8-Core Mac Pro (Clovertown) [Archive]

MacRumors

Sep 13, 2006, 07:49 AM

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

AnandTech took (http://anandtech.com/mac/showdoc.aspx?i=2832&p=6) a Mac Pro, which comes with two Dual-Core Xeon (Woodcrest) processors and replaced them with samples of two Quad-Core Xeon (Clovertown) processors.

We grabbed a pair of 2.4GHz Clovertown samples and tossed them in the system, and to our pleasure, they worked just fine. Our samples used a 1066MHz FSB, although we're expecting the final chip to use a 1333MHz FSB, but the most important part of the test is that all 8 cores were detected and functional.

The Mac Pro appeared to run fine with the Quad-Core processors, effectively givem them a 8-Core machine. While they are unable to give performance numbers due to non-disclosure agreements, the machine was reportedly stable. It also gives hope for current Mac Pro owners that they will be able to later upgrade the processors on their machine in the future. Clovertown Quad-core processors are expected to be available in late 2006.

Patch^

Sep 13, 2006, 07:52 AM

cool!! They should hopefully increase speed :)

I like the fact that you can upgrade the processors now, but Xeons are pretty expensive.

Alexsaru

Sep 13, 2006, 07:54 AM

MIDI_EVIL

Sep 13, 2006, 08:04 AM

Is there a possibility to assign single cores, or even pairs of cores to certain jobs?

For example, have two cores rendering, two cores encoding, two cores processing regular actions, and two cores making breakfast?

Does my PowerBook G4 have 8 Cores? What is an 8-Core?

Rich.

milo

Sep 13, 2006, 08:05 AM

A bit pointless given that no software utilises the extra cores yet.

Not true, according to the article. They said it wasn't easy, but they were able to max out all 8 cores. You can see the Activity Monitor graph all filled up.

It would be nice if 10.5 would allow a more 'blind' method to utilize these cores, versus having programmers specificly program for multi-core. Now that would be extremely helpful and allow a more simultanous workflow.

That's how it is now, at least with multiple apps. I bet it's possible to program for an unspecified number of multiple cores, and there may be apps doing it already.

I was interested to see that they were unable to max out CPU utilization on all 8 cores in the system. I hope it's due to the software these days not being ready to fully utilize more than one or two cores and not due to OSX's ability to scale to larger core counts. Since that's obviously where we're heading. Does anyone know about the potential for scalability of OSX to large numbers of CPU's/cores? I know some *nix varieties and BSD varieties do this really well, but one wonders if they were thinking this far in the future when they developed OSX. It'll be interesting to see...

Read the article again, they WERE able to max them out, just not easily. Based on that, OSX seems to be able to scale already. Developers just need to start writing apps that are more MP friendly.

prady16

Sep 13, 2006, 08:11 AM

thats a killer system!
Kinda like supercomputers for small businesses!

dornoforpyros

Sep 13, 2006, 08:13 AM

savar

Sep 13, 2006, 08:14 AM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

I'm still getting used to having two cores in my laptop!

Man, I don't know why people keep saying this. On OS X, *all software utilizes the extra cores*. The only way it wouldn't is if you have less than 8 processes running, which I guarantee you that you don't. (System alone requires 20-30 processes to run.)

Granted, 8 cores won't make Mail open up faster, but there are still plenty of ways to use those cores, and that's only going to increase as apps are re-written to be more heavily multi-threaded.

iGary

Sep 13, 2006, 08:14 AM

savar

Sep 13, 2006, 08:17 AM

I was interested to see that they were unable to max out CPU utilization on all 8 cores in the system. I hope it's due to the software these days not being ready to fully utilize more than one or two cores and not due to OSX's ability to scale to larger core counts. Since that's obviously where we're heading. Does anyone know about the potential for scalability of OSX to large numbers of CPU's/cores? I know some *nix varieties and BSD varieties do this really well, but one wonders if they were thinking this far in the future when they developed OSX. It'll be interesting to see...

Older versions of OS X had severe limitations due to kernel re-entrancy...or lack thereof. There were only two locks for the entire kernel (also known as "funnels")...but Apple has revised the kernel for 10.5 and will be implementing much more granular locks, which should alleviate the re-entrancy problem.

whooleytoo

Sep 13, 2006, 08:22 AM

Man, I don't know why people keep saying this. On OS X, *all software utilizes the extra cores*. The only way it wouldn't is if you have less than 8 processes running, which I guarantee you that you don't. (System alone requires 20-30 processes to run.)

Actually, it's even less than 8 - as a process can have several threads each of which can be moved to idle processors. Safari alone on my Mac currently has 23 threads at the moment, my system overall 277.

Platform

Sep 13, 2006, 08:25 AM

WOW...I have one core....now 8 cores...:eek:

Brandon Sharitt

Sep 13, 2006, 08:25 AM

While Clovertown and Kenstfield ar interesting as probably going to be the first consumer quad core CPUs, they are currently much like Intels initial stabs at Dual Core, which were largely two Pentium chips jammed together. What will be interesting to see is the second generation Intel quad cores and (probably) first generation AMD quad core CPUs, which should be to Kenstfield and Clovertown as the Core Duo family is to the dual core Pentium 4s, though maybe not quite the same performance jump.

hob

Sep 13, 2006, 08:31 AM

That's a joke! Incredible. Shame they couldn't say much about the performance though...

For the performance of Clovertown you'll have to wait a bit longer as we're not allowed to disclose it just yet

whooleytoo

Sep 13, 2006, 08:37 AM

Evangelion

Sep 13, 2006, 08:42 AM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

Most people run more than one app at once.

iMikeT

Sep 13, 2006, 08:55 AM

I'll just wait for the 16 core model.:p

Shagrat

Sep 13, 2006, 08:56 AM

and this got negative votes because...??????????

Yeesh!

epitaphic

Sep 13, 2006, 08:59 AM

What I couldn't understand - I couldn't see it explained in the article - why is the dual core Mac Pro (i.e. with current Mac Pro with 2 cores disabled) faster in so many tests than the 4 core Mac Pro.

probably due to latency involved in distributing the load across the two processors. that's the same problem a single Clovertown would have. Only true quads wouldn't suffer from these problems (earliest seems to be Harpertown in 2009, don't know if there are any non-MCM Xeons scheduled before then)

bob5820

Sep 13, 2006, 09:13 AM

cool!! They should hopefully increase speed :)

I like the fact that you can upgrade the processors now, but Xeons are pretty expensive. I believe the article stated that quad core is not likely to reach dual core speeds. Some situations will get better performance from faster dual core CPUs while others may get better performance from slower quad core CPUs.

ezekielrage_99

Sep 13, 2006, 09:14 AM

Do you really need that amount of power? I'm guessing yes ;)

I'd love to see the performance on the Clovertown Mac Pro.

Eidorian

Sep 13, 2006, 09:17 AM

DAMN :eek:

so 2-3 years from now are people going to be asking "why am I stuck with dual core in my macbook?" oh yeah I'll mostly be surfing the web and maybe editing a photo once and a while" :rolleyes:Fixed

RichP

Sep 13, 2006, 09:19 AM

Nice! Im with iGary and others, soon as they are out, Im buying. That should correspond nicely with the release of CS3.

Im doing work in Alias these days, I can only imagine how 8 cores could do a render!

Suprised the MacPro could handle the heat of 8 cores with its 4-core heatsink design. I read the article about Kentsfield on Tom's Hardware, and that chip made ALOT of heat; these are basically the same chip.

EDIT: Who knows, the hot setup may be a refurb'ed MacPro 2.0Ghz, then drop in better CPUs!

RedTomato

Sep 13, 2006, 09:49 AM

Who knows, the hot setup may be a refurb'ed MacPro 2.0Ghz, then drop in better CPUs!

Damn, yes, that would be the ultimate future hot machine - a second hand or refurb Mac Pro 2ghz, - I reckon by October they will about £1200 second hand, then in the new year, down to about £800, get one, then pop in 2 x 4 cores.

Anyone know if the chipset will be pin-compatible with the next generation quad-cores?

Seems we are stuck as far as increasing MHZ goes, 3ghz seems to be near the realistic limit. Now we're just adding more and more cores in. How many cores will be the feasible limit?

After that, what will be the next method of radically increasing computing throughput?

macenforcer

Sep 13, 2006, 09:52 AM

Now this is what I am talking about. YEAH!

guzhogi

Sep 13, 2006, 09:53 AM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

I'm still getting used to having two cores in my laptop!
This is a bit of a chicken and the egg problem. Hardware companies don't want to release multicore hardware b/c no software is out to support it and software people don't want to ship multiprocessor software b/c no multiproc hardware is out. Since there are a few multiproc computers out now, some software companies might start multithreading their apps, most people probably don't have it them yet.

Also, some apps won't really benefit from being multithreaded. Take a basic calculator. Why would you really need it to take advantage of multiproc computers? It isn't that processor heavy to do 2+2. AV software, like iTunes, Final Cut Pro, etc., could greatly benefit from multiproc systems. One core can do the audio while the other does the video for FC or iTunes could use one core for playing music and the other for ripping.

Something I'd like to see is to have AV stuff offloaded to the sound or graphics card to speed it up. I've heard of a company called Aspex Semiconductors (www.aspex-semi.com) that designs PCI cards that speed up MPEG encoding. Might be nice for video pros.

rawdigits

Sep 13, 2006, 09:57 AM

I guess I'll wait until Tigerton. I want to buy a merom MBP when it comes out.
The architecture of Tigerton is without the Frontside Bus. More in direction of AMD. Much more efficiency than put just 8 Cores to the 1.3 FSB. Clovertown alos has slower RAM.

:rolleyes:

ImAlwaysRight

Sep 13, 2006, 10:01 AM

Great news. I can't afford nor do I need that much power, but great to see it can be done.

zero2dash

Sep 13, 2006, 10:10 AM

DStaal

Sep 13, 2006, 10:12 AM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

Mac OS X distributes threads and processes across cores/CPUs to optimize performance already. (Subject to some limitations, as noted already.)

Many Mac programs which can benifit from mutiple threads already use this, and will automatically get boosts from 8 cores depending on the amount of cocurrency they support.

On the other hand, not everything is suitable for cocurrent execution. Photoshop editing an image would love to have a core per pixel. BBEdit couldn't care less, most likely. It all depends on what you are doing.

Plenty of Mac software would use the extra cores, if they were avalible.

(Note: I keep specifying 'Mac' here. There is a reason. Windows isn't as good at multithreading/processing yet...)

WildPalms

Sep 13, 2006, 10:12 AM

*sigh* My poor, poor wallet.....may as well call it iWallet for the use it gets buying Apple gear constantly....:o

Platform

Sep 13, 2006, 10:13 AM

Most people run more than one app at once.

Most are multi-threaded though and if I am not incorrect it doesn't matter for Photoshop if there are two or 72 cores...;)

RHutch

Sep 13, 2006, 10:21 AM

and this got negative votes because...??????????

Yeesh!

I was wondering the same thing. How can people not be pleased with this? The fact that the processors can be swapped, the fact that the OS recognizes all 8 cores, and that it was so difficult to use all of the processing power. What is there to complain about?

mmmcheese

Sep 13, 2006, 10:23 AM

aricher

Sep 13, 2006, 10:31 AM

RichP

Sep 13, 2006, 10:33 AM

After that, what will be the next method of radically increasing computing throughput?

Personally, I still see data transfer, namely from storage media, as a huge bottleneck in performance. Unless you are doing something really CPU intensive (vid editing, rendering, others) Most of the average "wait-time" is the damn hard drive.

DStaal

Sep 13, 2006, 10:35 AM

It would be nice if 10.5 would allow a more 'blind' method to utilize these cores, versus having programmers specificly program for multi-core. Now that would be extremely helpful and allow a more simultanous workflow.

How much more 'blind' do you want it? All the programmer has to do at this point is use multiple threads. Even if they don't, multiple cores will be automatically used for system and other processes.

Splitting one thread so that it ran cocurent with itself is a recipie for massive trouble. Mac OS X is about as blind as any system out there for the programmer. There may be some more optimizations that the system could make in it's own handling of multiprocessing, but from a programmer's perspective it doesn't matter how many cores the system has. (Unless you really want it to.)

Mundy

Sep 13, 2006, 10:38 AM

Are these processors 32 or 64 bit? I told one of my PC-lovin' IT guys about the 8 core Mac this morning and he said, "32 bit processors are ancient technology no matter how many you stuff into a box, but I guess they are OK for entertainment computers." :rolleyes:

Clovertown is a 64-bit CPU.

Ask your PC-loving IT guy if he uses Windows XP64 and more than 4 gigabytes of RAM. If not, then 32-bit processors are apparently okay for him, too.

sososowhat

Sep 13, 2006, 10:50 AM

One could run a Folding@Home process on each core :D

ncook06

Sep 13, 2006, 10:50 AM

I'm just wondering if I can drop one of these into an iMac... Are they pin-compatible? Also sort of wondering about a heat issue.

aricher

Sep 13, 2006, 10:56 AM

Clovertown is a 64-bit CPU.

Ask your PC-loving IT guy if he uses Windows XP64 and more than 4 gigabytes of RAM. If not, then 32-bit processors are apparently okay for him, too.

hee hee - thanks.

DStaal

Sep 13, 2006, 11:03 AM

I'm just wondering if I can drop one of these into an iMac... Are they pin-compatible? Also sort of wondering about a heat issue.
Nope - Different processor archetecture. Even if the pins were the same, the motherboards are different.

Both Clovertown and Woodcrest are 'Xeon' chips, which is a particular processor family. Chips in one processor family may be replaceable with others from that family, but generally different families are not replacable with each other. (Unless specifially designed to do so.)

Also, the iMac is a 32-bit computer, and these are 64-bit chips, reducing any possiblity to zero.

brianus

Sep 13, 2006, 11:05 AM

probably due to latency involved in distributing the load across the two processors. that's the same problem a single Clovertown would have. Only true quads wouldn't suffer from these problems (earliest seems to be Harpertown in 2009, don't know if there are any non-MCM Xeons scheduled before then)

What about Tigerton (2007)? Isn't that a "true" quad?

AoWolf

Sep 13, 2006, 11:05 AM

I remember thinking "My goodness what would I ever do with all this power" when playing with the apple store quad woodcrests. The future is very bright :-)

RedTomato

Sep 13, 2006, 11:11 AM

Personally, I still see data transfer, namely from storage media, as a huge bottleneck in performance. Unless you are doing something really CPU intensive (vid editing, rendering, others) Most of the average "wait-time" is the damn hard drive.

Arrays of cheap RAM on a PCIe card?

The RAM companies don't seem interested in making wodges of slow cheap hi-cap ram, only in bumping up the speed and upping the capacity. For the last 10 years, a stick of decent RAM has always been about £100/ $100 no matter what the capacity / flavour of the moment is.

Even slow RAM is still orders of magnitude faster than a HD, hence my point. There's various historical and technical factors as to why we have the current situation.

I've also looked at RAID implementations (I run a RAID5) but each RAID level has its own problems.

I've recently seen that single-user RAID3 might be one way forward for the desktop, but don't really know enough about it yet.

ChrisA

Sep 13, 2006, 11:14 AM

Very cool. Now to find apps (os10.5 direct blind support?) that can make use of all those cores. :cool:

One app would be iTunes. I noticed iTunes was running 14 threads last night. Any time you have a multithreaded application or are running multiple single thread aplications more cores can help.

Some server applications (the Apache web server and many DBMS systems) use a "process per client" model where a new process (another instance) of the server is created for each client connection. A bussy web server might have 100 copies of apache all running at once. 8 cores would help there.

Half Glass

Sep 13, 2006, 11:26 AM

Wow...a user upgradable Mac. Good stuff indeed.

I am anxiously awaiting better utilization of all the cores, but the ability to multitask without hiccups is still great for now!

--HG

gallinger

Sep 13, 2006, 11:27 AM

does anyone know how much the clovertown chips are going to be?

mdntcallr

Sep 13, 2006, 11:30 AM

this is pretty neat news.

means people like me can buy a mac pro tower with the 2.0 ghz core. good video card.

then upgrade later on when i have more money. that and it will be powerful as hell.
super nice!

epitaphic

Sep 13, 2006, 11:53 AM

What about Tigerton (2007)? Isn't that a "true" quad?

Intel has two lines of Xeon processors:

* The 5000 series is DP (dual processor, like Woodcrest, Clovertown)
* The 7000 series MP (multi processor - eg 4+ processors)

Tigerton is supposed to be an MP version of Clovertown. Meaning, you can have as many chips as the motherboard supports, and just like Clovertown its an MCM (two processors in one package). 7000's are also about 5-10x the price of 5000's.

So unless the specs for Tigerton severely change, no point even considering it on a Mac Pro (high end xserve is plausible).

ChrisA

Sep 13, 2006, 11:54 AM

Arrays of cheap RAM on a PCIe card?

The RAM companies don't seem interested in making wodges of slow cheap hi-cap ram, only in bumping up the speed and upping the capacity. For the last 10 years, a stick of decent RAM has always been about £100/ $100 no matter what the capacity / flavour of the moment is.

Even slow RAM is still orders of magnitude faster than a HD, hence my point. There's various historical and technical factors as to why we have the current situation.

I've also looked at RAID implementations (I run a RAID5) but each RAID level has its own problems.

I've recently seen that single-user RAID3 might be one way forward for the desktop, but don't really know enough about it yet.

The reason for the RAM improvoments in speed and size are that RAM (not CPU) is the main bottle neck in preformance. A CPU can only execute instructions as fast as they can be pulled out of RAM. Now you go and put multiple cores inthe box and the demand on RAM doubles.

As for RAID. I think the way forward is Sun's "ZFS" file system. There was talk of that moving into Mac OSX and we know it is being ported to BSD Unix and Linux. Basically ZFS makes the RAID layer just go away

Read more here...
http://www.sun.com/2004-0914/feature/index.html

Sun has released this as Open Source. so it will get ported around to other OSes. I hear Sun's Dtrace is already in Leopard

moogs

Sep 13, 2006, 11:57 AM

Would it be smart to wait for these 8 core mac pros or are they still a long ways away?

epitaphic

Sep 13, 2006, 12:02 PM

Read more here...
http://www.sun.com/2004-0914/feature/index.html

Logically, the next question is if ZFS' 128 bits is enough. According to Bonwick, it has to be. "Populating 128-bit file systems would exceed the quantum limits of earth-based storage. You couldn't fill a 128-bit storage pool without boiling the oceans."
wow. boiling the oceans. there's a thought that never crossed my mind ;)

RedTomato

Sep 13, 2006, 12:04 PM

Quoting myself, bad boy,

Arrays of cheap RAM on a PCIe card?

http://www.superssd.com/products/tera-ramsan/indexb.htm

That's one answer. 1 TB of DDR on a (rather big) card. Takes 2500 watts to power, but gives you 32GB/sec continous bandwidth.

Would that be enough to feed an 8-core Mac Pro? (4GB/sec per core, running through the entire 1TB in 32 seconds.... hmmm)

Wonder when products like that will filter down?

There's a rather sad Gigabye Ramdisk card at

http://www.gigabyte.com.tw/Products/Storage/Products_Overview.aspx?ProductID=2180&ProductName=GC-RAMDISK

Costs only £100 but has a max capacity of 4GB. You'd be better off spending the money on more system RAM.

Dr.Gargoyle

Sep 13, 2006, 12:05 PM

and this got negative votes because...??????????
All the people that just coughed up $3k for a quad core MacPro.

Mr. Anderson

Sep 13, 2006, 12:05 PM

meaning that unless you have a way of really stressing 8 cores, you may be better off with 4 faster cores in your Mac Pro

drool - i'll take 8 cores for my 3D rendering :D

I think I'll be selling my quad G5 next year for a 8 core Mac Pro.

D

daver969

Sep 13, 2006, 12:05 PM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

I'm still getting used to having two cores in my laptop!

What I couldn't understand - I couldn't see it explained in the article - why is the dual core Mac Pro (i.e. with current Mac Pro with 2 cores disabled) faster in so many tests than the 4 core Mac Pro.

I think part of the reason so many people seem to be hung up on the "software doesn't utilize multiple cores" mantra is because benchmarks tend to test only one software component at a time. If a given app isn't multithreaded, then it doesn't benefit from multiple cores in these tests. But that doesn't mean that multiple cores don't affect the overall system speed.

What we need is some kind of a super benchmark: How fast is my computer when I'm watching a quicktime stream of Steve demoing the latest insanely great stuff, while ripping my CD collection to iTunes, while surfing complex Cnet.com pages (w/animation), and compiling the latest version of my Java app, every once in a while flipping over to Dashboard (dashboard seems to take up a lot of system resources every time I invoke it, not just on startup).

At this point I would rather push towards more cores than more raw speed in a single core, since I don't tend to wait on any single process. If something is taking a long time, like loading a page or compiling code, I switch to something else and come back later. I would much rather have the whole system retain its responsive feel than have one app finish its task a few seconds quicker.

DStaal

Sep 13, 2006, 12:12 PM

Sun has released this as Open Source. so it will get ported around to other OSes. I hear Sun's Dtrace is already in Leopard

Great. Um... What's their patent licensing scheme on this? (Since they proudly announce they've patented parts of it...)

relimw

Sep 13, 2006, 12:13 PM

Do you mean like how BeOS did things?
Sorry, never played with BeOS so I don't know how they did it :)

dante@sisna.com

Sep 13, 2006, 12:22 PM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

I'm still getting used to having two cores in my laptop!

Not pointless at all if a person uses a lot of applications. You can justify all 8 cores right now. For sure. My quad core shines in multitasking.

Evangelion

Sep 13, 2006, 12:30 PM

shawnce

Sep 13, 2006, 12:36 PM

Do you mean like how BeOS did things?

Yeah BeOS had this great feature called magic pixel dust. :rolleyes:

All that BeOS had was separate threads per window at the UI level. This does nothing for parallelizing compute tasks. These extra thread that BeOS had spent most of their time doing absolutely nothing.

What Mac OS X has now is several operating services that will automatically scale up to use as many cores possible (while still making sense). Many of the "Core" framework do this without any work by application authors other then then those authors deciding to use those services instead of rolling their own.

For example ColorSync color correction, audio conversion, audio mixing, etc.

...and yes Mac OS X 10.5 is expanding the OS services that will do the right thing (TM) as well as making it easier for developers to transparently and directly utilize the cores available in a system.

Multimedia

Sep 13, 2006, 12:37 PM

I wouldn't want to say I told you so but... :eek: :p :D Where's Multimedia? This is exciting!
Wow...a user upgradable Mac. Good stuff indeed.

I am anxiously awaiting better utilization of all the cores, but the ability to multitask without hiccups is still great for now!Must Crush Video...Must Crush Video...Must Crush Video...Must Crush Video...Must Crush Video...Must Crush Video...Must Crush Video...

I'm still gonna wait for the Clovertown option to appear in the BTO page, then price retail Clovertowns a Fry's before I decide if I'll let Apple to my upgrade or do it myself according to which way cost less. But I really don't want to kill my warranty on day one. So it'll be academic since they are going retail in a month prolly before Apple adds the Clovertown option to the BTO page although they were pretty Johnny On The Spot with the C2D iMacs.

aegisdesign

Sep 13, 2006, 12:40 PM

Most people run more than one app at once.

Yes, that's true.

It's also true that most of the time, most people aren't even maxing out ONE core never mind eight.

And when they do, their program won't get any faster unless it's multithreaded and able to run on multiple cores at once.

radesousa

Sep 13, 2006, 12:40 PM

So the question I have is can the latest iMac be CPU upgraded like the MacPro?

scottlinux

Sep 13, 2006, 12:41 PM

Blender http://www.blender.org/ can uses 8 cores.

Multimedia

Sep 13, 2006, 12:44 PM

shawnce

Sep 13, 2006, 12:48 PM

Yes, that's true.

It's also true that most of the time, most people aren't even maxing out ONE core never mind eight.

And when they do, their program won't get any faster unless it's multithreaded and able to run on multiple cores at once.

Lets not forget things like Spotlight that can now run more rigorously without affecting CPU resource much. You will get more intelligent software that can prepare for what you want to do so that when you go to do it it will be much more responsive. In other words just because some tasks cannot be easily broken up to leverage multiple cores doesn't mean that tasks such as those cannot be speculative run by software on idle cores in preparation for you doing the task.

aegisdesign

Sep 13, 2006, 12:54 PM

All that BeOS had was separate threads per window at the UI level. This does nothing for parallelizing compute tasks. These extra thread that BeOS had spent most of their time doing absolutely nothing.

Whilst true in that regard, BeOS also had threads for event queues too if you used BLooper, which could also be overused.

I think the threaded-ness just gave everyone the impression it was fast and not waiting on anything to a large extent rather than it actually being fast. Most of the speed just came from it being very lightweight and the apps written for it being written by good programers that knew how the thread.

Antares

Sep 13, 2006, 12:55 PM

and this got negative votes because...??????????

Yeesh!

One reason this is negative is because there were no performance numbers given.

aegisdesign

Sep 13, 2006, 12:55 PM

Lets not forget things like Spotlight that can now run more rigorously without affecting CPU resource much. You will get more intelligent software that can prepare for what you want to do so that when you go to do it it will be much more responsive. In other words just because some tasks cannot be easily broken up to leverage multiple cores doesn't mean that tasks such as those cannot be speculative run by software on idle cores in preparation for you doing the task.

Yes, that's definitely true. And I'd be happy to divert a whole core just to frickin WindowServer. :D

jrhone

Sep 13, 2006, 12:58 PM

Man, I don't know why people keep saying this. On OS X, *all software utilizes the extra cores*. The only way it wouldn't is if you have less than 8 processes running, which I guarantee you that you don't. (System alone requires 20-30 processes to run.)

Granted, 8 cores won't make Mail open up faster, but there are still plenty of ways to use those cores, and that's only going to increase as apps are re-written to be more heavily multi-threaded.

NOT TRUE....The Quad core G5 people are in an uproar because Logic Pro only uses 2 cores on the G5....they updated Logic Pro so it uses 4 cores, but the G5 Quad still only uses 2 cores....there are also photoshop actions that are NOT multi core aware so will only run on one core.....Hopefully 10.5 will make all this irrelevant.

shawnce

Sep 13, 2006, 01:04 PM

Whilst true in that regard, BeOS also had threads for event queues too if you used BLooper, which could also be overused.

Mac OS X has runloops which are flexible event processing constructs that can be run per thread. So nothing really unique in regards to BeOS in that regard.

aegisdesign

Sep 13, 2006, 01:05 PM

Also, the iMac is a 32-bit computer, and these are 64-bit chips, reducing any possiblity to zero.

Only the Yonah based Core Duo iMacs are 32bit (Well, and the G3/G4 too). G5 and the new iMac Core 2 Duo models on sale now are 64bit. Not that it matters per se.

shelterpaw

Sep 13, 2006, 01:05 PM

Yes, that's true.

It's also true that most of the time, most people aren't even maxing out ONE core never mind eight.

And when they do, their program won't get any faster unless it's multithreaded and able to run on multiple cores at once.
The Mac Pro isn't for most people. It's for professionals and professional applications, which are usally multithreaded, and will take advantage of the capabilities.

If you have a complaint about all these cores and not being able to take advantage of them, then this is not the computer for you. You're probably not using the software that will take advantage of them, so let it go and stop whining about it. For the those of us that do, this is great news.

daver969

Sep 13, 2006, 01:10 PM

Yes, that's true.

It's also true that most of the time, most people aren't even maxing out ONE core never mind eight.

And when they do, their program won't get any faster unless it's multithreaded and able to run on multiple cores at once.

I'm underutilizing my cpu nearly all of the time, but that's irrelevant-what really matters to me is that fraction of the time when I *am* asking it to do 4 things at once, and I want it do them at the same speed that each could be done individually.

zero2dash

Sep 13, 2006, 01:11 PM

How is this Apple "innovating"? Anandtech just put pre-release quad-core Intel-processor in to an Apple-computer. Apple itself had nothing to do with it. They could have used quad-core Dell-machine just as well.

The OS takes advantage of the extra 4 cores already therefore its ahead of the technology curve, correct? Gee, no innovation here...please move along folks. :rolleyes:

As for using a Dell, sure they could've used that. Would Windows use the extra 4 cores? Highly doubtful. Microsoft has sketchy 64 bit support let alone dual core support; I'm not saying "impossible" but I haven't read jack squat about any version of Windows working well with quad cores. You think those fools (the same idiots who came up with Genuine Advantage) actually optimized their OS to run in an 8 core setup? Please pass along what you're smoking. :rolleyes:

epitaphic

Sep 13, 2006, 01:14 PM

I'd be happy to divert a whole core just to frickin WindowServer. :D
going out on a limb here and assuming you have a heavily cluttered desktop

Multimedia

Sep 13, 2006, 01:15 PM

So the question I have is can the latest iMac be CPU upgraded like the MacPro?Only to 2.33GHz Merom C2D which is the forseeable top speed. :(

aricher

Sep 13, 2006, 01:17 PM

Clovertown is a 64-bit CPU.

Ask your PC-loving IT guy if he uses Windows XP64 and more than 4 gigabytes of RAM. If not, then 32-bit processors are apparently okay for him, too.

This was his response:

"Cloverton is not 64, Cloverton MP (Tigerton) is 64 and is still on the drawing board last I heard.

Using Windows 2003 x64 with dual AMD Opterons and 32 GB RAM. It'll smoke any of you antique Mac server wannabees..."

Multimedia

Sep 13, 2006, 01:21 PM

The Mac Pro isn't for most people. It's for professionals and professional applications, which are usally multithreaded, and will take advantage of the capabilities.

If you have a complaint about all these cores and not being able to take advantage of them, then this is not the computer for you. You're probably not using the software that will take advantage of them, so let it go and stop whining about it. For the those of us that do, this is great news.Thank you shelterpaw. And Bravo! Couldn't have said it better. Those who don't see the point of a lot of cores are not doing anything like what those of us who do are. :)I'm underutilizing my cpu nearly all of the time, but that's irrelevant-what really matters to me is that fraction of the time when I *am* asking it to do 4 things at once, and I want it do them at the same speed that each could be done individually.Zactly. This is the most reason - not that you need this level of performance 24/7. Thank you for that daver.http://images.amazon.com/images/P/B0007US79Y.01._AA240_SCLZZZZZZZ_.jpg
MORE POWER!

Multimedia

Sep 13, 2006, 01:23 PM

This was his response: "Cloverton is not 64, Cloverton MP (Tigerton) is 64 and is still on the drawing board last I heard.

Using Windows 2003 x64 with dual AMD Opterons and 32 GB RAM. It'll smoke any of you antique Mac server wannabees..."Your IT guy can't even spell it correctly - so how do you expect him to have a clue?

ClovertownHow do you know the IT dude typed that vs. the poster just typing what he said?I did a direct copy-paste from my IT guy's email. What a knucklehead - him not you.Do you think he has a clue telling him/her that Clovertown is only a 32 bit processor? He's supposedly a professional IT guy. I think it's pretty outrageous he doesn't know better.Anyone seen this?

http://images.dailytech.com/nimage/1775_large_longtermroadmap.png

The real architecture changes are coming June 2008, then June 2010, then June 2012. With derivatives in the years between.

So Merom(Merom Santa Rosa)/Conroe/Woodcrest(Clovertown) are the end of the road of separate chips. No more mobile/desktop/sever chip... all are the same (should expect mobiles to have the lowest MHz, then desktop, then toping out with server)

And what's interesting is that each architecture change will be a leap in performance similar to Pentium D to Conroe transition. (source) (http://www.dailytech.com/article.aspx?newsid=2649)

Screw Tigerton, Penryn's next (probably June 2007)Thanks. Very helpful.

NewbieNerd

Sep 13, 2006, 01:27 PM

Your IT guy can't even spell it correctly - so how do you expect him to have a clue?

Clovertown

I think we can all read at normal size. Besides, how do you know the IT dude typed that vs. the poster just typing what he said?

aegisdesign

Sep 13, 2006, 01:30 PM

The Mac Pro isn't for most people. It's for professionals and professional applications, which are usally multithreaded, and will take advantage of the capabilities.

If you have a complaint about all these cores and not being able to take advantage of them, then this is not the computer for you. You're probably not using the software that will take advantage of them, so let it go and stop whining about it. For the those of us that do, this is great news.

It was just a general point, not a whine, so don't get your panties in a bunch. And some of the applications that don't take advantage of multiple cores currently are Adobe Photoshop and Quicktime which both rarely use more than two cores and sometimes only one. Both pretty important to professionals.

puuukeey

Sep 13, 2006, 01:32 PM

http://images.amazon.com/images/P/B0007US79Y.01._AA240_SCLZZZZZZZ_.jpg
MORE POWER!

kingtj

Sep 13, 2006, 01:33 PM

He's totally mistaken! The Cloverton CPUs will *all* be 64-bits, as Woodcrest (found in current Mac Pros) is. Intel is not going to ever go back to a 32-bit Xeon class CPU.

The difference between Woodcrest and "Tigerton" is that Woodcrest CPUs achieve their "dual core" status by basically placing two complete Xeon CPUs under one outer casing, and making them communicate with each other through the front-side bus on the motherboard.

Cloverton will be the same way, but with 4 cores packed into one casing, instead of just two.

"Tigerton" will finally allow both cores to interconnect with each other through an internal interface built into the CPU, instead of slowing communications down by routing it off one CPU core, through the motherboard's front-side bus, and back onto the other core.

This was his response:

"Cloverton is not 64, Cloverton MP (Tigerton) is 64 and is still on the drawing board last I heard.

Using Windows 2003 x64 with dual AMD Opterons and 32 GB RAM. It'll smoke any of you antique Mac server wannabees..."

aegisdesign

Sep 13, 2006, 01:35 PM

going out on a limb here and assuming you have a heavily cluttered desktop

Yes, I know. it takes me a couple of days to really clutter up my desktop whilst I'm working on a project and the desktop's the handiest place to stash stuff. I've also usually got 30-40 windows open too.

If I'm busy I don't have the time to de-clutter and get back teh snappy.

RedTomato

Sep 13, 2006, 01:36 PM

I read the link above about the ZFS filesystem.

Hmm this could remove a lot of the pain I currently have juggling disks on the cheap.

(I hold a lot of footage of deaf people signing for a project, and don't really have any budget to pay for disk storage. I currently have about 200 GB left on a 1 TB RAID5 system inside a Powermac G3)

It seems the concept of individual volumes will vanish, and instead ZFS creates a common pool of filespace and looks after the checksums etc itself. New drives can just be thrown into the array and ZFS will look after optimising the array I/O.

Mixing 15k rpm speed demon drives with 5400rpm storage hog drives mmmm...

I look forwards to being able to buy a cheap chassis with just a power unit and space for 10 drives, and being able to put that next to my G3, and having ZFS sort out what to do with the 8-9 drives in there.

Something like that hooked up to a Cloverton should give significant HD speedup. Not as much as a ramdisk tho :)

One thing, the article says ZFS can cope with drives being removed from the pool. I'd like to see more detail on that. It surely copes with 1 out of 4 drives failing - what about 3 out of 4? What if 3 x 20GB 15k rpm drives fail and the 1x750GB 5400rpm drive is still up?

relimw

Sep 13, 2006, 01:36 PM

How much more 'blind' do you want it? All the programmer has to do at this point is use multiple threads. Even if they don't, multiple cores will be automatically used for system and other processes.

Splitting one thread so that it ran cocurent with itself is a recipie for massive trouble. Mac OS X is about as blind as any system out there for the programmer. There may be some more optimizations that the system could make in it's own handling of multiprocessing, but from a programmer's perspective it doesn't matter how many cores the system has. (Unless you really want it to.)

Programming in pthreads is a bear (at least to me) an easier method would be nice. However, when I was looking up something today I came across OpenMP (http://www.openmp.org/) which seems to greatly simply setting up threads and the like. I suppose I was just thinking of run-time parallelization.

whatever

Sep 13, 2006, 01:41 PM

All the people that just coughed up $3k for a quad core MacPro.
I'm one of those people who dropped $4K for a quad core MacPro and basically I'm happy that I did. It blows away everything else that is out there today and will be the top performing Mac until 2007. Apple will not be releasing an upgrade to the Mac Pro this year. No matter what anyone says.

Why you might ask, well they don't need to!

But what if the competition releases these super fast machines, won't Apple be left behind. No! What OS will these machines be running, Windows XP. One of the things that seperates Apple from everyone else is their OS. They have an OS which takes full advantage (important word is full) of the hardware. It's the big advantage that they have over Dell and HP, they create the software that runs on the computer.

So if I want to run Final Cut Pro as fast as possible on an optiomized machine, then I'll have to run it on a Mac. Alright, that's a bad example, but in a way it's not, because a lot of the people buying Mac Pros also live in Apple's Pro apps.

The next new computer we'll see from Apple anytime soon will be the MacBook Pro which will be redesigned (featuring the MacBook's keyboard), upgrades to the MacBook won't happen until January (however Apple may try to get them out in December).

Apple's goal is to have everything 64-Bit before Leopard is uncaged.

THX1139

Sep 13, 2006, 01:42 PM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

I'm still getting used to having two cores in my laptop!

Why do people think the computing world always revolves around them? Extra cores WILL be recognized by most 3d applications and will speed up rendering. There are many other applications for multiple core use that don't include web browseing or writing email to your grandma.

peace

aricher

Sep 13, 2006, 01:44 PM

I think we can all read at normal size. Besides, how do you know the IT dude typed that vs. the poster just typing what he said?

I did a direct copy-paste from my IT guy's email. What a knucklehead - him not you.

epitaphic

Sep 13, 2006, 01:47 PM

Anyone seen this?

http://images.dailytech.com/nimage/1775_large_longtermroadmap.png

The real architecture changes are coming June 2008, then June 2010, then June 2012. With derivatives in the years between.

So Merom(Merom Santa Rosa)/Conroe/Woodcrest(Clovertown) are the end of the road of separate chips. No more mobile/desktop/sever chip... all are the same (should expect mobiles to have the lowest MHz, then desktop, then toping out with server)

And what's interesting is that each architecture change will be a leap in performance similar to Pentium D to Conroe transition. (source) (http://www.dailytech.com/article.aspx?newsid=2649)

Screw Tigerton, Penryn's next (probably June 2007)

ergle2

Sep 13, 2006, 01:53 PM

Only the Yonah based Core Duo iMacs are 32bit (Well, and the G3/G4 too). G5 and the new iMac Core 2 Duo models on sale now are 64bit. Not that it matters per se.

The Chipset in the Yonah iMac is still 945, which works fine with Merom's long mode (64bit/EM64T more).

faustfire

Sep 13, 2006, 01:54 PM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

A lot of 3d programs will use as many cores as are available when rendering.

And I would say that the next versions of many programs will be better suited for multiple core processors.* They are way too common for software developers to ignore them any longer.

adamfilip

Sep 13, 2006, 01:57 PM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

I'm still getting used to having two cores in my laptop!

No software such as, Cinema 4D, Motion, Aperture, Final Cut Pro etc

craig jones

Sep 13, 2006, 01:58 PM

Arrays of cheap RAM on a PCIe card?

The RAM companies don't seem interested in making wodges of slow cheap hi-cap ram, only in bumping up the speed and upping the capacity. For the last 10 years, a stick of decent RAM has always been about £100/ $100 no matter what the capacity / flavour of the moment is.

Even slow RAM is still orders of magnitude faster than a HD, hence my point. There's various historical and technical factors as to why we have the current situation.

I've also looked at RAID implementations (I run a RAID5) but each RAID level has its own problems.

I've recently seen that single-user RAID3 might be one way forward for the desktop, but don't really know enough about it yet.

Slow RAM may be faster than hard disk but it's too slow for main memory. It could be useful for disk cache but products like that came and went. If such hardware could actually result in performance improvements to justify their costs then you'd see products that used them.

As for RAID 3, it has been used before but really has no place considering modern disk drives and workloads. RAID 3 and 4, in order to work properly, require spindle sync. Workstations have no business implementing any parity-based RAID scheme. Servers used RAID 5 when they have high capacity needs and aren't sensitive to write performance.

relimw

Sep 13, 2006, 02:00 PM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.

Hehe, everybody else cited you, I suppose I will as well.

It's not that those cores won't be used. The average Joe user won't need them, it won't help you type letters any faster, and it'll do very little to help you websurf any faster (unless people keep putting bloat-ware browsers out there).
What it will help with, is people using HPC apps (BLAST comes to mind), or multi-threaded apps.

ergle2

Sep 13, 2006, 02:00 PM

So Merom(Merom Santa Rosa)/Conroe/Woodcrest(Clovertown) are the end of the road of separate chips. No more mobile/desktop/sever chip... all are the same (should expect mobiles to have the lowest MHz, then desktop, then toping out with server)

I think you've misunderstood.

Merom/Conroe/Woodcrest are one microarch now. That's Intel's point -- the core is essentially the same. Then they package as appropriate for a given market. Merom is lower-voltage/lower-clockspeed, Woodcrest has the external pins for multi-processor exposed and has a higher FSB, etc.

There will still be different chips for different markets, but the arch is the same across the board. This is a significant difference from the Pentium-4/Pentium-M days, where the arch was very, very different.

The other big difference is the support chipsets -- the Xeon range use a different chipset that supports FBDIMM vs DDR2 for the Core 2 branded chips. This is the reason Intel kept the memory controller off the CPU die, so that they had more flexibility with memory types.

mmmcheese

Sep 13, 2006, 02:02 PM

Dagless

Sep 13, 2006, 02:02 PM

Forget 3 monitors - 8 CORES. Lordy.

The move to intel was the best decision Apple made. Or just one of the very good ones.

Trekkie

Sep 13, 2006, 02:08 PM

does anyone know how much the clovertown chips are going to be?

if it follows typical intel transitions price point replace. So the same price as woodcrests. They might introduce faster ones though that cost more. We'll see before the end of the year.

Would it be smart to wait for these 8 core mac pros or are they still a long ways away?

Quad core is supposed to be out before EOY 2006.

Will Apple release it before then is the question...

Evangelion

Sep 13, 2006, 02:10 PM

The OS takes advantage of the extra 4 cores already therefore its ahead of the technology curve, correct? Gee, no innovation here...please move along folks. :rolleyes:

Uh, last time I checked, Windows can take advantage of multiple cores just fine. Do you think that multithreading is some Black Magic that only MacOS can do? Hell, standard Linux from kernel.org can use 512 cores as we speak!

Related to this: Maybe not 512-way SMP, but here (http://www.linux-mips.org/wiki/IP27_boot_messages) is what it looks like when Linux boots on 128-way SGI Origin supercomputer. Note, the kernel that is booting is 2.4.1, which was released in early 2001. Things have progressed A LOT since those day.

OS X works with quad core == "Ahead of technology curve"... puhleeze!

As for using a Dell, sure they could've used that. Would Windows use the extra 4 cores? Highly doubtful. Microsoft has sketchy 64 bit support let alone dual core support

Windows works just fine with dual-core. It really does. To Wndows, dual-core is more or less similar to typical SMP, and Windows has supported SMP since Windows NT!

I'm not saying "impossible" but I haven't read jack squat about any version of Windows working well with quad cores.

Any reason why it wouldn't work? And did you even read the Anandtech-article? They conducted their benchmarks in Windows XP! So it obviously DID work with four cores! And it DID show substantial improvement in performance in real-life apps! Sheesh! Dial tone that fanboysihness a bit, dude.

craig jones

Sep 13, 2006, 02:10 PM

The OS takes advantage of the extra 4 cores already therefore its ahead of the technology curve, correct? Gee, no innovation here...please move along folks. :rolleyes:

As for using a Dell, sure they could've used that. Would Windows use the extra 4 cores? Highly doubtful. Microsoft has sketchy 64 bit support let alone dual core support; I'm not saying "impossible" but I haven't read jack squat about any version of Windows working well with quad cores. You think those fools (the same idiots who came up with Genuine Advantage) actually optimized their OS to run in an 8 core setup? Please pass along what you're smoking. :rolleyes:

How do you know these things? Is Windows' 64-bit support sketchier than OS X's? Of course not. OS X has little 64-bit support and none at all for Intel. Windows also supports far more than 2 or 4 cores (although there are license restrictions). Windows has run on far more than 8 cores for a long, long time. You realize they have an actual presence in the server market, don't you?

janstett

Sep 13, 2006, 02:11 PM

Sheesh...just when I'm already high up enough on Apple for innovating, they throw even more leaps and bounds in there to put themselves even further ahead. I can't wait 'til my broke @$$ can finally get the money to buy a Mac and chuck all my Windows machines out the door.

I'm sure we'll see similar efforts from other PC manufacturers eventually, but let's see the software use those extra cores in Windows land. Ain't gonna happen...not on the level of what Apple's doing at least.

First, this is INTEL innovating, not Apple.

Second, Apple has been the one lagging behind on multiprocessor support. Pre OSX it was a joke of a hack to support multi CPUs in Mac OS and you had to have apps written to take advantage of it with special libraries.

On Windows, the scheduler automatically handles task scheduling no matter how many processors you have, 1 or 8. Your app doesn't have to "know" it's on a single or multiple processor system or do anything special to take advantage of multiple processors, other than threading -- which you can do on a single processor system anyway. Most applications are lazy and unimaginative, and do everything in a single thread (worse, the same thread that is processing event messages from the GUI, which is why apps lock up -- when they end up in a bad state they stop processing events from the OS and won't paint, resize, etc.). But when you take advantage of multithreading, there are some sand traps but it's a cool way to code and that's how you take advantage of multiple cores without having to know what kind of system you are on. I would assume OSX, being based on BSD, is similar, but I don't know the architecture to the degree I know Windows.

In Windows, you can set process "affinity", locking it down to a fixed processor core, through Task Manager. Don't know if you can do that in OSX...

Half Glass

Sep 13, 2006, 02:15 PM

No software such as, Cinema 4D, Motion, Aperture, Final Cut Pro etc

I don't know about this statement.

From my usage of FCP, Compressor, Aperture and DVDSP, they work very well with the MacPro but I haven't seen them approach usage of even 3 full cores.

Ability to multistask is great but I would not say that any one of the above is using all cores the way we want them to. I would contend that this is coming and pointed out in another thread that some of the FCP benchmarks on Apple's MacPro performance page are footnoted that the figures given were using Beta version of FCP.

--HG

ergle2

Sep 13, 2006, 02:18 PM

Multimedia

Sep 13, 2006, 02:21 PM

A bit pointless given that no software utilises the extra cores yet. But nice to know, I guess.No software such as, Cinema 4D, Motion, Aperture, Final Cut Pro etcNo software such as Toast 7.1, Handbrake UB. More to the point is not how many cores an application can use but rather how many things you can get done at once. :rolleyes:From my usage of FCP, Compressor, Aperture and DVDSP, they work very well with the MacPro but I haven't seen them approach usage of even 3 full cores.

Ability to multistask is great but I would not say that any one of the above is using all cores the way we want them to. I would contend that this is coming and pointed out in another thread that some of the FCP benchmarks on Apple's MacPro performance page are footnoted that the figures given were using Beta version of FCP.I think in the next few months the full FCS and Logic will get an update to address this.

shawnce

Sep 13, 2006, 02:22 PM

As for using a Dell, sure they could've used that. Would Windows use the extra 4 cores? Highly doubtful. Microsoft has sketchy 64 bit support let alone dual core support; I'm not saying "impossible" but I haven't read jack squat about any version of Windows working well with quad cores. You think those fools (the same idiots who came up with Genuine Advantage) actually optimized their OS to run in an 8 core setup? Please pass along what you're smoking. :rolleyes:

:confused:

Yeah because whatever you [zero2dash] are smoking is really screwing with your mind... best to get something else.

ergle2

Sep 13, 2006, 02:30 PM

No software such as Toast 7.1, Handbrake UB. More to the point is not how many cores an application can use but rather how many things can you get done at once. :rolleyes:I think in the next few months the full FCS and Logic will get an update to address this.

One thing to note is that IO may become a more limiting factor than number of cores under heavy multitasking, or even just particularly data-heavy apps (multiple streams of raw hires video, for example).

Multimedia

Sep 13, 2006, 02:34 PM

One thing to note is that IO may become a more limiting factor than number of cores under heavy multitasking, or even just particularly data-heavy apps (multiple streams of raw hires video, for example).When we hit that wall, I'll let you know. :p

Hey everybody the Big News is
NBC Today Show went High Definition today!

CBS Morning now looks totally pathetic. What a revolting development. CBS hires Katie and don't advance the news department to high-def while NBC totally goes High Def with Merideth. Wow! Talk about a cou-de-gras!

NBC Nightly News can't be far behind. I'm so excited. Getting my Mom a High Definition Set For Christmas.

janstett

Sep 13, 2006, 02:37 PM

The OS takes advantage of the extra 4 cores already therefore its ahead of the technology curve, correct? Gee, no innovation here...please move along folks. :rolleyes:

As for using a Dell, sure they could've used that. Would Windows use the extra 4 cores? Highly doubtful. Microsoft has sketchy 64 bit support let alone dual core support; I'm not saying "impossible" but I haven't read jack squat about any version of Windows working well with quad cores. You think those fools (the same idiots who came up with Genuine Advantage) actually optimized their OS to run in an 8 core setup? Please pass along what you're smoking. :rolleyes:

Sorry to burst your reality distortion field, but see my previous post. I ran a dual processor Pentium II NT setup ten years ago and Windows handled it just fine THEN -- back when Apple barely supported it with a hack to its cooperatively-multitasked OS and required specially written applications with special library support.

BTW my 2 year old Smithfield handles 4 processors fine (Dual Core Pentium Extreme with hyperthreading = 4 cores).

The only limit with Windows is they keep the low end XP home to 2 processors on the same die. There is probably an architectural limit on both OSX and XP and if it's not 8 it's 16. It's probably 8.

patrick0brien

Sep 13, 2006, 02:37 PM

I smell it an option for Rev. B.

As Mac Daily News says: "Mac Pro Octo-Core. For when you absolutely, positively have to sequence the entire human genome before lunch."

Naaaaaaaaaaaaaaarrrrrrrrrrrrrrr!

gugy

Sep 13, 2006, 02:47 PM

NBC Nightly News can't be far behind. I'm so excited. Getting my Mom a High Definition Set For Christmas.

Hey Multimedia,
I have a question for you.
Do you record HDTV with EyeTV 500 then encode to H.264 using Handbrake and then do you add it to itunes to manage and organize those shows or movies?

I think this is a neat idea with you have the spare HD room and want to keep shows or events for long time and want to access it fast and easy.

Cheers!

freezerburrn

Sep 13, 2006, 02:50 PM

After reading the Anandtech article, I'm curious to know how much of a concern the FBD memory latency is to some of you, seeing that the Core 2 extreme came out on top in many of the benchmark tests.

scottlinux

Sep 13, 2006, 02:54 PM

ergle2

Sep 13, 2006, 02:58 PM

The only limit with Windows is they keep the low end XP home to 2 processors on the same die. There is probably an architectural limit on both OSX and XP and if it's not 8 it's 16. It's probably 8.

There's a bunch of HP Superdome 64-way Itanium systems around running Windows Server 2003, mostly for MS SQL work.

Windows XP 64bit is based on the same core. Given the license is per-socket, not per-processor (currently, anyway) and the Pro editions support two sockets, it should in theory support the 8-way setup as described by Anandtech.

Whether it recognises quad-core CPUs as such may of course be a different matter.

epitaphic

Sep 13, 2006, 03:00 PM

I think you've misunderstood. Merom/Conroe/Woodcrest are one microarch now. That's Intel's point -- the core is essentially the same.

Conroe and its derivatives are a step away from Intel's former flagship NetBurst, but even these processors are a bit of a dying breed: during Intel's shift to 45nm, the company will no longer focus on derived microprocessor cores in favor of refined unified core architectures.
So what do you think they meant with M/C/W being a derived arch and Penryn,etc being unified archs?

From what I understood, they'll stop having different characteristics (FSB,RAM,Cache) and instead just differentiate them with MHz and core count. Hence all the stories that future Intel chips (starting with Penryn I presume) won't use FSB.

savar

Sep 13, 2006, 03:35 PM

NOT TRUE....The Quad core G5 people are in an uproar because Logic Pro only uses 2 cores on the G5....they updated Logic Pro so it uses 4 cores, but the G5 Quad still only uses 2 cores....there are also photoshop actions that are NOT multi core aware so will only run on one core.....Hopefully 10.5 will make all this irrelevant.

You totally missed my point. Even if an application uses only one thread at all times, that application is still a separate process from all of the other processes you have running. At any given time you'll have at least 30 something processes, even when no user-land applications are running. OS X will spread out those processes to try to utilize all the cores as much as possible.

In reality, there are probably not too many non-Apple applications which routinely use 8 threads or more. In the near future I expect all applications to use at least 2-3 threads, even the most simple ones.

ergle2

Sep 13, 2006, 03:40 PM

So what do you think they meant with M/C/W being a derived arch and Penryn,etc being unified archs?

From what I understood, they'll stop having different characteristics (FSB,RAM,Cache) and instead just differentiate them with MHz and core count. Hence all the stories that future Intel chips (starting with Penryn I presume) won't use FSB.

I believe you've got it backwards. Penryn is a derived arch (check the diagram) -- it's derived from Conroe/Merom, etc., ie it's based on them with "more" -- faster FSB, more cache, a die shrink (which is technically less... :) ) etc.

Unified just means the micro-arch itself the same rather than the entire CPU. This is already true of Core2, and is significantly cheaper in terms production costs. Merom/Conroe are literally the same core in a different package, specified for different voltage/clockspeeds. I'm not sure if Woodcrest is but it seems highly likely.

The one oddity I am aware of is Allendale isn't a Conroe with half the cache disabled, it's actually a specific die. The rest of the microarch itself is the same, however.

Nehalem, etc. aren't derived because they're a new microarch. (Interestingly, Nehalem was originally intended for launch early 2007).

CSI replacing FSB was originally planned for 2006 in older roadmaps. It now looks like a 2008 debut with Tukwila (Itanium, not x86), and will no doubt work its way down from there.

ergle2

Sep 13, 2006, 04:02 PM

You totally missed my point. Even if an application uses only one thread at all times, that application is still a separate process from all of the other processes you have running. At any given time you'll have at least 30 something processes, even when no user-land applications are running. OS X will spread out those processes to try to utilize all the cores as much as possible.

In reality, there are probably not too many non-Apple applications which routinely use 8 threads or more. In the near future I expect all applications to use at least 2-3 threads, even the most simple ones.

Sure, but all those background processes take next to no time to execute -- the extra latency of having more processors will probably slow things down far more than you gain from having up to 8 of those 30 be able to run at any one time.

I'm not saying there's no need for 8 cores -- markets such as databases, media production, rendering, etc. can already make use of that kind of power.

Regular desktops, not so much.

Many simple apps are already mutithreadedto some dgree, but it's to make them non-blocking rather than to spread processor load. If you look at Windows, you'll find a very high number of threads in even just a media player, but some of it's just there to repaint the GUI etc.

amin

Sep 13, 2006, 04:13 PM

The MP is so overkill for my needs right now, I wonder if I'd even notice the difference. I think I'll wait for 32 cores before I update!

ATD

Sep 13, 2006, 04:20 PM

A lot of 3d programs will use as many cores as are available when rendering.

And I would say that the next versions of many programs will be better suited for multiple core processors.* They are way too common for software developers to ignore them any longer.

Yep. Not all of the software I use taps all the cores but the 3D renders I do shallow every inch of the CPUs. I have Maya with Mental Ray hooked to 2 computers, a quad and a dual. When I hit render the CPU usage hits 100% on all 6 processors. While having all these processors working is great I have noticed that my quad has lots of pauses in the finder doing simple things, even if nothing is running. Everyone else I know that has a quad has the same issue. I have to believe that there is a trade off for having all these processors, it seems they trip over each other on the small stuff. I hope the next version of OSX will take a look at this, in light of the fact we will be jumping to 8 or more processors.

ergle2

Sep 13, 2006, 04:58 PM

The MP is so overkill for my needs right now, I wonder if I'd even notice the difference. I think I'll wait for 32 cores before I update!

In my experience, single to dual is a huge step up because the interactive response feels much snappier (esp on Windows, but that's not terribly relevant).

After that it gets less noticable -- but then, most long-term CPU bound work isn't interactive anyway, games aside :)

iSee

Sep 13, 2006, 05:38 PM

Damn! That's just cool (well, cool in a very geeky sort of way :rolleyes: ).

bob5820

Sep 13, 2006, 05:57 PM

Trekkie

Sep 13, 2006, 06:42 PM

According to tha Anandtech article its likely that the Clovertown family will be clocked slower then the Woodcrests

clock speed isn't everything. workload dependant of course.

suneohair

Sep 13, 2006, 07:05 PM

Sorry to burst your reality distortion field, but see my previous post. I ran a dual processor Pentium II NT setup ten years ago and Windows handled it just fine THEN -- back when Apple barely supported it with a hack to its cooperatively-multitasked OS and required specially written applications with special library support.

BTW my 2 year old Smithfield handles 4 processors fine (Dual Core Pentium Extreme with hyperthreading = 4 cores).

The only limit with Windows is they keep the low end XP home to 2 processors on the same die. There is probably an architectural limit on both OSX and XP and if it's not 8 it's 16. It's probably 8.

Didn't you get the memo, Hyperthreading was a joke.

On to this post. The current Mac Pro is not going to be upgraded. Nor will Quad be making its debut, at least at the current price points, anytime soon. It may be offered as a very expensive upgrade but thats about it looking ahead a year.

So for those who bought now, it was a good choice. When the time comes and cost is low they can take that next step and upgrade.

suneohair

Sep 13, 2006, 07:22 PM

if it follows typical intel transitions price point replace. So the same price as woodcrests. They might introduce faster ones though that cost more. We'll see before the end of the year.

Dual core isnt transitioning into quad core anytime soon. Quad core chips will probably be a very high end part for quite sometime.

This isnt going to be akin to the Core Duo --> Core 2 Duo changeover. Cores are increasing which means price will increase.

suneohair

Sep 13, 2006, 07:26 PM

clock speed isn't everything. workload dependant of course.

You are right. However, you try to tell consumers "Well we are moving to 2.4Ghz chips" after you just had 2.66Ghz and 3.0Ghz chips. It isnt going to work.

If today, Dell decided to move there whole line back to 1Ghz processors, nobody would buy. Unfortunetly the Ghz myth is a strong as its ever been. Taking a step backward is not an option.

Another example would be this: Today Apple decides to go back to plain, bulky ipods, no color, no photos. Just monochrome and music. Would anybody go for it? Probably not. You just can't step back in tech today.

Don't get me wrong, I am sure the octo core would out perform the current quad anyday given the right apps. But when people see that Ghz number go down...

aarond12

Sep 13, 2006, 07:50 PM

Didn't you get the memo, Hyperthreading was a joke.

Obviously, since Intel is no longer creating new processors with HT.

By the way, previous poster, HT does not double the number of cores. Just the number of virtual cores. A Pentium 4 system with HT will run slower than a dual Pentium 4 system (with HT disabled) at the same clock speed.

Dual-core means there are essentially two separate CPUs on a single die.

Oh, and to the IT person who didn't know Clovertown is 64-bit... congratulations. You've shown everyone that you have an exceedingly small penis in having to argue that a $3000 Mac is slower than a $30,000 Windows DataCenter Edition PC system.

-Aaron-

ssamani

Sep 13, 2006, 08:15 PM

DAMN :eek:

so 2-3 years from now are people going to be asking "do I need a quad core or an 8 core macbook? oh yeah I'll mostly be surfing the web and maybe editing a photo once and a while" :rolleyes:

I'll be mostly surfing the web to simultaneously stream multiple TV channels, download HD movies and video chat with friends with a live video substitution background and maybe editing a 16 Megapixel camera phone photo and the odd HD camcorder movie once in a while.

Hey, as Steve said, 128K is more than anyone will ever need...

ergle2

Sep 13, 2006, 08:19 PM

Obviously, since Intel is no longer creating new processors with HT.

By the way, previous poster, HT does not double the number of cores. Just the number of virtual cores. A Pentium 4 system with HT will run slower than a dual Pentium 4 system (with HT disabled) at the same clock speed.

Actually, many tasks were faster.

HyperThreading was thrown in to mask other deficiencies in the NetBurst arch by exploiting resources that were otherwise wasted.

There were a few cases where HT ran slower when HT first debuted, but with OS scheduler tweaks and BIOS updates (microcode changes, likely), HT was a net win in most cases.

Core 2 doesn't have the same design issues - mostly down to that excessively long pipeline - that Prescott had, and hence HT makes no sense.

The problem, however, lay with Netburst as a whole, rather than HT -- which offered a minor improvement in performance - a band-aid if you will.

ssamani

Sep 13, 2006, 08:28 PM

I have to say that I think a lot of people have missed the point around various software not using all the cores. There is a simple reason. Developing multi threaded apps is hard. And until recently (~1 yr since G5 Quad) developing for more than 2 simultaneous threads (cores or processors) was pretty pointless on a Mac and completely pointless on a PC. Why would a developer bother to develop for more than two thread unless they expect threads to get blocked easily or there to be more than two cores? That's why browsers are one of the few genuine multithreaded apps, rather than 2-4 threads, the threads block easily, so having multiple makes sense. In a lot of cases where there was the opportunity for parallelism using a vector processing unit often makes more sense than multiple threads. Why spend development and testing effort on a solution with no hardware to even test it on, let alone deliver and make use of?

The whole industry is taking a 90 degree turn and the tools for developing multithreaded applications are not up to scratch. Look at some of the commentary about XBox 360 and PS3 development.

The dev tools will come, the software will catch up with the hardware, in the meantime just be glad that you can play your stolen MP3's and browse your pr0n without interfering with each other and stop whinging.

Willis

Sep 13, 2006, 08:37 PM

One could run a Folding@Home process on each core :D

ooooor.... use multiple cores to do one fold... 4 days like my g5 would cut down to like 16 hours... thats mental. awesome... but god damn mental!!

stapler

Sep 13, 2006, 08:57 PM

I doubt anybody runs more than eight really hardcore apps at once.

Multimedia

Sep 13, 2006, 09:23 PM

Hey Multimedia, Do you record HDTV with EyeTV 500 then encode to H.264 using Handbrake and then do you add it to itunes to manage and organize those shows or movies?

I think this is a neat idea with you have the spare HD room and want to keep shows or events for long time and want to access it fast and easy.I don't use H.264 because the previous max res allowed to go on an iPod with it was 320x240 which would upscale to a TV poorly. Now I will continue to not use it because the file sizes are more than twice what I can make without it.

The author of Handbrake is going to have to make some changes to it before we can use H.264 judiciously. I currently use the FFmpeg 2-pass encoding at bitrates around 750kbps for SD @ 544x400 and 1000kbps @ 624x352 for HD. These are dimensions that fall within the upper limit pixel count that will still load and play on an iPod 230,000. This way they upsacle to a TV well and still play on iPod.

I don't use iTunes much. I am burning DVDs of all these files. 12 one hour shows fit on a DVD @ 351 per 42 minute Ads edited out episode. Movies I can get to 702MB - one CD size each. Personal message me if you need more details.

drsmithy

Sep 13, 2006, 09:41 PM

Mac OS X distributes threads and processes across cores/CPUs to optimize performance already. (Subject to some limitations, as noted already.)

[...]

(Note: I keep specifying 'Mac' here. There is a reason. Windows isn't as good at multithreading/processing yet...)

Uh, no. Windows NT is better at multithreading - and particularly multiprocessor scheduling - largely because it's been doing it for a lot longer and on a lot more powerful hardware. NT was running on quad-processor machines a decade ago.

Prior to 10.4, OS X had roughly the same level of SMP support Windows NT had back around the 1993 - 95 timeframe, with Windows NT 3.x.

The improvements in 10.4 start to put it in the ballpark of NT 4.0, ca. 1996.

10.5 will probably put it on par with Windows 2000, maybe XP.

drsmithy

Sep 13, 2006, 10:08 PM

Would Windows use the extra 4 cores?

Yes. Windows NT was running on machines with eight processors several years before OS X was even released.

Microsoft has sketchy 64 bit support let alone dual core support;

Windows supported 64 bit platforms and dual core CPUs long before OS X did.

I'm not saying "impossible" but I haven't read jack squat about any version of Windows working well with quad cores.

That's probably because you're not interested in reading anything that might portray Microsoft in a non-negative light.

You think those fools (the same idiots who came up with Genuine Advantage) actually optimized their OS to run in an 8 core setup?

The people who have been running Windows on 8, 16, 32 and 64 processor systems for many years already seem to think they did.

Windows has much better multiprocessor capabilities than OS X does - and has done since before OS X was released.

Multimedia

Sep 13, 2006, 10:18 PM

Hey everybody the Big News is
NBC Today Show went High Definition today!The Today show is an embarrassment. The US major tv networks do not have any real morning news programs. How to trim your dog's ears and an inside look into American Idol contestants is NOT NEWS. It is an entertainment talk show.My post of this news has nothing to do with content. It is a historic moment in the evolution of television. If you don't think so, then you are paying way too much attention to the content and not enough to the process by which they are conveying that content.

iJawn108

Sep 13, 2006, 10:22 PM

:p very cool.

I may purchase an 8 core mac pro if they become available. I just love things in 8s

scottlinux

Sep 13, 2006, 11:32 PM

If you don't think so, then you are paying way too much attention to the content and not enough to the process by which they are conveying that content.

Well if the content is crap, who cares to watch? Content of TV is more important to me. I'd rather see a fascinating news show or program over rabbit ears than watch the Today Show in HD.

My post of this news has nothing to do with content.

This wasn't clear the first time. You sounded like a crazed American Idol fan with your original post. And HD broadcasts are nothing new...

slabbius

Sep 13, 2006, 11:44 PM

you know what? since my dinosaur of a desktop (3yr old :rolleyes: 3Ghz P4 HT that can't even run a retail 3DSMax without me getting fatal exception blue screen of death errors on winxpsp2) the time value of money says that a new Mac Pro Quad Core machine is still worth more now than a Mac Pro Octo Core machine in the future. Reason is I need a much more viable means of work NOW, not later. I can always upgrade, and besides, the new chips will probably be rather pricey, therefore causing a rise in the current mac pro price? I'm no analyst so don't flame me if i'm wrong. ;)

Besides I'm a young full sail student that just got an educational loan to purchase a computer and a camera.... and maybe an ipod :) Don't try to give me the "if you wait" lecture, either.

FFTT

Sep 14, 2006, 12:07 AM

The current Mac Pro and even the Core Duo 2 iMacs are both amazing upgrades
to what most people were running 2 years ago.

Generally, I recommend waiting for those who already have a working OSX Tiger
machine.

However, if you NEED a new Pro workstation, the Mac Pro will pay for itself
in productivlty.

We also don't know how cool these new quad core processors will run.

At least we do know that the current Mac Pro is extemely quiet with power to spare.

Multimedia

Sep 14, 2006, 12:20 AM

Well if the content is crap, who cares to watch? Content of TV is more important to me. I'd rather see a fascinating news show or program over rabbit ears than watch the Today Show in HD.

This wasn't clear the first time. You sounded like a crazed American Idol fan with your original post. And HD broadcasts are nothing new...This is NEW because it is on a 3 hour weekday morning telecast. That makes it NEW and NEWS. Nothing about content. I NEVER watch American Idol. You are judgemental.

Millions watch that crap so your opinion of it is irrelevant to the market forces.

Evangelion

Sep 14, 2006, 02:14 AM

Didn't you get the memo, Hyperthreading was a joke.

At worst, it slowed performance down by few percent. At best, it gave substantial boost in performance. And multitasking-tests clearly benefitted from HyperThreading. That said, Intel dropped it, because it apparently consumed too much power. But we might see HT in some future Intel-CPU's at some point, you never know.

HT as such is not a bad idea. Sun UltraSparc T1 uses such a scheme extensively.

Voltes V

Sep 14, 2006, 03:51 AM

i think they're coming up with 2 dual octo-core.......

zero2dash

Sep 14, 2006, 09:33 AM

Yes. Windows NT was running on machines with eight processors several years before OS X was even released.

Windows supported 64 bit platforms and dual core CPUs long before OS X did.

On the server side.
Nevertheless, ok. Windows did it first.

That's probably because you're not interested in reading anything that might portray Microsoft in a non-negative light.

Couldn't be farther from the truth. I have no problem with Microsoft or Windows, evident by the fact that I've ran their operating systems for the last 10 years. I have a problem with all the crap they're putting in Vista, but otherwise - Win2k and XP Pro have left me primarily trouble-free.

Evangelion

Sep 14, 2006, 09:56 AM

On the server side.

Plenty of people ran NT on their desktops.

Nevertheless, ok. Windows did it first.

Admission of your mistakes is a good step in becoming a better person.

drsmithy

Sep 14, 2006, 11:05 AM

On the server side.

The server/desktop division with Windows - as with OS X - is one of marketing, not software. Windows "Workstation" and Windows "Server" use the same codebase.

Couldn't be farther from the truth. I have no problem with Microsoft or Windows, evident by the fact that I've ran their operating systems for the last 10 years. I have a problem with all the crap they're putting in Vista, but otherwise - Win2k and XP Pro have left me primarily trouble-free.

Well, if you can't find evidence of Windows running on well on machine with >2 processors, or of the significant low-level changes Microsoft have made to ensure it does, you aren't looking very hard.

Similarly, if you're one of the "Vista is just XP with a fancy skin" crowd, you've obviously not done much research. The changes in Vista are on par with the scale of changes Apple made to NeXT to get OS X.

Multimedia

Sep 14, 2006, 01:23 PM

i think they're coming up with 2 dual octo-core.......True That. But not until next summer 2007.

RedTomato

Sep 14, 2006, 01:32 PM

i think they're coming up with 2 dual octo-core.......True That. But not until next summer 2007.

:eek: :eek:

What's planned after that? 16 cores on a chip? Seriously?? :confused: :confused:

Half Glass

Sep 14, 2006, 01:49 PM

Well, if you can't find evidence of Windows running on well on machine with >2 processors, or of the significant low-level changes Microsoft have made to ensure it does, you aren't looking very hard.

Agreed.

Don't forget the new MacPros where XP runs very well (minus the MB chipset/SATA issue where there is a workaround.) It recognizes all four cores and seems very stable.

--HG

brianus

Sep 14, 2006, 01:56 PM

The server/desktop division with Windows - as with OS X - is one of marketing, not software. Windows "Workstation" and Windows "Server" use the same codebase.

True (today anyway; in the NT era they were indeed separate platforms though. Which brings me to my next point..)

Similarly, if you're one of the "Vista is just XP with a fancy skin" crowd, you've obviously not done much research. The changes in Vista are on par with the scale of changes Apple made to NeXT to get OS X.

I think people who say stuff like that are exhibiting a syndrome common to Mac folk who've never spent any time in the PC world -- they take negative comments they remember regarding versions of Windows or the PC experience from about 5 years back and assume they apply to today. XP, for example, really was for the most part a window-dressing of Windows 2000, but that is not the case for Vista. You see similar statements regarding "blue screens of death", overall system stability, etc, which suggest they haven't seen or used a PC since the late 90s/early 00's.

brianus

Sep 14, 2006, 02:10 PM

This is NEW because it is on a 3 hour weekday morning telecast. That makes it NEW and NEWS. Nothing about content. I NEVER watch American Idol. You are judgemental.

Well he did kind of have a point there at the end. You could stand to tone down your use of enormous, colored type if you don't want your posts "judged" as having an emotional content beyond their actual words.

ergle2

Sep 14, 2006, 02:17 PM

True (today anyway; in the NT era they were indeed separate platforms though. Which brings me to my next point..)

Point of total (and obnoxious) pedantry here -- XP and W2K3 Server aren't strictly the same codebase; The latter was a huge rewrite job with some fairly significant internal changes.

XP 64bit is based on W2K3, and Vista originally started out on the XP code base and then was scrapped, and was started over using the W2K3 codebase.

It doesn't invalidate your point in any way and the latter is most definitely descended from the former, but unlike previous products they weren't released in parallel. I mention it purely because I find it interesting, and it's also an example of how Windows is "evolving", so to speak.

I think people who say stuff like that are exhibiting a syndrome common to Mac folk who've never spent any time in the PC world -- they take negative comments they remember regarding versions of Windows or the PC experience from about 5 years back and assume they apply to today. XP, for example, really was for the most part a window-dressing of Windows 2000, but that is not the case for Vista. You see similar statements regarding "blue screens of death", overall system stability, etc, which suggest they haven't seen or used a PC since the late 90s/early 00's.

This is very common on both sides of the divide. Many Mac-only people seem to think Windows is still stuck in the Win9x days, and many of the Windows-only types seem to think MacOS is still in the 8.x days.

I guess it's a little like when your friend has kids and you don't see them for a few years, and you're surprised that instead of still being little kids they're teenagers... :)

aricher

Sep 14, 2006, 05:49 PM

He's totally mistaken! The Cloverton CPUs will *all* be 64-bits, as Woodcrest (found in current Mac Pros) is. Intel is not going to ever go back to a 32-bit Xeon class CPU.

The difference between Woodcrest and "Tigerton" is that Woodcrest CPUs achieve their "dual core" status by basically placing two complete Xeon CPUs under one outer casing, and making them communicate with each other through the front-side bus on the motherboard.

Cloverton will be the same way, but with 4 cores packed into one casing, instead of just two.

"Tigerton" will finally allow both cores to interconnect with each other through an internal interface built into the CPU, instead of slowing communications down by routing it off one CPU core, through the motherboard's front-side bus, and back onto the other core.

I got this great response this morning from my IT snob:
"Where in that linked article does it say 64bit? I see 65 nm, but not 64 bit. Duct taping two 32 bit cores together may get you Mac 64 bit processing... great for drawing cool pictures."

Anyone have a link that shows that Clovertown is 64 bit? Please help me to defeat this PC IT ogre

Trekkie

Sep 14, 2006, 06:22 PM

ergle2

Sep 14, 2006, 06:29 PM

I got this great response this morning from my IT snob:
"Where in that linked article does it say 64bit? I see 65 nm, but not 64 bit. Duct taping two 32 bit cores together may get you Mac 64 bit processing... great for drawing cool pictures."

Anyone have a link that shows that Clovertown is 64 bit? Please help me to defeat this PC IT ogre

Straight from the horse's mouth at Intel (http://idfemea.intel.com/moscow/download/moscow_final_ru_and%20_eng/SRV/SRVM01.pdf) (Clovertown supports EM64T), and again at Intel (http://www.intel.com/performance/server/xeon/intthru.htm) (5100 series supports EM64T), and once more from Intel (http://www.intel.com/pressroom/archive/releases/20050926corp_a.htm) (Xeon is 64bit, mentions Woodcrest).

Took me about 10 seconds. Your "friend" is either a troll or supremely ignorant -- especially when you bear in mind this is the same hardware that Windows runs on. Apple pretty much supplies a nice case and the OS at this point.

Edit: too slow... I was busy with RL in the background... ah, well... :)

Multimedia

Sep 14, 2006, 06:56 PM

:eek: :eek: What's planned after 8 core processors? 16 cores on a chip? Seriously?? :confused: :confused:Not sure about beyond 8 which can be paired into a 16 core Mac. Perhaps. Too far out to tell although it is casually mentioned in the roadmap.

ergle2

Sep 14, 2006, 09:09 PM

Not sure about beyond 8 which can be paired into a 16 core Mac. Perhaps. Too far out to tell although it is casually mentioned in the roadmap.

New micro-arch -- Nehalem is due 2008.

AidenShaw

Sep 14, 2006, 09:23 PM

Too far out to tell although it is casually mentioned in the roadmap.
The next versions of the roadmaps will be discussed at the Intel Developer's Forum (http://www.intel.com/idf/us/fall2006/index.htm) a week from Tuesday in the City By The Bay.

I've heard that as attendees we'll find a Kentsfield and a pair of Clovertowns taped under our seats at the keynote :cool: :cool: .

drsmithy

Sep 14, 2006, 09:23 PM

True (today anyway; in the NT era they were indeed separate platforms though. Which brings me to my next point..)

I think you're a bit arse-about-face there. Someone else has already pointed out the differences between XP and Windows 2003 aren't trivial, so I won't go into that. However, if you're sufficient vintage, you should remember the "outrage" when someone demonstrated that you could turn NT 4 Workstation into NT 4 Server (including the boot and login screens) just by changing a few Registry settings (although the part that usually doesn't get said is that those Registry settings then triggered a whole range of different tuning settings for the scheduler, memory management, etc). NT 3.5 & 3.51 were the same, and IIRC, NT 3.1 didn't even have a "Server" version.

ergle2

Sep 14, 2006, 09:42 PM

I think you're a bit arse-about-face there. Someone else has already pointed out the differences between XP and Windows 2003 aren't trivial, so I won't go into that. However, if you're sufficient vintage, you should remember the "outrage" when someone demonstrated that you could turn NT 4 Workstation into NT 4 Server (including the boot and login screens) just by changing a few Registry settings (although the part that usually doesn't get said is that those Registry settings then triggered a whole range of different tuning settings for the scheduler, memory management, etc). NT 3.5 & 3.51 were the same, and IIRC, NT 3.1 didn't even have a "Server" version.

The comments about separate platforms in the NT era I took to refer to NT3.x/4 vs Win9x.

Quite a few bits of XP Pro functionality can be enabled in XP home with some minor hex editing, too.

And of course, NT started as a reimplementation of VMS for a failed Intel RISC CPU...

Voltes V

Sep 14, 2006, 10:24 PM

:eek: :eek:

What's planned after that? 16 cores on a chip? Seriously?? :confused: :confused:

yeah, who would've thought we're having quad core 4 years ago.

amin

Sep 14, 2006, 11:08 PM

brianus

Sep 14, 2006, 11:23 PM

AnandTech is putting a lot of emphasis on this FB-DIMM issue. Their Conroe vs Xeon comparisons are poor given that they maximize the FB-DIMM latency "problem" by using a Mac Pro with only two RAM slots occupied. Seems as though they have an agenda to exaggerate the importance of this technical issue.

I have noticed this emphasis as well; not being an expert on this issue myself though, would you care to shed light on how their coverage is an exaggeration and why we shouldn't be worried about it?

The comments about separate platforms in the NT era I took to refer to NT3.x/4 vs Win9x.

Yes, this is what I was getting at. ("arse about face"? What is that, Swedish? :rolleyes: ). Noone other than a vintage Windows IT person would know there were further differences between versions of NT itself. Also when making comparisons I never mentioned Server 2003 (about which I know almost nothing); I was talking about XP and 2000 being relatively similar whereas, for example NT and 98 were not.

New micro-arch -- Nehalem is due 2008.

Really, completely new? As in, to Core 2 what the G5 was to G4? In just two years?? I guess they're really ramping things up... Core 3 Hexa Mac Pros, anyone?

ergle2

Sep 14, 2006, 11:49 PM

Really, completely new? As in, to Core 2 what the G5 was to G4? In just two years?? I guess they're really ramping things up... Core 3 Hexa Mac Pros, anyone?

Intel's stated plans as I understand them are thus:

A new micro-arch every 2 years. I don't think they mean brand new so much as "significant changes/improvements". Whether this is akin to Yonah->Conroe or Netburst->Conroe remains to be seen, but more like the former (or perhaps Pentium-M -> Merom -- Core Duo was very much a stop-gap). Little has been released about Nehalem, but at one time it was slated as "based on Banias/Dothan", due in 2005 and expected to ramp to 9/10GHz.

"Off" years will recieve derivative versions (e.g. Merom->Penryn), which appears to be mostly stuff like L2 cache increases, faster FSB speeds (at least while we have FSBs - 2008 looks like the year for DCI, finally), die shrinks, increasing the number of cores (expect at least one to be more cores on a single die instead of two dice/package), etc.

Die shrinks are currently scheduled for "off" years, in order to stablize the process ready for the new micro-arch in the following year so Intel doesn't need to deal with both new process and new arch at the same time, and presumably in part to keep speed increases coming in "off" years

Of course, roadmaps can change quite rapidly -- it's not that long ago that Whitfield was expected to debut late 2006 with DCI (FSB replacement). Whitfield was replaced by Tigerton which is now due sometime in 2007...

One thing's for sure, Intel appears to have learnt a great deal from the Netburst fiasco -- how not to do things, if nothing else. Unfortunately, they still estimate ~50% of processors shipping in 1Q2007 will be netburst-based (mostly Pentium-D).

amin

Sep 14, 2006, 11:53 PM

I have noticed this emphasis as well; not being an expert on this issue myself though, would you care to shed light on how their coverage is an exaggeration and why we shouldn't be worried about it?

I am no expert, and I am not denying that this issue matters. However, I see no cause for concern unless someone provides some decent evidence that it matters. It strikes me as odd that they (at AnandTech) put so much emphasis on explaining the theory behind a "problem" without making any competent effort at illustrating an example of the problem. When you go to configure a Mac Pro, the Apple page says the following about memory: "Mac Pro uses 667MHz DDR2 fully buffered ECC memory, a new industry-standard memory technology that allows for more memory capacity, higher speeds, and better reliability. To take full advantage of the 256-bit wide memory architecture, four or more FB-DIMMs should be installed in Mac Pro." Yet AnandTech chose a 1GB x 2 RAM arrangement to compare the Core 2 Extreme and Xeon processors. Using this setup, which effectively cripples the Mac Pro memory system, they find it to be at worst 10% slower than the Conroe Extreme (in a single non real world usage benchmark). Meanwhile in any comparison that utilizes the four cores, the quad Xeon whoops ass by a large margin.

Silentwave

Sep 15, 2006, 12:00 AM

One thing's for sure, Intel appears to have learnt a great deal from the Netburst fiasco -- how not to do things, if nothing else. Unfortunately, they still estimate ~50% of processors shipping in 1Q2007 will be netburst-based (mostly Pentium-D).

It is a shame, but sadly those are the real cheap chips right now. The good news is that they'll change those over soon enough with more Allendales, then millville and so on and so on taking on more segments of the market.

I think as they transition to 45nm we'll see more and more Core chips, simply because they'll want as much manufacturing to be on the new process as possible, and they don't need to scale the D's etc. down to it.

ergle2

Sep 15, 2006, 12:29 AM

It is a shame, but sadly those are the real cheap chips right now. The good news is that they'll change those over soon enough with more Allendales, then millville and so on and so on taking on more segments of the market.

I think as they transition to 45nm we'll see more and more Core chips, simply because they'll want as much manufacturing to be on the new process as possible, and they don't need to scale the D's etc. down to it.

Indeed. The Netburst chicken has been decapitated, it just hasn't yet stopped running around the marketplace...

I think Intel wants the transition to go as quickly as possible, given the aggressive pricing of Core 2 - not as cheap as Pentium D, but a much better bang for the buck, so to speak. Of course, that's also connected to trying to beat back the AMD surge of recent years...

AidenShaw

Sep 15, 2006, 07:40 AM

And of course, NT started as a reimplementation of VMS for a failed Intel RISC CPU...
A cancelled Digital RISC CPU.

Although, some of the ideas for the cancelled CPU ended up in the Alpha chips.

janstett

Sep 15, 2006, 08:41 AM

Didn't you get the memo, Hyperthreading was a joke.

Didn't you get the memo, PowerPC is dead. WTF does that have to do with anything? Do you just have this Pavlovian response to the word "Hyperthreading"?

I fully understand how Hyperthreading works -- in certain situations the processor can run two instructions simultaneously. Not all situations, however. So sometimes a single CPU can act like it is dual core, other times it cannot, depending on the independence of the two threads. It's like having an ambidextrous person instead of two people.

It was a top of the line processor when I bought it ~18 months ago. It is a DUAL CORE processor before Hyperthreading even comes into the picture. With Hyperthreading on it looks like 4 processors to Windows. So, what, should I turn off Hyperthreading just because you don't like it? Am I supposed to stop using the machine? Boob.

simontarr

Sep 15, 2006, 08:46 AM

I think all this quad and oct core stuff is fantastic (it would be even more fantastic if I have the money to get such gear...)

But at the moment it's the HDD that slows everything down. Your RAM may be able to send 4GB/s of data to the processor to deal with, but the HD can't write the said executed data at even a 10th of the speed.

I remember reading a BBC news article the other month about mRAM (or magnetic RAM) which has the same write speeds as RAM, but without its volatility. It doesn't loose it's data when the power is off. Ideal for fast HDDs they say.

On an unrelated note, wouldnt it been cool to effectivly install a whole OS on RAM. That would be noticably quicker....

janstett

Sep 15, 2006, 08:48 AM

The Today show is an embarrassment. The US major tv networks do not have any real morning news programs. How to trim your dog's ears and an inside look into American Idol contestants is NOT NEWS. It is an entertainment talk show.

The network morning "news" shows have always been fluff. What's worse is that the so-called "hard news" shows are just as bad, and not just in the morning -- CNN, MSNBC, and Fox News all run mindless fluff instead of news. And don't get me started with MSNBC airing Eye-Puss in the Morning.

janstett

Sep 15, 2006, 08:57 AM

I remember reading a BBC news article the other month about mRAM (or magnetic RAM) which has the same write speeds as RAM, but without its volatility. It doesn't loose it's data when the power is off. Ideal for fast HDDs they say.

Yeah, I think they're calling them "Hybrid drives" where they will have some fast static RAM built into a hard drive and store the most frequently accessed part of the drive in cache on the static RAM.

On an unrelated note, wouldnt it been cool to effectivly install a whole OS on RAM. That would be noticably quicker....

You used to be able to do that with ramdisks, but getting the files onto the ramdisk took more time than just booting from the disk. Sometimes you can force the OS to keep itself in RAM when it's loaded from disk (so the OS won't start swapping itself out when it needs memory), there's a setting for this in Tweak XP.

janstett

Sep 15, 2006, 09:07 AM

The server/desktop division with Windows - as with OS X - is one of marketing, not software. Windows "Workstation" and Windows "Server" use the same codebase.

True (today anyway; in the NT era they were indeed separate platforms though. Which brings me to my next point..)

No, that is not true, in fact it couldn't be more untrue. Now, the 95 family (95/98/ME) was a totally different codebase. But with the NT family (NT/2000/XP) the client and the server were identical, even identical in distributed code. In fact there was a big scandal years ago where someone discovered the registry setting where you could turn NT Workstation into NT Server. Back then all that was different was the number of outbound IP connections and possibly the number of CPUs supported. All they were trying to do with Workstation was prevent you from using it as a server (thus the outbound IP limit) and at some point they didn't give you full-blown IIS on Workstation. That's it.

janstett

Sep 15, 2006, 09:26 AM

And of course, NT started as a reimplementation of VMS for a failed Intel RISC CPU...

More pedantic details for those who are interested... :)

NT actually started as OS/2 3.0. Its lead architect was OS guru Dave Cutler, who is famous for architecting VMS for DEC, and naturally its design influenced NT. And the N-10 (Where "NT" comes from, "N" "T"en) Intel RISC processor was never intended to be a mainstream product; Dave Cutler insisted on the development team NOT using an X86 processor to make sure they would have no excuse to fall back on legacy code or thought. In fact, the N-10 build that was the default work environment for the team was never intended to leave the Microsoft campus. NT over its life has run on X86, DEC Alpha, MIPS, PowerPC, Itanium, and x64.

IBM and Microsoft worked together on OS/2 1.0 from 1985-1989. Much maligned, it did suck because it was targeted for the 286 not the 386, but it did break new ground -- preemptive multitasking and an advanced GUI (Presentation Manager). By 1989 they wanted to move on to something that would take advantage of the 386's 32-bit architecture, flat memory model, and virtual machine support. Simultaneously they started OS/2 2.0 (extend the current 16-bit code to a 16-32-bit hybrid) and OS/2 3.0 (a ground up, platform independent version). When Windows 3.0 took off in 1990, Microsoft had second thoughts and eventually broke with IBM. OS/2 3.0 became Windows NT -- in the first days of the split, NT still had OS/2 Presentation Manager APIs for it's GUI. They ripped it out and created Win32 APIs. That's also why to this day NT/2K/XP supported OS/2 command line applications, and there was also a little known GUI pack that would support OS/2 1.x GUI applications.

AidenShaw

Sep 15, 2006, 10:53 AM

More pedantic details for those who are interested... :).
Any description of the history of NT that doesn't say "Mica" and "Prism" is missing some major details ;) !

http://en.wikipedia.org/wiki/DEC_PRISM

ergle2

Sep 15, 2006, 01:25 PM

A cancelled Digital RISC CPU.

Although, some of the ideas for the cancelled CPU ended up in the Alpha chips.

NT was originally designed for the i860, which was codenamed the N-10 (hence NT).

Anything for Digital would have been while Cutler was at Digital, I imagine, rather than after he joined Microsoft.

brianus

Sep 15, 2006, 01:26 PM

No, that is not true, in fact it couldn't be more untrue. Now, the 95 family (95/98/ME) was a totally different codebase. But with the NT family (NT/2000/XP) the client and the server were identical, even identical in distributed code. In fact there was a big scandal years ago where someone discovered the registry setting where you could turn NT Workstation into NT Server. Back then all that was different was the number of outbound IP connections and possibly the number of CPUs supported. All they were trying to do with Workstation was prevent you from using it as a server (thus the outbound IP limit) and at some point they didn't give you full-blown IIS on Workstation. That's it.

Dude, how many times do I have to repeat myself before you myopic '90s-era IT geeks understand me? I was referring to the difference between Windows 9x and Windows NT. I neither knew, nor care, that there were different versions of NT itself. For. Christ's. Sake. I have said this three times now. Don't make me come over there.

On an unrelated note, wouldnt it been cool to effectivly install a whole OS on RAM. That would be noticably quicker....

I keep hearing about speculation that they'll start using NAND flash to help with startup times in laptops, things like that -- now, how would that work? Doesn't everything have to be on the boot volume? OS's seem to assume these days that the OS, programs and user directories are all going to be on one volume and you have to be kind of technically literate to do it differently..

ergle2

Sep 15, 2006, 01:50 PM

More pedantic details for those who are interested... :)

NT actually started as OS/2 3.0. Its lead architect was OS guru Dave Cutler, who is famous for architecting VMS for DEC, and naturally its design influenced NT. And the N-10 (Where "NT" comes from, "N" "T"en) Intel RISC processor was never intended to be a mainstream product; Dave Cutler insisted on the development team NOT using an X86 processor to make sure they would have no excuse to fall back on legacy code or thought. In fact, the N-10 build that was the default work environment for the team was never intended to leave the Microsoft campus. NT over its life has run on X86, DEC Alpha, MIPS, PowerPC, Itanium, and x64.

IBM and Microsoft worked together on OS/2 1.0 from 1985-1989. Much maligned, it did suck because it was targeted for the 286 not the 386, but it did break new ground -- preemptive multitasking and an advanced GUI (Presentation Manager). By 1989 they wanted to move on to something that would take advantage of the 386's 32-bit architecture, flat memory model, and virtual machine support. Simultaneously they started OS/2 2.0 (extend the current 16-bit code to a 16-32-bit hybrid) and OS/2 3.0 (a ground up, platform independent version). When Windows 3.0 took off in 1990, Microsoft had second thoughts and eventually broke with IBM. OS/2 3.0 became Windows NT -- in the first days of the split, NT still had OS/2 Presentation Manager APIs for it's GUI. They ripped it out and created Win32 APIs. That's also why to this day NT/2K/XP supported OS/2 command line applications, and there was also a little known GUI pack that would support OS/2 1.x GUI applications.

All very true, but beyond that -- if you've ever looked closely VMS and at NT, you'll notice, it's a lot more than just "influenced". The core design was pretty much identical -- the way I/O worked, its interrupt handling, the scheduler, and so on -- they're all practically carbon copies. Some of the names changed, but how things work under the hood hadn't. Since then it's evolved, of course, but you'd expect that.

Quite amusing, really... how a heavyweight enterprise-class OS of the 80's became the desktop of the 00's :)

Those that were around in the dim and distant will recall that VMS and Unix were two of the main competitors in many marketplaces in the 80's and early 90's... and today we have OS X, Linux, FreeBSD, Solaris, etc. vs XP, W2K3 Server and (soon) Vista -- kind of ironic, dontcha think? :)

Of course, there's a lot still running VMS to this very day. I don't think HP wants them to tho' -- they just sent all the support to India, apparently, to a team with relatively little experience...

ergle2

Sep 15, 2006, 01:54 PM

I keep hearing about speculation that they'll start using NAND flash to help with startup times in laptops, things like that -- now, how would that work? Doesn't everything have to be on the boot volume? OS's seem to assume these days that the OS, programs and user directories are all going to be on one volume and you have to be kind of technically literate to do it differently..

Intel's "Robson" technology.

It's just a cache of certain files in FLASH. It's trivial to have the system check there first and then the boot volume afterwards. Like everything else, those implementing it need to be technically literate, but once its done, it's done. Users don't need to know what's going on.

Vista already has some feature that allows caching etc. to any flash devices connected to the system, btw.

ergle2

Sep 15, 2006, 02:08 PM

On an unrelated note, wouldnt it been cool to effectivly install a whole OS on RAM. That would be noticably quicker....

The OS would be faster but unless you had tons of RAM, the Apps ... :)

Modern OSes use RAM not used by apps to cache recently used files/data, since it makes more sense to keep around stuff the system mind need again. Most OS files aren't needed (just look at the size of the OS itself on any system!).

Of course, back in my Amiga days, pretty much all the OS was running from ROM/RAM, and it had pre-emptive multitasking but no VM system. As a result, it was incredibly snappy to use, despite being a 7.14MHz 68K. I've occasionally seen real Amigas since then and I'm always impressed by how "fast" it feels, even if the system itself seems rather primative by modern standards.

I imagine the early Macs were somewhat similar in this regard, but I didn't use one properly til the early 90's, by which time I was more interested in Unix, VMS, etc.

pavetheforest

Sep 15, 2006, 10:44 PM

Dude I'm going to sell my dell.

Demoman

Sep 15, 2006, 11:52 PM

Uh, last time I checked, Windows can take advantage of multiple cores just fine. Do you think that multithreading is some Black Magic that only MacOS can do? Hell, standard Linux from kernel.org can use 512 cores as we speak!

Related to this: Maybe not 512-way SMP, but here (http://www.linux-mips.org/wiki/IP27_boot_messages) is what it looks like when Linux boots on 128-way SGI Origin supercomputer. Note, the kernel that is booting is 2.4.1, which was released in early 2001. Things have progressed A LOT since those day.

OS X works with quad core == "Ahead of technology curve"... puhleeze!

Windows works just fine with dual-core. It really does. To Wndows, dual-core is more or less similar to typical SMP, and Windows has supported SMP since Windows NT!

Any reason why it wouldn't work? And did you even read the Anandtech-article? They conducted their benchmarks in Windows XP! So it obviously DID work with four cores! And it DID show substantial improvement in performance in real-life apps! Sheesh! Dial tone that fanboysihness a bit, dude.

I think the same applies to you, Bill. You seem to be here to act as a Microsoft evangelist.

AidenShaw

Sep 16, 2006, 01:09 AM

Dude I'm going to sell my dell.
And buy a new Dell with these same chips and features ;)

janstett

Sep 16, 2006, 11:10 AM

Any description of the history of NT that doesn't say "Mica" and "Prism" is missing some major details ;) !

Well, come on! I wrote a synopsis that was already too lengthy. I felt it sufficient to say that Dave Cutler's life at DEC gave him OS Guru status and left it at that. I didn't mention Gordon Letwin either. On either point it's rather like mentioning Brian Kernighan and Rob Pike in a history of OSX -- technically accurate but of marginal relevance.

janstett

Sep 16, 2006, 11:14 AM

Dude, how many times do I have to repeat myself before you myopic '90s-era IT geeks understand me? I was referring to the difference between Windows 9x and Windows NT. I neither knew, nor care, that there were different versions of NT itself. For. Christ's. Sake. I have said this three times now. Don't make me come over there.

Well then, if you are so consistantly misinterpreted, have you ever stopped to think you should CLARIFY yourself, or that you must not be communicating your point clearly? The truth is Microsoft has dealt with two simultaneous families of operating systems from 1987-2003, and the survivor is NT/2K/XP, and it was always the better of the two operating system families that geeks like us would be concerned with, so naturally that's the one most people think of when projecting back in history.

JeffDM

Sep 16, 2006, 04:50 PM

One app would be iTunes. I noticed iTunes was running 14 threads last night. Any time you have a multithreaded application or are running multiple single thread aplications more cores can help.

iTunes is generally so low-impact that it could be single threaded and you probably wouldn't notice. If the main thread is bogged down, I still get the spinny color disc with iTunes on occasion. It seems to do this sometimes when I sync an iPod, iTunes sometimes won't let me do anything else.

An eight-core system should be able to eight single threaded programs running at 100% of one CPU without issue. What I hope is that more programs that need the processing power can use the full power of more than one CPU so you don't need to multitask so heavily to take advantage of the power available.

JeffDM

Sep 16, 2006, 04:56 PM

As for using a Dell, sure they could've used that. Would Windows use the extra 4 cores? Highly doubtful. Microsoft has sketchy 64 bit support let alone dual core support; I'm not saying "impossible" but I haven't read jack squat about any version of Windows working well with quad cores.

Bad dual core support? Citations please. I think this is a case where a Mac fan is simply speaking out of ignorance of their "enemy" platform.

I've been using dual processor Windows computers for a few years now and it works fine, I can't imagine dual core being any different. For quad core, I think THG showed that a Kentsfield showed significant performance benefits over a Conroe for many Windows programs. The media encoders showed very nearly a 2x performance difference.

JeffDM

Sep 16, 2006, 05:39 PM

You are right. However, you try to tell consumers "Well we are moving to 2.4Ghz chips" after you just had 2.66Ghz and 3.0Ghz chips. It isnt going to work.

If today, Dell decided to move there whole line back to 1Ghz processors, nobody would buy. Unfortunetly the Ghz myth is a strong as its ever been. Taking a step backward is not an option.

It's already happened, just not in as a melodramatic way as you suggest (back to 1GHz? geez). AMD took a small step back, Hz wise when they introduced dual core, though it still advanced their "+" processor ratings I suppose that few noticed the actual clock reduction. Intel took a major step back Hz wise between Netburst and Core 2. The 5000 and 5100 series Xeon CPUs demonstrate this, you can get a Dell precision 690 with 3.73GHz Netburst based chips or the same 690 with 3.0GHz Core2 based chips.

So I don't think that a quad core Xeon running at 2.66GHz is going to be hurt too much in comparison to a dual core 3.0GHz, it's still a much more powerful processor.

Didn't you get the memo, PowerPC is dead. WTF does that have to do with anything? Do you just have this Pavlovian response to the word "Hyperthreading"?

PPC isn't dead, it's just not in new desktops anymore. IBM is making them (or at least co-designed them) for all the next generation game consoles and a lot of huge supercomputers.

zero2dash

Sep 18, 2006, 02:44 PM

Plenty of people ran NT on their desktops.

Admission of your mistakes is a good step in becoming a better person.

Key word being DESKTOPS.
MP machines were server based long before they were included in desktops. I'd like to see where people had dual Xeon based DESKTOPS 'cause I've never seen it. It's not impossible but it's also not a good cost-based answer either. :p

The server/desktop division with Windows - as with OS X - is one of marketing, not software. Windows "Workstation" and Windows "Server" use the same codebase.

I never said otherwise.
The hardware they run on is where it differentiates.
Most people/corporations run server-based OS on servers and workstation-based OS on desktops (or "workstations" in the business world). It's not impossible to run a server OS on a desktop or a workstation OS on a server but it is incredibly stupid.

Well, if you can't find evidence of Windows running on well on machine with >2 processors, or of the significant low-level changes Microsoft have made to ensure it does, you aren't looking very hard.

Bad dual core support? Citations please. I think this is a case where a Mac fan is simply speaking out of ignorance of their "enemy" platform.

I erronously bundled in "dual core" with "sketchy 64-bit support". Don't know why. From what I hear, 64-bit support in XP64 is sketchy because of device driver issues (and drivers not being natively 64-bit). I don't have any true 'dual core' systems myself but my P4 3.0C HT works fine in XP Pro. I apologize for lumping in "dual core" in.

Similarly, if you're one of the "Vista is just XP with a fancy skin" crowd, you've obviously not done much research. The changes in Vista are on par with the scale of changes Apple made to NeXT to get OS X.

User Account Protection is a big change. I've seen the list of "new features" and it doesn't do anything for me. UAP is nice...it's just really late. I'm sure there's changes "under the hood" like the ones implemented in XP sp2 to prevent buffer/stack overflows, etc. and I'm sure that's what you're referring to.

I think people who say stuff like that are exhibiting a syndrome common to Mac folk who've never spent any time in the PC world -- they take negative comments they remember regarding versions of Windows or the PC experience from about 5 years back and assume they apply to today. XP, for example, really was for the most part a window-dressing of Windows 2000, but that is not the case for Vista. You see similar statements regarding "blue screens of death", overall system stability, etc, which suggest they haven't seen or used a PC since the late 90s/early 00's.

So - are you inferring that Windows 2000 or Windows XP never blue screen? Because (if you are) that's a load of crap. I've seen blue screens in both OS's. Granted it's usually tied to hardware only, but it still happens. I've had an external USB drive blue screen in XP every time I turned it on, tried on 3 XP computers. Hardware fault, no doubt. Lately my HP Laptop dvd drive has been causing XP Pro to blue screen every other time I insert a dvd-r. Again - hardware fault.

Otherwise are both OS's stable? Damn straight. But problems do occur and I hope you're not suggesting otherwise. No OS is without its flaws.

Trekkie

Sep 18, 2006, 03:19 PM

zero2dash

Sep 18, 2006, 03:26 PM

The Thinkpad X40 I'm typing from Bluescreened on me no longer than three weeks ago. My crime? coming out of suspend mode.

Windows Crashes.

Believe it or not, Mac OS X can crash too. While it is prettier, it's still a crash.

Pretty funny reading the last few pages, thanks for the laughs.

I'll never forget at my old job (Kinkos) when our dual G4 running Panther had that system crash screen come up that is gray and basically says "your FUBAR'D" in like 8 languages...we were all stunned. :D Good times...never thought I'd see a bad crash like that in OSX. Or back in 2000 when our workstations were all running Windows 95b - I lost track of how many blue screens we'd get in a day. Man Win95 was garbage. :p Big for it's time - utter garbage now.

ergle2

Sep 19, 2006, 12:57 AM

Key word being DESKTOPS.
MP machines were server based long before they were included in desktops. I'd like to see where people had dual Xeon based DESKTOPS 'cause I've never seen it. It's not impossible but it's also not a good cost-based answer either. :p

I've known many people with multi-processor machines on their desktop, with a variety of processor families -- including Intel -- going way back over the best part of a decade. If your requirements include applications that can make use of it, it can make sense. Time is money and all that.

One market I'm aware of is the fluid dynamics market, which pretty much eats all the processor time you can throw at it.

I even had an x86 dual CPU machine at home back in 1999... I still have it, it's just not that fast any more...

Of course, these days everyone and his dog has dual-core, pretty much...

Edit: 2nd para clarified

Evangelion

Sep 19, 2006, 07:17 AM

Key word being DESKTOPS.

Again: NT was widely used on desktops. Maybe not by your Average Joe, but LOTS of people used it on the desktop. I used NT-workstation back when I studied, my friend used NT on his PC, lots and lots of companies ran NT, the list goes on. Hell, there were propably an order of magnitude more NT-desktops out there that there were Macs of any type!

I still don't know personally anyone who uses OS X. Does that mean that no-one uses it?

MP machines were server based long before they were included in desktops. I'd like to see where people had dual Xeon based DESKTOPS 'cause I've never seen it.

There were plenty of people running SMP-systems. I personally knew two guys who had SMP-PC's. Just because you haven't seen anyone use one, does not mean that they weren't there.

padré

Sep 19, 2006, 11:18 AM

ergle2

Sep 19, 2006, 01:14 PM

so... after reading here for a while i got a question, its kinda stupid, i'm good at that,
first off, i was doubting between the 24" and the macpro so i disided that for my needs i should realy go with a macpro, but know that i'm hearing things about this 8 core macpro, i'm realy doubting about ordering my quad macpro this month,

has anybody got an idea of how long it would be before apple launches " a macpro octo " :confused:

thx for your time :)

It's Apple. No-one has any idea when they'll do anything. :)

It could be as soon as January, could be a lot later -- but I seriously doubt it'd be at the same price as a quad is now. I'd figure on a fairly major premium. It wouldn't surprise me if the OEM price of processors was in the $1200-1500 range alone (current 3.0GHz 5160's are around $900) for a lower clock-speed version.

Which is fastest will very much depend upon how well your specific applications scale -- fewer, faster cores can often bear more slower cores, and scaling isn't linear -- traditional thought on SMP was that the first extra core you add adds 80-90% to the speed (for fully-threaded apps, obv.), the second adds about 60-70%, the third about 40%, and so-on... diminishing returns. This will be more so because each chip has a finite amount of bandwidth that is shared between all the cores -- more cores = more contention for the available bandwidth.

Of course, the Mac Pro CPUs are socketed, so you can always go Octo at a later date if you so choose...

padré

Sep 19, 2006, 01:39 PM

ergle2

Sep 19, 2006, 01:56 PM

thx for your reply,

i will go for the mac pro quad know (i'm updating my home computer wich is a G3, but i'm used to work on a dual G5 for my projects) and yeah i will allways be able then to update later, but how about ram, when DDR3 comes out, i read that its going to replace FB-DIMMs so will that be upgradeble too???
cause these FB-DIMMS are so ********** expensive :) thx

DDR and FB aren't mutually exclusive technologies. Current FB-DIMMs use DDR2 chips. Newer FB-DIMMs will use DDR3 chips but due to the way FB-DIMMs work -- the buffer and control hardware essentially sit between the memory and the MCH -- you should still be able to use them in a Mac Pro. Note, should, not will.

You won't be able to use regular DDR3 DIMMs just like you can't use regular DDR2 DIMMs tho'. That's something that'd require a major design change to the motherboard.

So, memory will still be damn' expensive compared to that found in the cheap desktops...

padré

Sep 19, 2006, 02:12 PM

then i will have to work just a little bit harder i'm afraid ;)

i'll start with the 1 gig ram, maybe 2, and later i will be upgrading,it can't stay expansive forever.

thx again for your reply

ergle2

Sep 19, 2006, 09:38 PM

then i will have to work just a little bit harder i'm afraid ;)

i'll start with the 1 gig ram, maybe 2, and later i will be upgrading,it can't stay expansive forever.

thx again for your reply

DailyTech (http://www.dailytech.com/article.aspx?newsid=4217) has a mention of the Core 2 Quadro processors.

Pricing mentioned was a little lower than I expected, but it's processors in the Conroe line rather than the Xeon. Having said that, the 3GHz Xeon is slightly cheaper than the 2.93GHz Conroe.

As expected, the highest rated speed mentioned is 2.67Ghz.

technicolor

Sep 19, 2006, 09:49 PM

DailyTech (http://www.dailytech.com/article.aspx?newsid=4217) has a mention of the Core 2 Quadro processors.

Pricing mentioned was a little lower than I expected, but it's processors in the Conroe line rather than the Xeon. Having said that, the 3GHz Xeon is slightly cheaper than the 2.93GHz Conroe.

As expected, the highest rated speed mentioned is 2.67Ghz. This intel crap updates far too frequently...ugh

:mad:

Silentwave

Sep 19, 2006, 10:16 PM

well they're working through a pretty active period right now, what with the major architecture change. It'll sorta taper out a bit in a while, and so the next 'big thing' besides more cores will be 45nm, followed by the Common System Interface for the Xeons in '08.

ergle2

Sep 19, 2006, 10:56 PM

This intel crap updates far too frequently...ugh

:mad:

Speaking personally, I don't see a problem with it.

They've always had updates roughly every 3-6 months, of one kind or another.

technicolor

Sep 19, 2006, 11:05 PM

Speaking personally, I don't see a problem with it.

They've always had updates roughly every 3-6 months, of one kind or another.
Yeah and they werent in Macs, so I didnt care.

ergle2

Sep 19, 2006, 11:08 PM

Yeah and they werent in Macs, so I didnt care.

Why do you care?

technicolor

Sep 19, 2006, 11:15 PM

Why do you care?
Why shouldnt I?

ergle2

Sep 19, 2006, 11:17 PM

Why shouldnt I?

Why should it bother you that new processors come out?

technicolor

Sep 19, 2006, 11:35 PM

Why should it bother you that new processors come out?
why shouldnt it?

ergle2

Sep 19, 2006, 11:43 PM

why shouldnt it?

Ah, a mature, intelligent, well reasoned reply.

THX1139

Sep 20, 2006, 03:39 AM

Ah, a mature, intelligent, well reasoned reply.

What did you expect? Didn't you look at his avatar? Cool, Homer is a member of Macrumors. :D

ksz

Sep 20, 2006, 04:29 AM

What did you expect? Didn't you look at his avatar? Cool, Homer is a member of Macrumors. :D
...except that he's a she...a demi-goddess. Frequent updates are a good thing. I would not want to stop the march of progress just so I could personally feel better about a little money I spent.

technicolor

Sep 20, 2006, 05:12 AM

Ah, a mature, intelligent, well reasoned reply.
No, one that just ignores you and your inquiries because it was already clear where you were coming from..thus I feel no obligation to engage you in my thought process and your self important questioning. Has nothing to do with my maturity, and everything to do with my lack of caring about you or your opinion.

ergle2

Sep 20, 2006, 02:50 PM

No, one that just ignores you and your inquiries because it was already clear where you were coming from..thus I feel no obligation to engage you in my thought process and your self important questioning. Has nothing to do with my maturity, and everything to do with my lack of caring about you or your opinion.

Ah, a response that is pompous, imperious, petulant, inconsistent AND incorrect -- you didn't ignore me, after all, you were simply obnoxious and evasive with a serious of ridiculous replies.

Nice projection with the "self-important", though.

Where I was coming from was purely from a sense of curiosity as to why someone would apparently feel that technology can advance too quickly. It's not unobvious, I grant you, but a rather unobjectionable query to my mind.

Interestingly, it would appear you do care enough about my opinion - or perhaps that of other forum denizens? - to respond in a way you apparently consider to be "clever". Feel free to ignore me -- really, I can take it! I mean, I'm sure I'll probably suffer a couple of seconds of heartbreak sometime around, say, 2020, but I'll survive.

Oh and "thought process" - ROTFL - lovely!

patrick0brien

Sep 20, 2006, 03:10 PM

Umm. What happened in here?

Can we reurn to some common respect please? This spat isn't constructive.

ergle2

Sep 20, 2006, 03:23 PM

...except that he's a she...a demi-goddess.

She certainly has the attitude of one.

Frequent updates are a good thing. I would not want to stop the march of progress just so I could personally feel better about a little money I spent.

The only real downside I see is that Intel Macs are unlikely to hold their value anywhere near as well as the PPC line did due to the quicker changes we'll see now.

I keep systems til they fall apart, pretty much, but there's quite a few on the various forums who say they always buy and sell 2-3 years later to upgrade.

ergle2

Sep 20, 2006, 04:51 PM

Umm. What happened in here?

Can we reurn to some common respect please? This spat isn't constructive.

True enough.

I ... well, I won't go there, too likely to throw more fuel on the fire.

I'll drop it if she does, fair enough?

brianus

Sep 20, 2006, 05:07 PM

So - are you inferring that Windows 2000 or Windows XP never blue screen? Because (if you are) that's a load of crap. I've seen blue screens in both OS's. Granted it's usually tied to hardware only, but it still happens. I've had an external USB drive blue screen in XP every time I turned it on, tried on 3 XP computers. Hardware fault, no doubt. Lately my HP Laptop dvd drive has been causing XP Pro to blue screen every other time I insert a dvd-r. Again - hardware fault.

Otherwise are both OS's stable? Damn straight. But problems do occur and I hope you're not suggesting otherwise. No OS is without its flaws.

Huh? When did I say they never, ever experience any crashes whatsoever? Good god, I have never seen such a collection of mind-bendingly literal-minded people in one thread. Yikes. No idiot would ever say they never ever crash. As was painfully obvious, I was comparing Mac users' perceptions of older Windows OS's to the more recent ones and saying their impressions were inaccurate. I've been dealing with OS X kernel panics and CarbonLib issues all day, but I would never suggest things are as bad as in the OS 8 days when you'd get that little "bomb" at the system would shut down.

It's already happened, just not in as a melodramatic way as you suggest (back to 1GHz? geez). AMD took a small step back, Hz wise when they introduced dual core, though it still advanced their "+" processor ratings I suppose that few noticed the actual clock reduction. Intel took a major step back Hz wise between Netburst and Core 2. The 5000 and 5100 series Xeon CPUs demonstrate this, you can get a Dell precision 690 with 3.73GHz Netburst based chips or the same 690 with 3.0GHz Core2 based chips.

One thing I've noticed is that store ads no longer quote GHz like they used to, but rather processor model numbers. Makes sense: most people will not bother to investigate further, but if they did see the GHz numbers of Pentiums on the same sale ad as those of Core 2's, they might not be so hot on the latter. And please, everyone for the love of god, do not treat me to 5 replies in which you remonstrate me for not getting that the Core 2's are actually faster - I GET IT.

mdntcallr

Sep 20, 2006, 05:10 PM

hey ill be happy as apple keeps the mac pro on the cutting edge, but anything to be able to bring the ram cost down would be awesome.

this buffered ram is expensive.

all the other ram out there is getting cheaper, but not this stuff they want in the mac pro.

ksz

Sep 20, 2006, 05:11 PM

The only real downside I see is that Intel Macs are unlikely to hold their value anywhere near as well as the PPC line did due to the quicker changes we'll see now.

I keep systems til they fall apart, pretty much, but there's quite a few on the various forums who say they always buy and sell 2-3 years later to upgrade.
I should have been more thorough in my previous reply. What I really like about these frequent updates are the following:

1. The motherboard has socketed processors (except for the laptops).

2. Even though Intel is updating processors every 6 months or so, the motherboard and chipset seem to support the next processor version.

Yonah can be replaced with Merom.
Woodcrest can be replaced with Clovertown.

Your computer does not become obsolete in 6 months. Instead, it gains new life if you decide that you need the new processor.

Every 12 to 18 months or so a new chipset may become necessary. Only then does your computer lose the upgrade potential. If you buy Merom, you may not be able to upgrade to the next processor. Likewise if you buy Clovertown. New chipsets will be required beyond Merom and Clovertown.

In any event, this is based on trailing history of just 1 year. Future events may unfold differently.

ergle2

Sep 20, 2006, 07:44 PM

I should have been more thorough in my previous reply. What I really like about these frequent updates are the following:

1. The motherboard has socketed processors (except for the laptops).

Yeah, an upgradable processor socket is a wonderful thing :)
It's a shame the laptops are soldered, but it makes sense given the design...

2. Even though Intel is updating processors every 6 months or so, the motherboard and chipset seem to support the next processor version.

Yonah can be replaced with Merom.
Woodcrest can be replaced with Clovertown.

Your computer does not become obsolete in 6 months. Instead, it gains new life if you decide that you need the new processor.

Every 12 to 18 months or so a new chipset may become necessary. Only then does your computer lose the upgrade potential. If you buy Merom, you may not be able to upgrade to the next processor. Likewise if you buy Clovertown. New chipsets will be required beyond Merom and Clovertown.

In any event, this is based on trailing history of just 1 year. Future events may unfold differently.

Yeah -- tho' some of this might not please some due to philosophy.

Bear in mind part of the Mac philosophy from the start was "no user servicable parts inside" -- think of it as the computing equivalent of a toaster, in a sense. Jobs and Raskin were both proponents of that concept, and it lives in in some of the userbase.

I suspect that part of the userbase would prefer being able to sell an old system and buy a new one.

Now, that's not my worldview, but it's definitely out there.

Going back, often newer processors are release, at least initially, in multiple forms of package. Take the Pentium-4, which appeared for some versions as both a S478 and S775 (I think? or was there one inbetween?) chip. So even when there's a new chipset, it's not always required, it'll just give you some whizz-band new features.

With Merom, you're likely right, since that's part of the mobile line, and Intel sells the mobile line by platform (well, you can get it OEM too, but it's a lot cheaper if you just buy the platform).

ksz

Sep 20, 2006, 08:44 PM

Jobs and Raskin were both proponents of that concept, and it lives in in some of the userbase.
I had the pleasure of meeting Jef Raskin at his home in Pacifica a year before he passed away. He loved to play musical instruments and performed a short recital on his piano. Later that evening, after showing his Apple I in a wooden box, he encouraged me to read his book The Humane Interface and let him know what I thought about it. Sadly, I wasn't able to do that in time. But the conversation we had made it clear that he was not a fan of Steve Jobs. They both had strong opinions on various aspects of UI design. Even though I rather like OS X, Raskin politely argued against the inefficiencies of that design.

It was some time ago and I don't remember all the details from that night, but Raskin, I think, was more scientific in his approach. He preferred to study user response rates, time-to-decision, amount of eye movement, amount of pointer movement, number of mouse clicks, and various other factors that might contribute to 'dead' or wasted time.

ergle2

Sep 21, 2006, 04:17 PM

I had the pleasure of meeting Jef Raskin at his home in Pacifica a year before he passed away. He loved to play musical instruments and performed a short recital on his piano. Later that evening, after showing his Apple I in a wooden box, he encouraged me to read his book The Humane Interface and let him know what I thought about it. Sadly, I wasn't able to do that in time. But the conversation we had made it clear that he was not a fan of Steve Jobs. They both had strong opinions on various aspects of UI design. Even though I rather like OS X, Raskin politely argued against the inefficiencies of that design.

It was some time ago and I don't remember all the details from that night, but Raskin, I think, was more scientific in his approach. He preferred to study user response rates, time-to-decision, amount of eye movement, amount of pointer movement, number of mouse clicks, and various other factors that might contribute to 'dead' or wasted time.

It doesn't surprise me he was no fan of Jobs, especially given the history of the original Mac. From people I know who've worked with Jobs, he's not always easy to get along with.

From what I've read, the Mac was fundamentally quite different from Raskin's original vision after Jobs took over the project, though some of his ideas were obviously incorporated into it. (I believe Raskin wanted to go with the cheaper but obviously slower 6809).

His book's been one I've meant to track down for some time now. You know how it is, so many things to do/see...

The Archy interface modelled on his concepts is quite interesting, too.

I was sorry when we lost Jef, I feel he was one of those people striving to make the world a better place.

What did you think of The Humane Design?

zoran

Oct 14, 2006, 03:50 PM

well its said that clovertown will be here early 2006, is that early/late november or early/late december, any new rumors regarding this subject?

AppliedVisual

Oct 14, 2006, 05:02 PM

Intel is supposed to start shipping Clovertown CPUs into the retail channel by mid-november. We should see systems based on them by then too -- HP is claiming to have their first Clovertown workstations available on the 15th, so only a month away. I bet we'll see the 8-core Mac Pro systems by the end of November.

zoran

Oct 15, 2006, 01:45 PM

HP is claiming to have their first Clovertown workstations available on the 15th, so only a month away. I bet we'll see the 8-core Mac Pro systems by the end of November.
Why would Apple show their Clovertown workstations after HP and not simultaneusly with HP?

AppliedVisual

Oct 15, 2006, 01:59 PM

Why would Apple show their Clovertown workstations after HP and not simultaneusly with HP?

Because that's usually how it works. :confused:

HP is Intel's main launch partner for the quad-core Xeon and I think they have secured the first of the major shipments.

l4t13

Oct 15, 2006, 02:02 PM

still very good news for high performance users

zoran

Oct 15, 2006, 02:05 PM

How long did macPro delay compared to HPs similar workstation?

ruutiveijari

Oct 15, 2006, 02:06 PM

Why would Apple show their Clovertown workstations after HP and not simultaneusly with HP?
Because HP is a much bigger company with much bigger sales volume and probably gets all the new processors before Apple does.

AppliedVisual

Oct 15, 2006, 02:08 PM

How long did macPro delay compared to HPs similar workstation?

HP, Dell and IBM all had dual Core 2 Xeon workstation systems available 2~3 weeks ahead of Apple's Mac Pro release. Apple has yet to release their new Xserve. HP, Dell, IBM and others have had dual (and even some quad CPU configurations) of Core 2 Xeon 1U servers and blades available for months now...

Oct 15, 2006, 03:15 PM

... hmmm ... i just ordered a mac pro quad 3ghz ... 8 cores would be somehow nicer ;)
.a

zoran

Oct 15, 2006, 03:55 PM

... hmmm ... i just ordered a mac pro quad 3ghz ... 8 cores would be somehow nicer ;)
.a
cancel the order while its not too late! :)

AppliedVisual

Oct 15, 2006, 04:47 PM

... hmmm ... i just ordered a mac pro quad 3ghz ... 8 cores would be somehow nicer ;)
.a

The 8-core Mac Pro @ 2.33GHz should be about the same price as the quad-core 2.66GHz. Theoretically, the 8-core 2.66GHz should be about the same price as what you just ordered.

Before you seriously consider canceling, just be sure that your workflow can benefit from the various CPU cores. Very few applications can take advantage of dual-core CPUs, let alone quad-core. In most situations, you need to be running various instances or multiple apps at once that can handle 2 or more threads to benefit from these newer multi-core systems. If you do any 3D animation or heavy rendering, scientific computing, visualization, massive database management/development, etc... You may be a candidate. Depending on your requirements, a quad-core 3GHz may still be the best performing system for you.

ddekker

Oct 22, 2006, 02:21 PM

AppliedVisual

Oct 22, 2006, 04:14 PM

I heard Leo Laporte talking about this on his KFI podcast... exciting... one question... how many softwares take advantage of multi cores? I understand that the OS can deal with it for multi tasking, but how many programs multi thread?

DD

Unfortunately not many multithreaded apps - yet. For a long time most of the multi-threaded apps were just a select few pro level things. 3D/Visualization software, CAD, database systems, etc.. Those of us who had multiprocessor systems bought them because we had a specific software in mind or group of software applications that could take advantage of multiple processors. As current CPU manufacturing processes started hitting a wall right around the 3GHz mark, chip makers started to transition to multiple CPU cores to boost power - makes sense. Software developers have been lazy for years, just riding the wave of ever-increasing MHz. Now the multi-core CPUs are here and the software is behind as many applications need to have serious re-writes done in order to take advantage of multiple processors. Intel tried to get a jump on this with their HT (Hyper Threading) implementation that essentially simulated dual-cores on a CPU by way of two virtual CPUs. Software developers didn't exactly jump on this and warm up to it. But I also don't think the software industry truly believed that CPUs would go multi-core on a mass scale so fast... Intel and AMD both said they would, don't know why the software industry doubted. Intel and AMD are uncommonly good about telling the truth about upcoming products. Both will be shipping quad-core CPU offerings by year's end.

SMM

Oct 22, 2006, 05:15 PM

HP, Dell and IBM all had dual Core 2 Xeon workstation systems available 2~3 weeks ahead of Apple's Mac Pro release. Apple has yet to release their new Xserve. HP, Dell, IBM and others have had dual (and even some quad CPU configurations) of Core 2 Xeon 1U servers and blades available for months now...

It makes sense for Apple not to offer the 5000 series with their 1066 Max FSB. They were pretty quick to market for the MP with the 5100 Woody's release. It is curious that the XServe is not released yet. Perhaps, they are waiting for the Clovertown? I cannot imagine any unusual technical issues for not making the 5100 available by now.

janstett

Oct 23, 2006, 12:44 PM

Unfortunately not many multithreaded apps - yet. For a long time most of the multi-threaded apps were just a select few pro level things. 3D/Visualization software, CAD, database systems, etc.. Those of us who had multiprocessor systems bought them because we had a specific software in mind or group of software applications that could take advantage of multiple processors. As current CPU manufacturing processes started hitting a wall right around the 3GHz mark, chip makers started to transition to multiple CPU cores to boost power - makes sense. Software developers have been lazy for years, just riding the wave of ever-increasing MHz. Now the multi-core CPUs are here and the software is behind as many applications need to have serious re-writes done in order to take advantage of multiple processors. Intel tried to get a jump on this with their HT (Hyper Threading) implementation that essentially simulated dual-cores on a CPU by way of two virtual CPUs. Software developers didn't exactly jump on this and warm up to it. But I also don't think the software industry truly believed that CPUs would go multi-core on a mass scale so fast... Intel and AMD both said they would, don't know why the software industry doubted. Intel and AMD are uncommonly good about telling the truth about upcoming products. Both will be shipping quad-core CPU offerings by year's end.

What you're saying isn't entirely true and may give some people the wrong idea.

First, a multicore system is helpful when running multiple CPU-intensive single-threaded applications on a proper multitasking operating system. For example, right now I'm ripping CDs on iTunes. One processor gets used a lot and the other three are idle. I could be using this CPU power for another app.

The reality is that to take advantage of multiple cores, you had to take advantage of threads. Now, I was doing this in my programs with OS/2 back in 1992. I've been writing multithreaded apps my entire career. But writing a threaded application requires thought and work, so naturally many programmers are lazy and avoid threads. Plus it is harder to debug and synchronize a multithreaded application. Windows and Linux people have been doing this since the stone age, and Windows/Linux have had usable multiprocessor systems for more than a decade (it didn't start with Hyperthreading). I had a dual-processor 486 running NT 3.5 circa 1995. It's just been more of an optional "cool trick" to write threaded applications that the timid programmer avoids. Also it's worth noting that it's possible to go overboard with excessive threading and that leads to problems (context switching, thrashing, synchronization, etc).

Now, on the Mac side, OS 9 and below couldn't properly support SMP and it required a hacked version of the OS and a special version of the application. So the history of the Mac world has been, until recently with OSX, to avoid threading and multiprocessing unless specially called for and then at great pain to do so.

So it goes back to getting developers to write threaded applications. Now that we're getting to 4 and 8 core systems, it also presents a problem.

The classic reason to create a thread is to prevent the GUI from locking up while processing. Let's say I write a GUI program that has a calculation that takes 20 seconds. If I do it the lazy way, the GUI will lock up for 20 seconds because it can't process window messages during that time. If I write a thread, the calculation can take place there and leave the GUI thread able to process messages and keep the application alive, and then signal the other thread when it's done.

But now with more than 4 or 8 cores, the problem is how do you break up the work? 9 women can't have a baby in a month. So if your process is still serialized, you still have to wait with 1 processor doing all the work and the others sitting idle. For example, if you encode a video, it is a very serialized process. I hear some work has been done to simultaneously encode macroblocks in parallel, but getting 8 processors to chew on a single video is an interesting problem.