run Windows 98 in Q for laughs. I liked Windows 98...Exactly. A perfect example where one application topping out on only one core leaves the entire other one for other stuff. Probably a good thing to have that limit.

In fact, in future, I could see where application developers let the user in preferences tell the application how many cores to be allowed to use. Give the user a choice of how many cores he/she wants a particular process to use. That would be a way cool improvement in all application preferences. Would prevent any one applicaiton from hosing the computer due to core hogging.

Eidorian

Sep 9, 2006, 03:24 PM

Yes that's right. I always have Activity Monitor on so I can see exactly what's going on with my four cores. I have the sort on the percentage column on the left followed by the application name and then I stick it in the lower right corner of my two screens. By keeping it open I can make sure nothing has crashed.It looks like Activity Monitor is the best approach then. Thanks though!

Well I am 100% certian four is not enough. But I have read there may be diminishing returns beyond 16 perhaps even beyond 8 - I have no idea.We haven't hit that wall yet. ;)

Exactly. A perfect example where one application topping out on only one core leaves the entire other one for other stuff. Probably a good thing to have that limit.

In fact, in future, I could see where application developers let the user in preferences tell the application how many cores to be allowed to use. Give the user a choice of how many cores he/she wants a particular process to use. That would be a way cool improvement in all application preferences.I have to agree. Single or multi core affinity would let you pick and maximize your usage.

flopticalcube

Sep 9, 2006, 03:26 PM

In fact, in future, I could see where application developers let the user in preferences tell the application how many cores to be allowed to use. Give the user a choice of how many cores he/she wants a particular process to use. That would be a way cool improvement in all application preferences. Would prevent any one applicaiton from hosing the computer due to core hogging.
That would be an OS issue, would it not? It would be up to the OS to allocate cores to processes.

Eidorian

Sep 9, 2006, 03:30 PM

That would be an OS issue, would it not? It would be up to the OS to allocate cores to processes.Yes an OS issue.

http://www.math.purdue.edu/~abarreno/affinitydlg.gif

Multimedia

Sep 9, 2006, 03:39 PM

Yes an OS issue.

http://www.math.purdue.edu/~abarreno/affinitydlg.gifWow so if that's in XP already it's gotta be a feature in Leopard.

You call that Application Core Affinity or what's the correct full termonology? And where in the OS do you choose the applications to assign x number of cores with that dialog box. Looks like they're ready for a lot of cores coming up?!?! :eek:

32. I'd say that's planning ahead.

Are we hijacking this thread? Don't mean to go so far Off Topic folks. :D

Eidorian

Sep 9, 2006, 03:40 PM

Wow so if that's in XP already it's gotta be a feature in Leopard.

You call that Application Core Affinity or what's the correct full termonology? And where in the OS do you choose the applications to assign x number of cores with that dialog box. Looks like they're ready for a lot of cores coming up?!?! :eek:

32. I'd say that's planning ahead.I think it might be in Windows 2000 as well. It's found via Task Manager under Processes. Right click on a process in the list and you can assign its affinity. Some programs crash when the encounter hyper threading or multi core machines. So you have to assign the process to a single CPU/core. More then likely on a dual processor machine from back then a multi-core one.

Multimedia

Sep 9, 2006, 03:48 PM

I think it might be in Windows 2000 as well. It's found via Task Manager under Processes. Right click on a process in the list and you can assign its affinity. Some programs crash when the encounter hyper threading or multi core machines. So you have to assign the process to a single CPU/core. More then likely on a dual processor machine from back then a multi-core one.Seems like the application developers could add a link to such a feature in their code so the user could just assign core volume in each application prefs that would tell the system what amount to assign to that process. Maybe even have the system do that automatically to all applicaiton preferences so the choice appears in all general preference panes of each application.

Eidorian

Sep 9, 2006, 03:52 PM

Seems like the application developers could add a link to such a feature in their code so the user could just assign core volume in each application prefs that would tell the system what amount to assign to that process. Maybe even have the system do that automatically to all applicaiton preferences so the choice appears in all general preference panes of each application.I don't know if it's up to Apple or the program developers to allow CPU/core assignment.

Could you get on an IM? I'd like to talk about the Guide we were talking about.

Half Glass

Sep 9, 2006, 03:55 PM

Benchmarks look good...real good for a laptop chip! I wonder how the laptops will fare!

Loving the MacPro, and just like you guys waiting for some of the software to be seriously optimized for more than 2 cores. I've seen ~289% in spurts but not more from an app while encoding some video (I think the 289% was Handbrake). FCP seems to keep itself under 2 cores.

Bring on the truly optimized software and lets roll!

--Half Glass

AidenShaw

Sep 9, 2006, 04:00 PM

Wow so if that's in XP already it's gotta be a feature in Leopard.

You call that Application Core Affinity or what's the correct full termonology? And where in the OS do you choose the applications to assign x number of cores with that dialog box.
Yes, Windows has APIs to set affinity masks (a mask representing a set of one or more cores) at both the process and thread level. Thread affinity must be a subset of process affinity.

The task manager lets you set that from the GUI.

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dllproc/base/setthreadaffinitymask.asp
SetThreadAffinityMask

Sets a processor affinity mask for the specified thread.

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dllproc/base/setthreadaffinitymask.asp
SetProcessAffinityMask

Sets a processor affinity mask for the threads of the specified process.

Looks like they're ready for a lot of cores coming up?!?! :eek:

32. I'd say that's planning ahead.
On 64-bit Windows, it can show up to 64 processors....

By the way, the image a few back showed an old version of Windows, the current version only shows the number of CPUs available on the current system (you only see 64 on a big box ;) ).

EagerDragon

Sep 9, 2006, 04:06 PM

Well, wasn't the iMac G5 restricted to 2GB, yet it was a 64-bit processor? A 32-bit computer can take up to 4GB, but due to the hardware Apple was/is using, they can't even take this.

What i find odd is that it appears to allow 1 or 2GB in either slot, but no more than 3GB in total. That is obviously the maximum the board can take, but it would have made a little more sense to allow 2GB in each. This will not really effect it's ability to run 64-bit software, just restricts how much memory can be used. Remember that you have been able to get AMD systems with 64-bit processors for some time now. They won't take more than 4GB, but will allow you to run 64-bit OSes and Apps.

I'm hoping by the time I'm after an iMac, it will take at least 4GB, have Blu-Ray as an option, (although I may opt for standard Superdrive if it is an option and buy a Mac compatible external later), include bigger hard drives and stick to a similar price point to now.

I'm tempted by the 20" now, but am not buying yet and would want about 320-400GB in there for the same price, perhaps even 2GB RAM. I've got time to wait however.

I could be wrong but...I think Manic did hit it on the head. I think some of the hardware is still 32 bits. If the memory address leads are there for 3 Gigs then they are there to the max of 4 Gigs, above that is anyones guess. Yes it is very strange since either slot can take the 2 gig chip.

Obviously the Mac Pro does not have that issue.

EagerDragon

Sep 9, 2006, 04:30 PM

I like to see some benchmarks against G4 machines and see if they make me pull the credit card out.

epitaphic

Sep 9, 2006, 05:09 PM

I like to see some benchmarks against G4 machines and see if they make me pull the credit card out.

They've done that in this benchmark, tho only on the Speedmark test:
Speedmark 4.5 scores are relative to those of a 1.25GHz Mac mini, which is assigned a score of 100.
If you go by that, the new iMacs are about 2.5x faster.

BRLawyer

Sep 9, 2006, 05:35 PM

Seems like the application developers could add a link to such a feature in their code so the user could just assign core volume in each application prefs that would tell the system what amount to assign to that process. Maybe even have the system do that automatically to all applicaiton preferences so the choice appears in all general preference panes of each application.

Isn't that the same thing as assigning priorities to processes in OS X? Terminal or Developer Tools already do that, as well as several freeware apps...

Eidorian

Sep 9, 2006, 05:40 PM

Isn't that the same thing as assigning priorities to processes in OS X? Terminal or Developer Tools already do that, as well as several freeware apps...I believe Multimedia wants something a bit simpler then that though.

Oh and explanation/links for those tools/freeware?

freiheit

Sep 9, 2006, 06:20 PM

What i find odd is that it appears to allow 1 or 2GB in either slot, but no more than 3GB in total. That is obviously the maximum the board can take, but it would have made a little more sense to allow 2GB in each.

I wholeheartedly agree. I'm at a point (thanks to apps like Parallels) where I am ready to replace both my PC and my G5 PowerMac with a single system. I can't really justify the cost of the MacPro (even the low-end is over $2K) but if I'm to be running both Mac and Windows simultaneously, 4GB RAM capacity in the iMac would be very much appreciated. 3GB would do, but the cost difference between the 2GB and 3GB option right now is extremely high ($575 more to go from 2GB to 3GB, not $575 total for 3GB) which puts the 17" iMac right in the price range of the low-end MacPro.

Did anyone else also notice the verbage in the "how much RAM do I need" section on the iMac store page that says you can choose a "one SO-DIMM" option to keep 2 slots free? Why do I not see this as an option? There's 2x512MB, 2x1GB and 1x2GB+1x1GB.

GreatOne08

Sep 9, 2006, 06:27 PM

Those new iMac are looking better and better. Need to save up now.:D

iGary

Sep 9, 2006, 06:34 PM

Why does it say "MacCentral" when MacWorld did the benchmarks? :confused:

mashinhead

Sep 9, 2006, 06:48 PM

I want to see some unpacking pics of that 24inch model compared with the 20in. Soon enough I suppose.

yeah that 24 incher is a very tempting deal.

I have a Dual Core 2.0 G5 and a 20 inch monitor. I'm considering selling just the G5 and Getting a Low End Mac Pro. Or selling both and getting the iMac. I know that MacPro is better, but the price compared to what you get with the iMac is just too good. So torn.:rolleyes:

But also looking at My 20 inch ACD right now, I can't help but think how F*#king huge that iMac must be. If they made an all black one thouth it would be over. I wish they made black ACD's and MacPro's too now that i think bout it.

AidenShaw

Sep 9, 2006, 07:25 PM

Isn't that the same thing as assigning priorities to processes in OS X? Terminal or Developer Tools already do that, as well as several freeware apps...
No, not at all.

An affinity mask sets the set of CPUs that can be scheduled. A job won't be run on another CPU, even if the assigned CPUs are at 100% and other idle CPUs are available.

And that, by the way, is why setting affinity is usually a bad idea. Let the system dynamically schedule across all available resources -- or you might have some CPUs very busy, and others idle.

Win2k3 also has "soft" affinity masks, which define a preferred set of CPUs. If all of the preferred CPUs are busy, and other CPUs are idle, then soft affinity allows the system to run the jobs on the idle CPUs - even though the idle CPUs aren't in the preferred affinity mask.

firsttube

Sep 9, 2006, 07:41 PM

minimum post requirements suck, if you've been lurking for years, but never post you still can't post in the marketplace. sorry this is OT

Some_Big_Spoon

Sep 9, 2006, 07:47 PM

EagerDragon

Sep 9, 2006, 07:59 PM

I stopped at the Apple store this morning and tried out the 24 inch iMac and the Mac Pro. These are sweet machines. No did not buy anything.

The systems both had 1 gig on them and I compared them to a MacBook Pro. One weird thing.... the 24 incher had some stuttering on the iMoviedemo they all had. However the second time I tried it it was smoth as silk. I think it was not cached in memory and since the second time it was it ran smothly. I was also plesantly surpriced that the 24 incher screen was very readable at it highest setting even with my bad eyes. Nice screen realstate and resolution with nice easy to read fonts.

Im still waiting for Leopard to release these powerful anymals of their chains. By them the systems will be even better, maybe even incluse Santa Rosa.

imikem

Sep 9, 2006, 08:42 PM

I think we all knew that Merom would only bring modest performance gains.

Core 2 is a significantly different beast architecturally from Yonah to Merom. Merom has Intel's clone of AMD's cloned/extended x86 instruction set*, 64-bit instructions as well as long overdue changes to handling of old instructions, allowing this generation of CPUs to better utilize registers.

There are other enhancements in Core 2 as well, so I doubt that the current compilers are getting the full performance potential. It may be several months before updated compilers can properly optimize code for Core 2. Stay tuned.

* Sorry about that - x86 architecture is not pretty to look at. I sure liked the elegance of the PPC instruction set, but guess what $$Billions$$ can do?

BRLawyer

Sep 9, 2006, 08:58 PM

chatin

Sep 9, 2006, 08:59 PM

Core 2 is a significantly different beast architecturally from Yonah to Merom. Merom has Intel's clone of AMD's cloned/extended x86 instruction set*, 64-bit instructions as well as long overdue changes to handling of old instructions, allowing this generation of CPUs to better utilize registers.

The Yonah is not related to Intel's big disaster chip, the Pentium D 810, but was botched to the point that the engineers turned off EMT64!

JSchwage

Sep 9, 2006, 09:10 PM

All I can say is I can't wait for the Merom Macbooks! Speed increases are always nice things. :D

imikem

Sep 9, 2006, 09:29 PM

The Yonah is not related to Intel's big disaster chip, the Pentium D 810, but was botched to the point that the engineers turned off EMT64!

Really? I had understood that Yonah was close architecture wise to the previous Pentium M, while Merom represents the first "true" Core architecture.

Cheers.

chatin

Sep 9, 2006, 09:48 PM

Really? I had understood that Yonah was close architecture wise to the previous Pentium M, while Merom represents the first "true" Core architecture.

True, the Pentium M (Mobil Centrino) was a hudge sucess for Intel! The Pentium D (Desktop) was a dual-core disaster, pushing the old "NetBurst" Pentium 4 past all safe design limits.

Core 2 is the all new rework that saved Intel!

aafuss1

Sep 9, 2006, 09:49 PM

Nice Core 2 performance-just what we need.

Rocketman

Sep 9, 2006, 09:58 PM

Any word on the difference between the C2D 2.16 and the 2.33? Is it worth the upgrade price?

The benefit should be nearly linear, so not much. I suggest putting the price difference in a fund toward your next computer in 2 years, when something even more cool and hard top pass up will be available.

Rocketman

AidenShaw

Sep 9, 2006, 10:02 PM

But I am pretty sure the newest developer tools can cope with that, considering that multicore chips are a rather new thing in the mainstream market...

Try the Processor Preferences app contained in the Apple CHUD tools, for instance...
Please explain - I have no idea what "that" is....
---

Regardless of the tool, however, it is usually much better to let the OS dynamically schedule threads across the cores. Unless the programmer has some reason to try to control this, the alternative is some resources (CPUs) being overcommitted, while other CPUs are idle.

It doesn't matter who has the better tools - it's usually better to let the OS decide microsecond by microsecond how best to schedule the CPUs, than to have the developer make those decisions at edit time.

I've used the SetProcessAffinityMask APIs fairly often, but it's always been for specific test or benchmark situations. I have a hard time thinking of a situation where a general application would want to statically control the scheduler - it's just "bad think" to even try. (Except for those weird-a$$ NUMA Opterons - you can be really scr3wed if you have to go through HyperTransport to get to memory. I check NUMA topology, and use affinity to keep the AMD architecture from killing me.)

bluewire

Sep 9, 2006, 10:11 PM

Driving 1.5 hours to the Apple store this morning and the same on the way back. But I am not buying yet, just looking and getting a feel for the entire line. Oh I forgot.... and turning green with envy. Boy is going to be hard.

I've been calling around, there are 3 stores near me but none have a 24" iMac for me to look at yet. :( I'm looking to see what the annoucement is on Tuesday is...Cube redux? :eek:

dondark

Sep 10, 2006, 12:12 AM

I've been calling around, there are 3 stores near me but none have a 24" iMac for me to look at yet. :( I'm looking to see what the annoucement is on Tuesday is...Cube redux? :eek:

It is Metal iPod Nano, Macbook Nano, Full HD MBP.

houttbe

Sep 10, 2006, 01:53 AM

I stopped at the Apple store this morning and tried out the 24 inch iMac and the Mac Pro. These are sweet machines. No did not buy anything.

Is the 24" as quiet as the MacPro? Have you been able to compare to the 20"?

ergle2

Sep 10, 2006, 02:34 AM

True, the Pentium M (Mobil Centrino) was a hudge sucess for Intel! The Pentium D (Desktop) was a dual-core disaster, pushing the old "NetBurst" Pentium 4 past all safe design limits.

Core 2 is the all new rework that saved Intel!

Core 2 isn't "all new". It's an evolutionary design based on Core tho some parts are borrowed from other Intel designs (the Memory Disambiguation tech was originally designed for the unreleased, unlamented Tejas, for example).

Other changes include a full 128-bit path to the SSE registers, meaning that all SSE instructions can now complete in a single cycle, L2 shared cache instead of separate L2's per CPU, an extra integer unit, etc.

And, of course, the 64bit extensions :)

Sure, there's enough in the way of changes/additions to render it worthy of being considered a new microarch, but those changes are evolutionary.

Ironically enough, there's a direct line from Core 2 going all the way back to P6, whereas NetBurst really was "all new"!

ergle2

Sep 10, 2006, 02:41 AM

Please explain - I have no idea what "that" is....
---

Regardless of the tool, however, it is usually much better to let the OS dynamically schedule threads across the cores. Unless the programmer has some reason to try to control this, the alternative is some resources (CPUs) being overcommitted, while other CPUs are idle.

It doesn't matter who has the better tools - it's usually better to let the OS decide microsecond by microsecond how best to schedule the CPUs, than to have the developer make those decisions at edit time.

I've used the SetProcessAffinityMask APIs fairly often, but it's always been for specific test or benchmark situations. I have a hard time thinking of a situation where a general application would want to statically control the scheduler - it's just "bad think" to even try. (Except for those weird-a$$ NUMA Opterons - you can be really scr3wed if you have to go through HyperTransport to get to memory. I check NUMA topology, and use affinity to keep the AMD architecture from killing me.)

I've owned SMP machines in the past and often found it more useful to force CPU affinity of CPU-heavy tasks to a single processor, as Windows 2000 (which was current at the time) by default had a habit of swapping it between chips, resulting in a lot of cache-dirtying. I think it was the load balancing code, but it's been a while now and I don't have those machines handy currently. However, you could see some significant improvement in processing time on some non-parallelizable cpu-bound tasks.

I've no idea if MacOS does this, but at least in the case of Core 2 it shouldn't matter anywhere near as much, as the L2 is fully shared.

JZ Wire

Sep 10, 2006, 03:13 AM

I was credit card in hand when these were released, but I stopped myself. I'd like to wait a bit and see the 64 bit boost (if there is any), and Leopard in general.

I feel like these are speed demons, but I can't take advantage of a lot of it due to my heavy use of CS2 and the in-between feeling of Apple's apps/OS right now.

The second Leopard is out, I'm on the 24" iMac train.

Yea, I know what you mean. Apple needs to get on the ball with that already. They have been shipping dual core machines for a while yet OS X cant truly take advantage of it. Since intel will be using more cores as time goes by it only makes sense for OS X and it's apps to be able to harness the full power of all the cores/processors. I really really hope thats what they have planned for Leopard. Maybe its one of the " Super Secret Features" ??? :rolleyes:

BRLawyer

Sep 10, 2006, 06:02 AM

Please explain - I have no idea what "that" is....
---

Regardless of the tool, however, it is usually much better to let the OS dynamically schedule threads across the cores. Unless the programmer has some reason to try to control this, the alternative is some resources (CPUs) being overcommitted, while other CPUs are idle.

It doesn't matter who has the better tools - it's usually better to let the OS decide microsecond by microsecond how best to schedule the CPUs, than to have the developer make those decisions at edit time.

I've used the SetProcessAffinityMask APIs fairly often, but it's always been for specific test or benchmark situations. I have a hard time thinking of a situation where a general application would want to statically control the scheduler - it's just "bad think" to even try. (Except for those weird-a$$ NUMA Opterons - you can be really scr3wed if you have to go through HyperTransport to get to memory. I check NUMA topology, and use affinity to keep the AMD architecture from killing me.)

I also agree it's not the best strategy to deal with CPU scheduling...my example is linked to the following page, I presume...perhaps core affinity scheduling is also there:

http://developer.apple.com/documentation/Darwin/Reference/ManPages/man1/hwprefs.1.html

EagerDragon

Sep 10, 2006, 09:47 AM

Is the 24" as quiet as the MacPro? Have you been able to compare to the 20"?
The store is noisy, so it is hard to say. To me neither was making a sound.

EagerDragon

Sep 10, 2006, 10:07 AM

Yea, I know what you mean. Apple needs to get on the ball with that already. They have been shipping dual core machines for a while yet OS X cant truly take advantage of it. Since intel will be using more cores as time goes by it only makes sense for OS X and it's apps to be able to harness the full power of all the cores/processors. I really really hope thats what they have planned for Leopard. Maybe its one of the " Super Secret Features" ??? :rolleyes:

I think that Apple is still working on a gamer machine. I can only assume tat it will be Mac Pro based or a new enclosure. They currently do not support SLI (need support), are using buffered memory (slow), and needs to be user upgradable. I think it is coming but not untl some time next year.

Open GL suporting multiple cores and multiple GPU power agregation is likely to be the ticket so insane performance that blows the rest. But I think it will wait until Leopard to be released. I am sure the next revision 10.4.8 will take us part of the way, but multi cpu, multi-package, multicores are comming. In Jan we should see 2 Quad cores in the Mac Pro and an even better dual Quad in summer. Only Leopard will be able to unleash the power.

sam10685

Sep 10, 2006, 10:14 AM

Wonder how the 24" iMac at 2.33GHz will fare.

i don't know-- but i have a feeling it'll be really fast and a good seller. i'd go to the Apple Store down town and look at the 20" iMac and think "goodness... any bigger and it wouldn't fit on the table." now, for the same price as a 30'' ACD, you get a monitor that is just a little smaller than the 30" PLUS you get a really really good computer. if Apple does't sell a large ammount of these than something is wrong.

SiliconAddict

Sep 10, 2006, 05:08 PM

Remember everyone. Intel sucks and Steve Jobs is nuts for going with them. :rolleyes: Just a reminder of the comments typical after '05's WWDC.

AidenShaw

Sep 11, 2006, 12:48 AM

I've owned SMP machines in the past and often found it more useful to force CPU affinity of CPU-heavy tasks to a single processor, as Windows 2000 (which was current at the time) by default had a habit of swapping it between chips, resulting in a lot of cache-dirtying....
However, you could see some significant improvement in processing time on some non-parallelizable cpu-bound tasks.
I came to the opposite conclusion....

Running many compute-bound single-threaded benchmarks and apps - I saw how NT (pre Win2k) would balance across CPUs (that is, a "100%" compute-bound job would show each CPU running at 50%).

However, setting affinity so that one CPU was 100% and the other was 0% had no significant effect on the run times. (And by "significant" I mean statistically significant - I literally ran hundreds of runs in each configuration.)\\

By the way, with Win2k3 (and XP 64-bit, really the same system) you see much less "balancing" - a single-threaded app will stick to a CPU for much longer.

AidenShaw

Sep 11, 2006, 12:53 AM

They ... are using buffered memory (slow)
Have you seen the benchmarks?

The Xeon systems scream, even with the "slow" memory.

While some contrived tests showed real latency issues with the FB-DIMM memory, for real-life applications the faster busses and large L2 caches make it a non-issue.

Focus on *system* performance, not on a particular detail.

ergle2

Sep 11, 2006, 01:13 AM

I came to the opposite conclusion....

Running many compute-bound single-threaded benchmarks and apps - I saw how NT (pre Win2k) would balance across CPUs (that is, a "100%" compute-bound job would show each CPU running at 50%).

However, setting affinity so that one CPU was 100% and the other was 0% had no significant effect on the run times. (And by "significant" I mean statistically significant - I literally ran hundreds of runs in each configuration.)\\

By the way, with Win2k3 (and XP 64-bit, really the same system) you see much less "balancing" - a single-threaded app will stick to a CPU for much longer.

I suspect if any observable difference occurs depends upon the application, dataset, etc.

I'm guessing the 50% "balanced" method was done to try and keep a single CPU from heating up too much, and with the advent of multicore systems, it probably no longer matters which core is generating the heat due to them being in a single package.

It could also be MS found that certain circumstances (like mine) resulted in improvements in processing.

Interesting stuff.

MacinDoc

Sep 11, 2006, 01:34 AM

Have you seen the benchmarks?

The Xeon systems scream, even with the "slow" memory.

While some contrived tests showed real latency issues with the FB-DIMM memory, for real-life applications the faster busses and large L2 caches make it a non-issue.

Focus on *system* performance, not on a particular detail.
Aiden, it's just not like you to make a statement like this without adding the links...

ImAlwaysRight

Sep 11, 2006, 08:01 AM

I'm getting a 17" 2 ghz C2D iMac... Sounds exciting with all that speed! I don't know what to do with it! I guess my iTunes visualizations will be smoother, and I'll have the Dashboard ripple effect.
$1200K for a little eye candy? Woo-hoo! (Seriously, though, well worth the $ over an eMac G4 1GHz)

The overall score was 10% better in the new model

That's the bottom line, folks. Not a huge performance improvement, by any means. Anyone trading in a Yonah iMac for one of these Meroms has got to be nuts. If you are looking for any type of performance improvement, wait a year or two to upgrade.

gnasher729

Sep 11, 2006, 08:09 AM

now is the Xeon processor faster than the core 2 duo?

That is a rather pointless question since the Xeon processor in the MacPro _is_ a Core 2 Duo processor.

gnasher729

Sep 11, 2006, 08:42 AM

No, not at all.

An affinity mask sets the set of CPUs that can be scheduled. A job won't be run on another CPU, even if the assigned CPUs are at 100% and other idle CPUs are available.

And that, by the way, is why setting affinity is usually a bad idea. Let the system dynamically schedule across all available resources -- or you might have some CPUs very busy, and others idle.

Win2k3 also has "soft" affinity masks, which define a preferred set of CPUs. If all of the preferred CPUs are busy, and other CPUs are idle, then soft affinity allows the system to run the jobs on the idle CPUs - even though the idle CPUs aren't in the preferred affinity mask.

Another aspect of quad core systems like MacPro or future Kentfields: On these systems, two cores share one 4 MB cache. If an application runs on two threads, it can run on two cores on the same chip, or on two cores on different chips. Threads that run on the same chip can exchange data very quickly, because anything that is in one threads L2 cache is automatically in the other threads L2 cache, but both threads together have only 4 MB cache. Threads running on different chips cannot exchange data quickly; data that is exchanged needs to be transferred through main memory. However, _each_ chip has 4 MB cache, or 8 MB total.

In other words, some applications will run faster if using threads on the same chip, some will run faster if using threads on separate chip. It is quite hard for the OS to guess, but the application developer should have some idea.

AidenShaw

Sep 11, 2006, 06:43 PM

Aiden, it's just not like you to make a statement like this without adding the links...
http://www.spec.org/cpu2000/results/cfp2000.html ;)

Note the Dell Precision Workstation 390 (Conroe) and the Precision Workstation 690 (Xeon 5100).

3 GHz Xeon - 2775
2.93 GHz Conroe Extreme - 2872

That "horrible buffered memory" is about a 3.5% handicap....at most. (The memory, chipsets, motherboards are all different.)

And one shouldn't say "but the FB-DIMMs are clocked faster" - the buffering is what enables the faster clock, as well as what adds the latency. The two tend to balance out, and the net result is that you can put 64 GiB of RAM on the Xeon - which you couldn't do without buffering!