Intel's Quad Core CPU Reviewed

Gr8Apes is one of many to let us know that Tom's Hardware Guide has posted a review of Intel's new Kentsfield quad core processor. From the article: "Even expert opinions are deeply divided, ranging from 'more cores are absolutely necessary' to 'why do I need something more than my five-year-old PC system?' Although the Core 2 quad-core processors are not expected to hit retail channels before October, Tom's Hardware Guide had the opportunity to examine several Core 2 Quadro models in the test labs. We would like to make it clear that these samples were not provided by Intel."

Comments

Same old dilemma, new format.

[sdaemon]

Some applications will make use of it, some won't. More cores is pretty much the same as more CPUs.

Re:Same old dilemma, new format.

[fm6]

Gosh darn, you ended the flame war before it had a chance to start. Shame on you!

Re:Same old dilemma, new format.

[sdaemon]

My apologies. The jury will disregard my previous comment. Proceed with flames...

Re:Same old dilemma, new format.

[stonertom]

You speak from an educated perspective, think how many people you heard bragging about there new HT system without knowing what those letters are supposed to do. 4 cores may not help in any current game, especially no more than 2 (think background tasks) however think of the marketing deptartment.

Re:Same old dilemma, new format.

[Anonymous Coward]

Except that pretty much everyone is going to have multicore eventually while more CPUs has never been terribly usual.
Believe or not, this fact will affect the number of applications making use of it.

Re:Same old dilemma, new format.

[Redlazer]

Gosh darn, you ended the flame war before it had a chance to start. Shame on you!

Its ok, he must be new here.

-Red

Re:Same old dilemma, new format.

[bangenge]

so does this mean that the parent is gonna be modded flamebait?

Re:Same old dilemma, new format.

[Kjella]

Some applications will make use of it, some won't. More cores is pretty much the same as more CPUs.

And it was quite obvious which, parallelizable processes like media encoding saw 80% improvement. Some of the rendering tests saw very impressive improvement as well, but that's not really new and that's why those people have been paying for SMP setups in the past too. The gaming improvements were at the moment none, unlike the Core Duo which spanked the PIVs. They're really stretching at straws saying "it's a must for HDTV editing" - indeed, but how many are editing/compressing HDTV on their media PC? HDV home video is still rare as hell, HD-DVD / Blu-Ray can't be recompressed because of DRM. So maybe the ones with MythTV and OTA HDTV (or the few HDTV channels in Europe), but that's pretty much it. For what it's worth, I think the recommendation is to use hardware compression (or direct stream rips) anyway, and I imagine it won't take that long to get hardware cards which'll put out HDTV in one of the better HD-DVD/Blu-Ray formats, either VC-1 or H.264. At that point, who needs it? Maybe if they get the power consumption down to 2x Core 2 Duo (was that confusing or what) it'll be a nice way to pack them even tighter in servers. Beyond that... what?

Re:Same old dilemma, new format.

[SharpFang]

One core dedicated to enemy AI.
One core for physics calculations
One core for the game itself.
One core for OS, daemons and to bring them all and in the darkness bind them

Sure nothing for any -current- game.

Re:Same old dilemma, new format.

[ravenshrike]

So now, unless you code for 1,2, and 4 cores in the game you're making, you limit it unless the person has enough cores? Might work in 7-10 years, but not until then, and it's a lot easier to take code designed for one core and offload enough of it to optimize stuff for 2 cores but it's a whole 'nother ballgame to do so for 4 cores. Especially if you still have to design for 1 and 2 cores.

Re:Same old dilemma, new format.

[Amouth]

HT is an amazing tech. I know it is simplistic and in reality anything that takes advantage of SMP runs slower on HT than with out it.. but the nice thing about HT is that you can always get a second thread command through when a thread locks up and stops responding..

if you have a single threaded program that locks up you can always get commands to kill that thread via the second thread path via HT..

I always run HT on any cpu I have that supports it. mainly for the ability to always kill threads that might get locked up..

when you look at it from the perspective to stability it is a wonderful thing.

Re:Same old dilemma, new format.

[Gr8Apes]

I don't care who your are, now that's just funny!

Re:Same old dilemma, new format.

[Gr8Apes]

4 cores will sure help - I can now play 4 instances of solitaire simultaneously!!!!

Re:Same old dilemma, new format.

[Gr8Apes]

First off, don't think of them as cores, but parallel processes. Once you get over that hump, you can start coding with parallelism in mind, and then you automatically make use of however many cores/thread processing units you have available. Why should it be 1 process/core as implied above? That's inefficient. I'd want the AI using all available cores when needed, giving processing time up for higher priority processes, such as user IO processes (I personally never want to wait on the game). Once CPUs get fast enough, and I believe we're there already, such a system will allow almost instantaneous response to user input while delivering a seriously mean AI.

To the GP - nice tangential reference!

Re:Same old dilemma, new format.

[SharpFang]

Currently may work for any multi-core console. Splitting tasks between 3 cores of XBOX360 seems like a plausible strategy. Exploiting 8 "satellite" cores of Cell for varied serializable tasks would likely make sense too.

Re:Same old dilemma, new format.

[courtarro]

HDV home video is still rare as hell

Hell is rare?

Re:Same old dilemma, new format.

[Dr. Spork]

Look, you could launch the same objection against any CPU improvement: for most applications, the user won't even notice a difference between the best and, say my "budget" E6300 (overclockable but unoverclocked - I don't need it!). What are the exceptions? Games are one sort of thing where an excellent processor makes a qualitative difference - and media applications are the other. I'm hard-pressed to think of some other mainstream software that tests modern CPUs.

Games are a tough nut to crack. We'll see if game engine coders figure out how to separate threads. All claim that it's proving very hard. For now, your best gaming bang for the buck is still a fast single-core processor.

But the other mainstream application, media encoding, really does profit a great deal from multiple cores. Here, software support is easy. Many audio encoding apps let you run multiple tasks simultaneously, so you can keep arbitrarily many cores at 100% without need for crosstalk.

So my point is, an 80+% improvement in one out of two application categories that still need improving... ain't bad.

Now I think the real quesition is, how much media encoding do you think you're going to do? Because people who don't do much, and don't have a fetish for recent games, have no buisiness getting expensive CPUs. Those people should instead be focusing on reducing power consumption.

Re:Same old dilemma, new format.

[LWATCDR]

The question is more when is it fast enough. People that play games, work with huge graphics files, video editing, CAD, and simulation will probably never have enough CPU power.
However for 99% of business users anything over 1Ghz is probably over kill. At least until Vista comes out :)

Re:Same old dilemma, new format.

[cskrat]

You shouldn't have to code for a specific number of cores so long as your code is written with parallelization in mind. PC games will most likely be written in a manner that creates as many discrete processes as makes sense for the game. If you were to write a game that had, let's say, five discrete processes, the system should be able to schedule them as 2/3 or 1/4 on a dual core and 2/1/1/1 on a quad core. Considering that modern chips are very good at context switching, there wouldn't be that great of a penalty for running a multi-process game on a single core system compared to a game coded to run in a single process.

Re:Same old dilemma, new format.

[poot_rootbeer]

More cores is pretty much the same as more CPUs.

True, but where in the past multi-CPU commodity computers were rare, now they are becoming more and more ubiquitous. The chip manufacturers are adding cores to their designs now instead of just ramping up clock speed to improve performance. Software designers will have to follow suit.

Not every software problem has a solution that makes good use of multiple cores. But there are many that COULD, but simply haven't yet because of a lack of marketplace support. Soon they will.

Re:Same old dilemma, new format.

[mnmn]

Lets make a new game console using the Sun Ultrasparc T1 CPUs.

I was mildly tempted to buy one, and make a rudimentary game, with one core running the base game, anther just being the GPU and having 6 enemies with really good AI.

But then, the PS3 came along.

Re:Same old dilemma, new format.

[Schemat1c]

Hell is rare?

From what I've heard it's more like well-done.

Re:Same old dilemma, new format.

[ChrisA90278]

One core dedicated to enemy AI.
One core for physics calculations

That is NOT the way you design multi threaded aplications. No, what you do is to identify varius proceses that you need to do. Perhaps one of them is to read user inputs, read inputs comming in from the network (Internet). another to move the "camera" and then one process for each object to be simulated. So you might have 40 or 100 threads running. Then you would depend on the operating system to allocate physical "cores" to software threads. With this type if design performance can scale as you add cores. ASo in 5 to 10 years when we all are using 32 or 64 core machines it really will run 8 times faster How many threads is an enginerring decision. For example if you have four cars running on a raod do you use a thread for each car or one thread for all cars, another for all people. Thread per object is a radical design but thread per type of object is less radical

Re:Same old dilemma, new format.

[ckaminski]

Not for nothing, but HT is only good for threads that stall and make use of widely spaced uncacheable memory.

It also has nothing to do with the ability of an OS to preempt other threads to allow you to kill processes. That's just certain OSes being extremely dumb in how thread prioritization works.

Re:Same old dilemma, new format.

[lcsjk]

Funny, I don't get it!

Re:Same old dilemma, new format.

[try_anything]

creates as many discrete processes as makes sense for the game

I love this comment; wish I had mod points. Most of the difficulty with concurrent programming happens because programmers brutally hack a single-threaded program into an arbitrary number of pieces without looking for the easy, logical way to decompose it. A good programmer thinks like a butcher, cuts at the joints, and doesn't work half as hard as the programmer who saws through a program's bones and tries to knit them back together with mutexes.

Re:Same old dilemma, new format.

[smenor]

Some applications will make use of it, some won't.

I hear that argument all the time.

What I don't get about it is that most of us are running OSes that have more than one process running at the same time. Even if none of those processes is coded to support SMP and they were all single threaded, it's very useful to be able to run more than one process at the same time.

Even if only one or two are gobbling up the bulk of CPU usage and they're stuck using a single core, it's nice to have the UI remain responsive and to be able to do something else while the hogs are running in the background.

Re:Same old dilemma, new format.

[fm6]

However for 99% of business users anything over 1Ghz is probably over kill.

Dude, for 99% of business users, anything over 300Khz is overkill. It just just doesn't take a lot of processing power to run a spreadsheet or format a Word file. Most older machines would benefit more from a RAM upgrade or a defragramented swap file than they would from a faster CPU.

At least until Vista comes out :)

Actually, the official requirements for Vista emphasis the graphics processor more than the CPU — as they should. Of course, this won't prevent computer salespeople from using Vista to sell Quad-Core systems....

Re:Same old dilemma, new format.

[LWATCDR]

"It just just doesn't take a lot of processing power to run a spreadsheet "
that depends on the spreadsheet. I have seen some big honking spreadsheets in my day.
Actually I was going to say 500Mhz. 300khz is slower than a C-64.

Re:Same old dilemma, new format.

[elgatozorbas]

One core for OS, daemons and to bring them all and in the darkness bind them

You mean: "One core for BSOD"?

Re:Same old dilemma, new format.

[abradsn]

Another typical use for a threaded application would be to make one process that takes a long time to work, and help it finish up more quickly.


Ie.) sort through a terrain, for better level of detail.
Ie.) allow a bunch more AI to operate at the same time.


I agree with your design statement towards the end your post, but the multiple inputs anology that you make isn't very good since, it doesn't really apply very well nowadays, and didn't really apply very well 10 years ago either. Also, moving the camera is a fairly linear task and as far as I can see doesn't lend itself well to being parallelized.


Basically a good point, with a poor analogy. I'll try to improve on it a little here.

The multiple inputs that you mention while executing other tasks has been taken care of in hardware for quite a while already.

Another thing to point out, is that if you know you are going to have more than one core, then you can fairly easily plan for a bunch more cores. Threads and processes make this easy, since their paradigm has been around for a long time and programmers are already used to the idea.

The problems really start to creep in while you are testing to make sure that all cases work correctly. Though, I would argue that this is harder to do with any multithreaded application.

So, then, if you ask me, then the big waiting game we are playing is that there aren't a whole bunch of programmers with the kind of experience necessary to root out these very difficult bugs that occur. These are the kind of bugs that everyone has seen from time to time... Oh, that was wierd! ... Try it again ... hmmm, seems to work fine now. They just happen a lot more often when the threads are actually executing on different processors, instead of just in different time slices on the same processor.

So, I'm sure that in about 5 years there will be some good tools to help the other programmers figure out these issues more easily. I look forward to that and welcome my multi-core overlords.

Responding to the speedup, I just wanted to point out that their is often a diminishing return on the number of processors being thrown at a problem.


here is a hypothetical situation of a process that takes 10 seconds to run
We are assuming a perfrect speedup, at every additonal processor (which is not true in reality)
1 cpu = 10 seconds of runtime
2 cpu = 5 seconds of runtime
4 cpu = 2.5 seconds of runtime
everything below here is going to seem instantaneous to a web user or a database user.
8 cpu = 1.75 seconds of runtime
16 cpu = .83 seconds of runtime
32 cpu = .41 seconds of runtime
everything below here is going to seem instantaneeous to a user of a software application on a local machine

Re:Same old dilemma, new format.

[fm6]

Mega, Kilo, picky, picky, picky. Anyway, there's more to processor power than clock speed. As I recall, the 6510 in the C-64 didn't even have hardware multiplication!

I really think a 300Mhz system (or even a 100Mhz system) is capable of running most office software. What kills most of them is that they were sold with 64 to 128 MB of RAM in order to keep the price down. You need at least 256 MB to run MS Office without VM thrashing. That big honking spreadsheet you mention probably spends more time waiting for cells to get swapped in and out of memory than it does waiting to do the simple arithmetic you see in a typical spreadsheet.

Re:Same old dilemma, new format.

[LWATCDR]

"As I recall, the 6510 in the C-64 didn't even have hardware multiplication!"
that is correct and also no floating point. But it had zero wait-state memory and was actually pretty fast for it's day :)
Also only 2 registers and a single accumulator. Oh and those where all 8 bit.
You kids multiply? It is nothing but a repetitive add anyway.

Re:Same old dilemma, new format.

[Surt]

Games will have no problem using multiple threads as soon as the directx drivers support it ... the problem is that nearly all of the bottleneck is in the directx api which is all very single thread oriented. It'll probably be directx 11 before we really see this problem start to disappear in games.

Re:Same old dilemma, new format.

[sbohmann]

Ash Kôr Kr'Ashnaz AI Ash Kôr Kr'Fisikchkal Ash Kôr Kr'Darbanak Ash Kôr Kr'OS, Gûl, Thimbatuluk agh Burzum Ushi Krimpatul.

Re:Same old dilemma, new format.

[sbohmann]

Ash Kôr Kr'Ashnaz AI Ash Kôr Kr'Fisikchkal Ash Kôr Kr'Darbanak Ash Kôr Kr'OS, Gûl, Thimbatuluk agh Burzum Ishi Krimpatul.

Re:Same old dilemma, new format.

[sbohmann]

Ash Kôr Kr'Ashnaz
AI Ash Kôr Kr'Fisikchkal
Ash Kôr Kr'Darbanak
Ash Kôr Kr'OS, Gûl, Thimbatuluk agh Burzum Ushi Krimpatul. Argh... The preview betrayed me, but now it's right

Re:Same old dilemma, new format.

[sbohmann]

true, but in the case of those certain osses, any kind of dual processors, virtual or not, help significantly.

Could it be that W***o*s actually has one system queue per core?

Re:Same old dilemma, new format.

[Thundersnatch]

There are many issues with the "one thread per connection" or "one thread per object" sort of models. There is a lot of overhead per thread, not the least of which is stack space. On a modern OS, a typical thread will have around 1 MB of stack and other "housekeeping" space allocated to it, meaning at most a few hundred threads can fit in 2 GB of memory when you account for actual application data.

This is why all scalable web servers and database servers use "thread pooling" or "connection pooling". Do some Google searches on those phrases to learn the details.

The "everything is a thread" model is for lazy programmers, as surely as single-threaded programming is.

Re:Same old dilemma, new format.

[l33t gambler]

Actually, the official requirements for Vista emphasis the graphics processor more than the CPU as they should. Of course, this won't prevent computer salespeople from using Vista to sell Quad-Core systems....

Why should they? Because most of the GUI is bitmaps and vectors stored and rendered on the GPU? Then why is resizing an IE7 window so CPU intensive? Why is icon listing and resizing with File Explorer so jerky in Vista?

I had hoped Vista had a true 3d accellerated GUI but I've tested RC1 with a GeForce 7800 GS 256MB and its much more tearing and jerky then Windows XP is. It seems Vista just adds a texture with transparency, blurring and cubemapping ontop of the old GUI. I get 100% CPU usage in "classic" mode too. Using RC1 drivers from nvidia.com.

http://jooh.no/root/Vista_Report/resize_cpu_usage. png

Re:Same old dilemma, new format.

[fm6]

Then why is resizing an IE7 window so CPU intensive?

Have you actually counted the cycles?

Re:Same old dilemma, new format.

[sdaemon]

You kids multiply? It is nothing but a repetitive add anyway. Even simpler than that, if all you want to do is multiply by powers of 2 :) that you, brian?

Re:Same old dilemma, new format.

[LWATCDR]

A simple shift and it will even work for a divide.
And I know not of this brian of which you speak :)

Re:Same old dilemma, new format.

[try_anything]

Aside from working on well-designed apps written by other people and creating a few amusements, I have converted a single-threaded, single-process commercial application to a multithreaded application.

But who care? Everything I learned could be summed up in a bunch of platitudes anyway. Platitudes wouldn't be so damned popular if people didn't cause most of their own problems by ignoring known facts and techniques. So it is with threading.

It's the bandwidth stupid!

[siewsk]

It's the bandwidth stupid! Does not matter how fast the CPU is if it is bandwidth limited.

Re:It's the bandwidth stupid!

[osho_gg]

You are only saying this because the current state of the art is bandwith limited. It is much too trivialized generalization to say it "it is bandwith stupid!". Computer architecture and preformance is all about compromises, bottlenecks and parameter tuning. Not so far in future, there will be more bandwidth and then the limiting factor could be the speed/size of the memory. Or, it could be the power envelop of the entire system, Or, it could be back to the raw performance of CPUs. Or, it could be limiting opportunities for parallelization in most common workload for a typical office/home user. A little further out in future, there will be integrated GPUs, maybe even memory etc. and then it could be dinosaur hard-drive technology that may be limiting factors of overall system performance.

At any point of time in the history of computer performance, to say that, "it is stupid *anything*!" is much too simplistic point of view IMHO.

Osho

Re:It's the bandwidth stupid!

[ZachPruckowski]

From the looks of things, a Kentsfield on a 1066MHz FSB did similarly to a Kentsfield on a 1333MHz FSB, suggesting that bandwidth isn't the key issue. In fact, they scored the same on several tests (both were at the 2.67GHz standard speed). That surprised me, but I'm not complaining. Now, it may be that those tests never utilized all four cores, but it still scored head and shoulders above a E6700, suggesting that at least 3 cores were used.

Re:It's the bandwidth stupid!

[convolvatron]

of course you are right to a large degree. global memory bandwidth
is the cause of the day and given coherency, its not trivial to
architect around. the parent may have been a little terse, but
as you point out, overall throughput doesn't go up if all the
cores are too starved to issue.

however the memory latency picture isn't changing very much, and the
most compelling method to hide it for general purpose machines is through
thread parallelism (ignore vectors for a moment, its kind of a special
case of the same thing).

this is what makes the multicore picture interesting. assuming a workload
that can exploit it, you can really turn the scale knob pretty far up.
unfortunately the whole affair pivots on being able to get past the
crappy heavyweight thread-and-lock model the software people have grown
so fond of. and the software community isn't particularily light on its
feet.

Re:It's the bandwidth stupid!

[adam31]

Not so far in future, there will be more bandwidth and then the limiting factor could be the speed/size of the memory. Or, it could be the power envelop of the entire system, Or, it could be back to the raw performance of CPUs.

Ah, but the future is now. Cell has already addressed these issues: 25.6 GB/s main memory bandwidth, 256 kb of L1 cache per core, OoO sacrificed to minimize heat, maximal raw performance of CPUs in FP, integer, FP, load/store, FP, and main memory transfer (DMA engine) without any silliness like 8 GP registers (128). Even when multiple Cells are hooked together, it's over a 35 GB/s IOIF port.

Also for onboard multitasking, you forgot about being latency-bound by atomic operations, which is something that would be really bad with separated L2 caches. This issue is also elegantly handled by cell by having only a single bus-snooped L2.

It must be frustrating for the hardware guys. You address all the bottlenecks in a pretty uniform way, and they still criticize: "But... uh, the software guys need a refresher course in hypertasking..."

Re:It's the bandwidth stupid!

[julesh]

Note that this chip does have a faster FSB speed (333MHz x 4 words = 1333MHz) than the current-gen Core 2 Duo chips (266MHz x 4 words = 1066MHz). Possibly not *enough* faster to support twice as many cores, but that depends heavily on what your application is.

Re:It's the bandwidth stupid!

[uglyduckling]

OoO sacrificed to minimize heat

Surely it's not necessary to sacrifice openoffice.org, can't it just be tuned a bit to keep its processor useage down? Maybe it will be the bottleneck for a few more years to come, but eventually it will make better use of system resources, I'm sure. Or are you just a Microsoft Office fanboy?

Re:It's the bandwidth stupid!

[bangenge]

Does not matter how fast the CPU is if it is bandwidth limited

The bandwidth helps, but it's not limited to it. Even with all the advancements we've had in bandwidth, non-volatile storage is still the chokepoint. Most apps fit into main memory, and sometimes, cache. Most apps see a big jolt in performance with more cores (particularly properly coded multi-threading apps).

But the real culprit lies in the inherently non-parallel nature of the x86 architecture. All the advancements of processor architectures (out of order processing, hyper-threading, multiple cache) was designed to work around the limitations of x86. Itanic^W Itanium was supposed to help, but (un)fortunately, it was too late since a lot of the users were already entrenched in with their x86s. And unfortunately for us, only intel and probably AMD have the resources to make a dramatic shift in the architecture of the masses.

Re:It's the bandwidth stupid!

[masticina]

1333Mhz to connect 2x2 cores. I guess I have to agree, how fast will you hit the memory bandwidth limitations...

I have more faith in an HTT based solution that hooks up a quad core AMD processor then an FSB limited method.
And yes I know even HTT has it's limits..sure..it still won't hit those limits at fast as Intels FSB. Lets hope that Intel gets working on something alike HTT..or just license to use HTT. It would give their CPU's a direct speed up..by providing more then enough bandwidth!

Re:It's the bandwidth stupid!

[Joe The Dragon]

that is where amd's 4x4 and quad-cores will kick it's ass. Amd 4x4 has more ram and system bus band with them this it may let you have 2 full links in to the chip set. each 4x4 chip has 3 HT links.

~ = no link

      x16 ~~ x16
        | ~~~~ |
    ****** ~ ******
    |chip| ~ |chip|
    ****** ~ ******
        | ~~~ |
    ***** ~ *****
?-|cpu|----|cpu|-?
    ***** ~ *****
        | ~~~ |
      ram ~~ ram

? = extra ht link

Experts?

[Angst Badger]

Even expert opinions are deeply divided, ranging from 'more cores are absolutely necessary' to 'why do I need something more than my five-year-old PC system?

These are obviously experts who have never heard of servers.

I'm perfectly content with my 1.2GHz single-core single-processor laptop, but I'd sure as shit like to have more muscle in the database cluster I'm responsible for maintaining at work. Whether these chips are a good solution remains to be seen, but that's a separate question.

Re:Experts?

[Aadain2001]

Both AMD and Intel know they can't focus soley on the server or desktop market, but on both. While these chips are great for servers (what high powered chip isn't??), they will also be targeted at desktop machines (and may laptops in the future). With the performance of single core CPUs reaching their limit (thank you leakage current and high temps), multicore and multiproc systems will be the future of computing. Yes you can quote me on that.

Re:Experts?

[steveha]

The companies who are really serious about servers are particularly interested in CPU power compared to heat dissipation -- thermal density. This new Intel CPU is high performance with high heat--more of a gamer chip. At least so far it is; it's a very early sample and Intel hasn't had time to tune the power management features.

Intel's latest chips are fabbed at 65nm, while AMD is still only shipping chips fabbed at 90nm. This should give Intel a serious edge in the performance/heat ratio, but AMD's chips are so much more energy efficient that they are still competitive. (The current best performance/heat is the AMD Athlon64 X2 3800+ ADD chip.) When AMD finally ships 65nm Opterons, those ought to be really great for dense server installations.

It's telling that even Dell is planning to ship servers with AMD chips. They announced a 4-core server; two dual-core Opterons. It wouldn't surprise me if they will be 65nm Opterons when they finally are released.

The article says that Intel is going to transition from 65nm to 45nm sometime in 2007, and to 34nm sometime in 2009. They beat AMD to 65nm big-time. They may well be at 34nm before AMD can make it to 45nm! Just imagine some sort of server chip with 16 cores... or more likely, 8 cores and a whole bunch of cache.

But we shouldn't count those chickens before they hatch. Right now Intel is at 65nm and AMD will be there soon.

steveha

Re:Experts?

[julesh]

I've heard a lot of experts suggest that scaling outwards (i.e. adding more nodes to the cluster) is a better solution than improving the performance of individual nodes. They look at google as a model of how to build a high-performance database application.

I'm not convinced, but that's one point of view that's often expressed.

Re:Experts?

[kestasjk]

I've heard a lot of experts suggest that scaling outwards (i.e. adding more nodes to the cluster) is a better solution than improving the performance of individual nodes. They look at google as a model of how to build a high-performance database application. I'm not convinced, but that's one point of view that's often expressed.

Well obviously it depends on the application.. If you're running a database system which needs data from the whole database very quickly then splitting the database system up into lots of nodes will be much worse than trying to beef up individual nodes.

Re:Experts?

[MMC Monster]

Forget servers. I want lots of cores on the desktop.

Imagine tagging images using facial recognition. (I take a lot of photos.) Or creating three dimensional models out of multiple images. Perhaps with recommended angles to improve the 3d reconstruction. Or true speech-to-text and text-to-speech that is able to learn. Or true machine AI.

These are big software problems that need both coders and horsepower.

Re:Experts?

[dfghjk]

It is Tom's Hardware after all. The experts he's referring to are Mom, Dad, and his 14 year old brother.

Re:Experts?

[laffer1]

I've been hoping computer science programs will start addressing this issue. The masses have SMP and no one can use it effectively. I've never seen a great book on SMP. Some have coverage of basics, but not the type that lend to actual class work.

Re:Experts?

[99BottlesOfBeerInMyF]

I'm perfectly content with my 1.2GHz single-core single-processor laptop...

Back in the day, I was surprised to see how much a second processor helped for my workstation. Sort of the way having two eyes provides a significant advantage over one in nature, having two processors does the same for a workstation. Poorly written software that tries to monopolize CPU resources does not make the machine any less responsive. I'm less convinced of the benefits of more processors/cores until software development finds a better way to take advantage of them, but my old dual-533Mhz is still going strong as a low bandwidth server, media machine, and PVR. I'll be quite happy to have a second core in my laptop.

Re:Experts?

[Anonymous Coward]

For us budget minded dba's, multicore is nice. MS, for instance, sells per-processor licenses for SQL Server for somewhere around $4000. They've stated that for licensing purposes, a multicore chip shall be considered one processor. Therefore, if I can cram four cores under one license, I can get $16,000 worth of parallelism in my database processing for 1/4 the price. Since I'm budget constrained to begin with, I'm obviously not building anything dense enough to really care about flops/watts. I just want that multitable join done now, dammit.

I'm sure there are any number of other packages where the same kind of math holds true...

why do I need something more than my five-year-old

[SpaghettiPattern]

why do I need something more than my five-year-old PC system?
I would. Because these chips will eventually wind up in blade systems which will run Linix which will be used to port CPU intensive tasks.

Props for Intel for being early

[elh_inny]

Some nice example where more processing power (even in parralel) is nice is virtualisation, whether at home or on servers. Running multiple OSes in parallel will saturate all your processing power nicely.
What's more quad-core surely gives more processing power per watt and per cubic meter which is a very important factor for big folks like Google or whereever hosting space is expensive.
Even John Carmack who used to be very much against multi-cores for gamins recently elaborated much on this area in his keynote. Practically any modern (lets call it nextgen :D) gaming platform is now multi-core.

So I'd say overall it's nice that Intel is pushing this so fast, if developers start to realize that multi-cores are hitting mainstream, they will have to take that into account and by the time Intel and AMD launch 8-cores, there should be more software to take advantage of it.

Re:Props for Intel for being early

[AaronLawrence]

Running multiple OSes in parallel will saturate all your processing power nicely.

Only if you're actually DOING something CPU intensive in each OS. Virtualisation doesn't change the picture: you still have to be doing parallelisable tasks.

Re:Props for Intel for being early

[masticina]

Of course this is all due to our mono-core solutions hitting the "roof" that we can reach withour technology. As the old P4 articles already say..how smaller the production process how more the processor "leaks" energy. And how hotter it becomes..atleast in the silicon and strained silicon variants...

So to get more power after hitting such a roof your best bet becomes to improve things. Like we have seen with AMD K8..a processor with performance! Finally we moved over to Performance instead of speed ( like we moved from old 8liter engine cars to 2liter cars with just the same horse power in about 60 years ). And to get more speed adding more cores..was just around the corner!

Multi core is here to stay just because..Single cores are not going to cut it anymore..

Re:why do I need something more than my five-year-

[RuBLed]

I would too.. because 5 years ago, I got a Pentium 1 - 166mhz non mmx pc with 48mb of ram (upgraded) and an Imac DV 1999 candy models (G3 processor-400mhz)

but the real question is

[Rooked_One]

how many FPS can I get in quakeworld? With the +1000 FPS it would give, i'm sure I would be able to bunnyhop all the way across the 2fort5 outdoors area!

Re:but the real question is

[RuBLed]

it could be possible (maybe someone would tweak it in the future), but I doubt your existing crt/lcd would support that :)

What would improve is the quality of the games both in physics and in AI.

Re:but the real question is

[temojen]

I doubt your existing crt/lcd would support that

Just tweak the timings on an 80Hz CRT so the electron gun is only covering 1/125 the height of the screen. Run at 15x11@1000Hz mode. No problem. You could probably set this up with XVidtune or the right xorg.conf Mode lines.

On a related note, glxgears says it gets "6690 frames in 5.0 seconds = 1337.819 FPS" on my box, which isn't multi-core.

Re:but the real question is

[ichigo 2.0]

it could be possible (maybe someone would tweak it in the future), but I doubt your existing crt/lcd would support that :)

Disable VSync?

Re:but the real question is

[Rooked_One]

no one else said it, so I must... thats a really 1337 number you have there....

Re:why do I need something more than my five-year-

[Kangburra]

I think the idea was comparing a new system against a system bought new five years ago. If you bought that spec five years ago they must have seen you coming!

Games are going parallel

[savuporo]

Coincidentally, Gamasutra just two nice feature articles on rearchitecting the game engine flow to better parallelize the tasks so that multi-core can be taken advantage of, utilizing OpenMP
"Multithreaded Game Engine Architectures "
http://gamasutra.com/features/20060906/monkkonen_0 1.shtml
"Multi-Threaded Terrain Smoothing"
http://gamasutra.com/features/20060531/gruen_02.sh tml

A BIG question???

[RuBLed]

Why was it code named "Kentsfield"?

Did I miss something here... I thought it was called Kentsfarm?

Re:A BIG question???

[Nom du Keyboard]

Did I miss something here... I thought it was called Kentsfarm?

Kent's Farm is where Superboy grew up before he became Superman. It was a rights issue.

Re:A BIG question???

[vertinox]

Kent's Farm is where Superboy grew up before he became Superman. It was a rights issue.

See... We told everyone Trusted Computing was a bad idea.

Re:A BIG question???

[Gamefreak99]

I ain't gonna work on Kent's farm no more.
No, I ain't gonna work on Kent's farm no more.
Well, I wake in the morning,
Fold my hands and pray for cores.
Intel's got a head full of ideas
But everyone's a bore.
I ain't gonna work on Kent's farm no more.

One other thing

[iced_773]

Core Duo, Core Solo. What are we going to call this one? nVidia is already using the name Quadro.

Re:One other thing

[Jeremi]

Core Duo, Core Solo. What are we going to call this one?

Based on Intel's recent naming conventions, I think they'll call it the "Core 2 Duo Duo", so as to generate as much confusion as possible :^P

Re:One other thing

[evilviper]

Based on Intel's recent naming conventions, I think they'll call it the "Core 2 Duo Duo", so as to generate as much confusion as possible :^P

No, that's much to straight forward for Intel.

I expect something along the lines of: "Core 2 Tre Quad Pentium 405".

And AMD's AM3 5235+ 3.1G X4 Thunderon is faster and cheaper, anyhow.

Re:One other thing

[tygerstripes]

Perhaps the "Bi-die double Core 2 Duo 2".

Or perhaps not.

I'm rooting for the "Mach 4".

Re:One other thing

[PlusFiveTroll]

Four core and seven years ago chipzilla brought forth on this on this die, a new processor, conceived in vanity, and dedicated to the proposition that all CPUs shall require a nuclear power plant to function.

Guess they should have called the platform Gettysburg.

Re:One other thing

[bky1701]

I think they are natnaring with nintendo to bring you the Core Wii Wii.

Re:One other thing

[PlasticArmyMan]

It'll render graphics and give you a close shave at the same time!

Re:One other thing

[Fengpost]

Core 2 Quattro and it does mud road easily!

Re:One other thing

[tygerstripes]

"How can we make our CPUs more attractive without actually improving their performance?"
"Well, they've had twin-core - let's give them FOUR cores! And an extra one on the back!"
"It's... it's brilliant!"

Yes, Intel are the new Gillette. ;-)

Re:One other thing

[SharpFang]

Unfortunately, it's not MACH, it's ix86.

Re:One other thing

[Hast]

I hear AMD is going with 5 cores.

http://www.theonion.com/content/node/33930

Re:One other thing

[tygerstripes]

And I'm telling them to stick two more blades in there. I don't care how. Make the blades so thin they're invisible. Put some on the handle. I don't care if they have to cram the fifth blade in perpendicular to the other four, just do it!

God, I nearly pissed myself laughing at that! Apart from anything else, I'm pretty sure that's exactly what happened when the first Intel dual-cores came out... (what a botch THEY were)

Onion, how I've missed you.

Re:One other thing

[CrackedButter]

My friend i have just one word for you "Thundercougarfalconbird".

Re:One other thing

[ostiguy]

Will HP have a packaged cluster with shared scsi and 2 core 2 duo duo systems? Is there a SKU for that? That would be cool for exchange, but for sql, you might want 2 2 core 2 duo duo systems, if you need the extra cpu horsepower.

Re:One other thing

[clickclickdrone]

No, it will be the Intel four-play and you just know what's going to happen next..

Re:One other thing

[tygerstripes]

Ah, finally - SX technology!

(Anyone hoping for Woody Allen's "Sleepers"?)

Re:One other thing

[oc255]

I needed that laugh.

Re:One other thing

[oc255]

Yeah, major laughter here at the onion. Especially the "mew mew mew" line. A very "wtf" while laughing moment.

Re:One other thing

[bfree]

The Sexium.
How Intel's marketing ever let a Pentium II, III, IV out the door with that name in waiting I'll never know but I wonder how any people will name their four Core's Andrea, Caroline, Sharon and Jim!

Re:One other thing

[Rooked_One]

and the fifth layer will processor your caluculations into the future so that you don't need to calculate as often as before! Why do I feel I ripped that off of somewhere..... oh wait - it was an old SNL skit regarding razor blades (when they started adding two blades as a standard)

Re:One other thing

[ClamIAm]

More importantly, what will the next chips be called? The Core 2 Octo? I smell (heh) a tie-in with sushi restaurants! And don't get me started on the Core 3 Hexadeka...

Re:One other thing

[Dolly_Llama]

I hear AMD is going with 5 cores.

My chip goes to 11.

Who Needs This Much Power?

[Nom du Keyboard]

So who really needs this much computing power?

I, for one, have never heard how less cpu power is better than more, so I sincerely doubt that more is really going to hurt me.

'Nuff said on this issue.

260 Watts!

[dbIII]

260 Watts! We've been looking at the wrong factors - it appears that global warming is related to Moore's law.

Re:260 Watts!

[iknowcss]

Al Gore's next campeign: I'm here to tell you all .. about computer bear pig. It's totally serial.

Re:260 Watts!

[HoboMaster]

"In summary, we determined that the Core 2 Quadro requires the same amount of power in no-load operation that the Core 2 Extreme needs at full load."

Jesus. You say it as a joke, but you're not kidding. Seems they could have done better with power stepping when idle.

Also, I want the video card they say they're using on the test rig: "ATI X19800XTX - 512 MB"

Re:260 Watts!

[Michael Woodhams]

(1) Notice that this monster system consumes 260W under load. So why do so many people think they need a 600W power supply?
(2) 260W is for the entire system, not just the CPU. (But much of it is the CPU - ~100W *increase* from core duo to this chip.)
(3) It is an engineering sample. We can hope the final will do better than this.

If not Intel...

[abshnasko]

"We would like to make it clear that these samples were not provided by Intel"

Then who were they provided by, exactly?

Re:If not Intel...

[Anonymous Coward]

>"We would like to make it clear that these samples were not provided by Intel"

>Then who were they provided by, exactly?

The chip must have been leaked (possibly deliberately). From the fine article: "On the left the Core 2 Duo and on the right the Core 2 Quadro. The serial number has been distorted to protect secrecy."

Fine, but it is the serial number on the left that has been redacted.

So either Tom's doesn't know its right from its left, or they've just blown their source.

Re:If not Intel...

[Kjella]

What part of that was unclear? Obviously they were produced by Intel, and come either from some internal employee or one of the mobo companies or an employee thereof that get pre-production samples. Since the source is at least violating an NDA, do you think they'll come forward with the source?

Re:If not Intel...

[AcidLacedPenguiN]

*tinfoil hat*
 
AMD.

more power is often bad

[r00t]

Typically this requires more fans. Fans are moving parts, thus likely to fail. Your super-fast computer could crash because of a 13-cent Chinese fan. Dead computers only go fast when you drop them out the window.

Fans are noisy. This causes other people to accelerate your computer at 9.8 m/s/s. They can sneak up on you because you're going deaf from the noise.

Duo 2 Sexo?

[BikeRacer]

TFA spends a little time describing that Intel doesn't have enough package area to use this iteration of the Core 2 Duo to make a 4 die, 8 core part. So, my question is: Ignoring likely heat and bandwith issues, is there a SMP architectural reason they can't put 3 dies in one package?

Re:Duo 2 Sexo?

[julesh]

Apart from anything else, the dies are almost certainly square, so 3 of them would require as much space as 4.

Re:Duo 2 Sexo?

[dimfeld]

It's been done before; the Xbox 360's CPU is a tri-core PowerPC. I don't know much about Intel's specific implementation, but on a general architectural level there shouldn't be anything stopping them from putting 3 cores in a chip.

Re:Duo 2 Sexo?

[james b]

I've wondered in the past why multi-core/multi-processor systems usually seem to have a power-of-two number of cores. This quote is interesting:

Besides, it's very rare for users to need an odd number of processors (in any of the parallel codes I've seen at least). Most parallel problems are able to work in parallel by decomposing some sort of domain (be it physical, a mathematical matrix, etc.), and this decomposition usually happens in more than one dimension (this generally is done to optimize computation vs. communication). So generally worst case are prime numbers of nodes and the best cases are powers of two.

So perhaps it's a convention borne of parallel computing algorithm design? But there could be a more fundamental SMP architecture reason - anyone know?

Re:Duo 2 Sexo?

[julesh]

OK, clearly I should RTFA before I say something stupid like that.

Err... architectural reasons. Yeah. Probably.

Re:Duo 2 Sexo?

[Badfysh]

Not that I know too much about this, but my first thought using logic was SMP = Symmetric Multi Processor, and clearly 3 dies is not symmetrical. I imagine it's not worth the hassle to develop such an architecture, probably four dies is simply easier to implement.

Re:Duo 2 Sexo?

[TheRaven64]

When you are moving data between CPUs on a bus (or even with something like HT, since you don't always have a direct connection), you have to include the CPU's ID number. For two CPUs, you need one wire / one bit to do this. For three, you need two, but for four you also need two. For anything between five and eight you need three, etc.

Re:Duo 2 Sexo?

[Pharmboy]

I think the Symmetrical in SMP means the chips are more or less the same, not that they are in equal numbers. Any SMP system will always have different loads on each cpu at any given time, so they are not USED exactly the same amount either. The chips do NOT work in 'lock step' or anything. Technically, you can even Speed Step one down while the other is at full speed. The Symmetrical part of SMP is not as symmetrical as one would think.

This is the idea behind using 2 physical cpus with the same stepping in smp. Not for speed, but because different stepping have different features (that is what stepping means, after all). Also, although not commonly done, I have heard of SMP setups with different SPEED processors. I have run SMP systems with cpus that had different cache amounts (256k & 512k on PPro 200mhz), thus different stepping, although I had to run the 256k chip as CPU0 and the MB ignored the extra 256k cache on the CPU1, making them both virtual 256k cache chips. This was a limitation of the chipset, not the specification.

There is no reason you can't have odd numbers, although if I remember right, the traditional x86 specification makes it difficult to have more than 15 (yes, 15 ie: 16-1) cpus on a single motherboard. Its supposed to be a legacy issue, although that may be dealt with by now.

Re:Duo 2 Sexo?

[SharpFang]

One reasonable option for use of 3 or any odd number of CPUs is when you use 2^n + 1 - 2^n CPUs for the paralellized application, 1 for OS, demons and all the overhead. If you split the task in half, you'd like both halves to end at the same time. It won't happen if one of the CPUs gets additional work from other sources. So keep 2^n "workhorse" CPUs and one "manager" to coordinate them, provide interface, perform the load ballancing and all the other tasks other than the computation itself.

Re:Duo 2 Sexo?

[Peldor]

Three would just be odd.

Re:Duo 2 Sexo?

[red_flea]

The article mentioned that the 2MB cache per die pieces are smaller than the ones pictured and that by using them, up to 4 pieces of silicon could fit. So by chopping the cache a little, you can make the die smaller and allow 6 or 8 cores in the same socket. But they say that this would take way too much power to be feasable...

Re:Duo 2 Sexo?

[TheLink]

Well in theory it could be possible for 3 CPUs to take up spots at the 3 corners of the square and the cache to take up one huge spot at the 4th corner.

But that would mean a very different design/layout for a 3 CPU chip vs a 1 CPU chip.

Maybe they will start doing stuff like that when 1 CPU chips become rare.

Re:Duo 2 Sexo?

[xenocide2]

I'd imagaine it's much easier to design a symmetrical die than it is to cook up something asymmetrical like a 3 core die.

Re:Duo 2 Sexo?

[Tower]

From Wikipedia's SMP entry

Symmetric Multiprocessing, or SMP, is a multiprocessor computer architecture where two or more identical processors are connected to a single shared main memory. Most common multiprocessor systems today use an SMP architecture. SMP systems allow any processor to work on any task no matter where the data for that task is located in memory; with proper operating system support, SMP systems can easily move tasks between processors to balance the workload efficiently.

Another style is NUMA which dedicates different memory banks to different processors. This allows processors to access memory in parallel, which can dramatically improve memory throughput if the data is localized to specific processes (and thus processors).

Other systems include asymmetric or asymmetrical multiprocessing (ASMP), in which separate specialized processors are used for specific tasks.

Re:Duo 2 Sexo?

[Pharmboy]

two or more identical processors

Yes. Three is more than two, so three processors can exist in a SMP setup. Identical is relative, as I have stated, I have installed processors that were NOT identical, but very close, and they worked fine. I remember reading about "unsupported SMP" combinations in Intel's literature (was PDF or I could quickly google a copy).

Again, the specification doesn't require there be an EVEN number of chips, and yes, you CAN use different speeds of chips, which is even LESS "identical" or different stepping chips, which are NOT identical. Being connected to the same main memory is what makes it symmetrical, not the fact that the chips are exactly the same. Because they don't have to be, although YMMV with advanced features. Obviously, stability is better if they are truly identical, but go look up a SMP table for a processor family and you will see "compatible" steppings: Chips that are NOT identical, but identical enough for SMP and run stable.

Just because you read something in Wikipedia doesn't mean you understand it. Try working with SMP for a dozen years or so, then decide if the article is oversimplifying it for the masses.

Re:Duo 2 Sexo?

[neurojab]

I've wondered in the past why multi-core/multi-processor systems usually seem to have a power-of-two number of cores....

So perhaps it's a convention borne of parallel computing algorithm design? But there could be a more fundamental SMP architecture reason - anyone know?

The power of two rule seems to dominate the x86 and much of the UNIX space, but other server architectures, that have been doing SMP much longer than x86-based servers, don't really bother with it. For instance, the IBM zSeries mainframe is available 7 or 54 (and several other numbers) of physical CPUs installed. Depending on the configuration, certain cores may be switched off when you buy the machine, and a "capacity upgrade" will switch them on.

Note that the primary users of the zSeries are large businesses, not individual users, and a business manager might make the decision about how big a machine to buy, and business managers don't care about powers of two. I suspect the power of two rule for SMP is mostly to appeal to geeks that are just used to powers of two.

Re:Duo 2 Sexo?

[Tower]

Wow - it has been a while since I got flamed by the parent posted for posting a small bit of information that is roughly in agreement with said post. There was no intent to discredit your post, just to provide the contrast to what you already stated in case anyone wondered "What would non-SMP multiprocessing be". Clearly the "identical" label in Wikipedia could be interpreted differently, and certainly you took that as a slight against your post. Chill out. If I wanted to discredit anything someone writes, I will use my own words and cite the references, not just paste in some related information... I was providing a little bit of reference for others so they could read and make up their own minds is a totally different thing. Way to take the one word that didn't agree with your post and fly off of the handle.

I think someone needs a little less/more coffee today... or according to your .sig... maybe you just didn't have enough Tequila for breakfast.

To summarize....
Pharmboy: "I think the Symmetrical in SMP means the chips are more or less the same, not that they are in equal numbers"

Tower: (I agree with that statement - why don't I throw some info regarding that in there in support... how about a quick def of ASMP and NUMA too) [info regarding SMP, using the unforgivable 'identical' word inside, but otherwise providing a valid oversimplification for the masses...]

Pharmboy: [rant]

Re:Duo 2 Sexo?

[Pharmboy]

It has been the day for rants.

If you go back and look at your post, you might see it could be taken as an arguement, since the wikipedia article somewhat contradicts my statements by making absolute statements (ie: indentical). Looking back, now I see what you meant, but the meaning wasn't clear at first glance.

Also, on /., most posts that start with a Wikipedia link ARE arguements, thus the initial confusion. Appologies if I stepped on toes.

Re:Duo 2 Sexo?

[jafuser]

Why not design a multicore system so that if a defect occurs in only one core, the chip can still be sold at a discount as a n-1 core chip, instead of tossing it in the trash?

Spare Capacity Usage.....

[IHC Navistar]

Hmmmmm.....

Maybe now I will have the ability to put all of my excess computing power into figuring out why women are the way they are instead of searching for extraterrestrial intelligence...............oh wait.....

-----

Sig Sauer

Dear "Expert"

[hsmith]

>> Even expert opinions are deeply divided, ranging from 'more cores are absolutely necessary' to 'why do I need something more than my five-year-old PC system?'


stab yourself in the face. Sure, for average joe downloading itunes for their ipod, 4 cores aren't necessary. But people like me wanting to roll out a database cluster, getting a cpu with four cores could save me $75,000 per CPU running M$ SQL Server. Oh the dream of running 16 cores on 4 CPU's. I think I'd like that savings as would anyone looking to build a beefy server and not get raped by MS or Oracle liceneses.

Re:Dear "Expert"

[Rodyland]

Until such time as MS/Oracle/VendorOfChoice decides to (re-)institute per-core (or even better - per-virtual-core) licencing...

You act like being raped by MS/Oracle/VendorOfChoice isn't a priviledge and an honour.

Re:Dear "Expert"

[Vendekkai]

But people like me wanting to roll out a database cluster, getting a cpu with four cores could save me $75,000 per CPU running M$ SQL Server. Oh the dream of running 16 cores on 4 CPU's

Till Oracle and Microsoft revise their licensing terms to take into account multiple cores, that is.

Or do you think they're going to sacrifice all that potential revenue?

V.

Re:Dear "Expert"

[jimicus]

The only way to avoid being raped by MS or Oracle licenses:

    use free software

HAHAHAHAHAHAHAHAHA!

Seriously. There exists in this world proprietary software which the IT department doesn't get much of a say in. Finance departments are great at this, as are marketing ("We need Adobe Creative Suite!" "But you only use Acrobat." "We still need Creative Suite and if you don't give it to us we'll complain to the managing director and he's not technically literate enough to appreciate your point and you know it!") and even parts of the IT department. Free clue: the helpdesk just wants a system which allows them to log and track calls, not a religion.

This proprietary software frequently only supports one or two proprietary databases.

Yes, it sucks. But it sucks a whole lot less than getting sacked because you try forcing free software on the company like a religion rather than providing what the business needs to get on with its job.

Re:Dear "Expert"

[DrXym]

Using free software is easier said than done. In the real world there may be extremely good reasons for using commercial software from support, performance, scalability, tools, documentation etc. Besides, databases are extremely finnicky things - broadly similar in behaviour but extremely different in the details. Switching from one to another could cost far more in development effort than just paying Oracle the license to continue using theirs.

Re:Dear "Expert"

[merlin_jim]

quoth the grandparent:

But people like me wanting to roll out a database cluster, getting a cpu with four cores could save me $75,000 per CPU running M$ SQL Server. Oh the dream of running 16 cores on 4 CPU's

quoth the parent:

Till Oracle and Microsoft revise their licensing terms to take into account multiple cores, that is.

SQL Server already requires a license per CPU - not sure if a core counts or not, but I think the O/S sees it as a CPU, therefore it requires an extra license...

Re:Dear "Expert"

[temojen]

Le Voila!

Holy Crap

[mad zambian]

but those things are fast.
From TFA
40+ GigaFlop! (Sandra 2007 Arithmetic Float)
61 Giga Integer ops (Sandra 2007 Arithmetic ALU)
Wow.
Seriously impressive, and that is coming from a longtime dedicated AMD fanboy.

The question is, do we need this much processing grunt? My current AMD64 3500 is only exercised when I play with Fractal Explorer or Games; It spends most of its life at

Re:Holy Crap

[Overzeetop]

It really does depend on what you do with your PC. On the fun side, I dabble in graphics and video (just boring stuff, mostly - resizing and rotating digital photography, archiving shows off the TiVo, dice and splice videos of the family) - that alone is enough to want more power, even on a temporary basis. I also archive my music as FLAC and recode on the fly when I load my portable player. Just for grins I converted the whole collection to MP3 once, and it took 4 days. Copying the entire archive over my 1Gb network takes close to 1.5 hours, so it's not bandwidth limited. I've got a P4M-1.86 and 2 GB of memory on a "workstation" class laptop, and there are fairly few operations I do that don't involve getting up to watch TV between prompts.

For work, I run AutoCAD. Now, this may seem simple, but there is a "new" feature in Acad2006 which puts a semi-transparent version of the command line right next to the cursor. Sort of a heads-up display for CAD. With nothing but the basic "empty" autocad template loaded and all other programs at 0 cpu cycles, the heads-up command line slows the program down so much that it misses my normal click-speed operations. I've watched the CPU peg at 99% in the taskmanager just dragging the cursor around the screen. And that's with hardware acceleration on with an nVidia Quadra FX go1400 with memory that is 27x the size of the 32bpp frame buffer (256MB). It's either I slow down or the machine needs to get faster. I chose option 3: turn of the heads up display.

So, really, even on the desktop there are operations that really need more speed, and it will always be that way. I'm sure if I get a computer that's 4x as fast, I'll end up with a 28MP digital camera and I'll start shooting HD instead of DV video. And once again I'll be able to catch a segment of The Daily Show while the computer churns away at a large task.

With four cores...

[Bazman]

There's an ad for PC World or Currys or something where the salesguy explains to some students that having an Intel (bing-bong-bing-bong) dual core processor means it can do two things at once - like sending an email and downloading music!

Re:With four cores...

[Rodyland]

Talk about dumbing it down. Still, maybe they should put it in terms the average real computer user would understand.

How about, you can illegally download a DVD whilst watching another illegally downloaded DVD.

Or, more likely, you can play your pirate copy of Civ4 while compressing your CD collection so you can share it illegally over the internet.

Or, even more likely, you can watch illegally-downloaded pr0n on your monitor whilst watching more illegally-downloaded pr0n on your TV through your video-out.

I reckon with 4 cores hyper-threaded you could do all of the above at the same time. *swoon*

Re:With four cores...

[mei_mei_mei]

Dude!

Do you have two dicks?

Four???

Re:With four cores...

[mgblst]

Ha, I laugh at that. Finally, thanks to Intel, I don't have to stop winamp before I send an email - this has always been a pain. Having to go to the task manager and make sure that no only one program is running has sure been annoying. If only they had developed some sort of system sharing, lets call it multi-tasking.

Further proof why these stores really suck, and I will never go to them.

Re:With four cores...

[epp_b]

There's an ad for PC World or Currys or something where the salesguy explains to some students that having an Intel (bing-bong-bing-bong) dual core processor means it can do two things at once - like sending an email and downloading music!

Well, yeah, but that's just dual core. Now you can do four things at once! Imagine the possibilities!

Re:With four cores...

[mnmn]

So while the poor muck's MSDOS 6.22 Pegasus email is sending email, he's wondering where and how to download music.

Comment removed

[account_deleted]

Comment removed based on user account deletion

from intel's point of view

[Anonymous Coward]

it should be spelled the quad core due owe, they think it is their due for you to owe them a lot of money for something you really don't need.

The last couple of years I am finding it extremely hard justifying upgrades. I can surf fast, web pages render fast, can already watch a nice dvd movie or listen to fairly good sound, etc, and other multitasking things-all with a barely past 1 ghz cpu and half a gig of old slow RAM and a 45 dollar vid card. If I need some "upgrade", well, I still have two empty RAM slots, and that's cheap and probably a lot more cost effective in real world use than having to upgrade to some new chip that takes a brand new mobo as well. I am trying to see where having some webpage open 5 milliseconds faster or something like that is any sort of huge advantage that I should pay hundreds of dollars for. I's not like I would kick if it was only 20 bucks or something, but....not seeing it, computers hit a "good enough" level a few years ago now it seems, and I have heard that from any number of people in meatspace as well.

  I like a variety of gadgets, not just computers, and I'd say for the bulk of humanity out there, what they have right now (I am just generally speaking now) for computing power is actually filling the bill quite well. I realise that companies have to keep selling to stay in business, but this arms race is also creating a lot of planetary waste and over production of honestly unneeded stuff when it comes to computers. And cellphones are even more stupid is it REALLY necessary to get a new one and throw the old one away every few months? I would rather see this silicon (and peoples extra money they would have blown on an unneeded new computer) go to stuff like solar panels for the next few years ahead, that industry needs it a lot more and will help to drop prices there and help society as well.. I like consuming electricity and what it does, but I would like to step up to the plate and *produce* a little of it as well, especially with something that will come with a 30 year warranty like most PV panels do now.

Anyway, just a thought on this subject, I think it would be better if they told these chip places to just go into idle from a flat out race for a spell, give people a chance to really *use* what they already have, do some more R&D, and skip to every third generation or so to mass produce and sell.

Re:from intel's point of view

[aywwts4]

Damn, forgot the breaks, If you want to actualy Read my post, this might be more legible.

Vista will hopefully solve that problem, and 'progress' can continue unabated. ;)

But you have excellent points, I would have modded you up but I would rather respond.

Processor makers need to really work on energy efficiency of all their desktops, these speeds were achieved through sheer increases in heat and power consumption, and its really flatly unacceptable (My current desktop heats up my office to a toasty 89 degrees in the middle of winter if I close the door)

Truly computers have become terribly wasteful, Other than environmental impact (less lead and harmful metals) I doubt the $399 computer of today contributes less waste than the $2000 dollar computer of 1990, The only change is my 386 lasted for 7 years, and the 399 dollar computer, isn't very upgradable, comes crippled from the start, and gets thrown out when infested with spyware, because the price point makes it a disposable commodity.

The power saving features built into computers now are great, but its only so good if the processor is just short of having a nuclear powered core.

Not only that, but with your idea hopefully less money can be put into stopgap number increases down dead end roads and make processor manufacturers build their architectures with a little more forethought. Maybe if every 3-5 years there was a responsible and substantial leap in computing power people would upgrade in regular phases, without wasting so much energy, and without so many confusing acronyms and model numbers. Each generation would have a nice leap in power, and a new OS to match, each computer manufacturer would keep models to a minimum, and just slide the prices down during that generation until it hits everyone's sweet spot. Upgrades during the generation can focus on power consumption instead of computing power, programmers can aim and test for certain generations of computers, and system requirements will be simplified greatly, Less confusion, less waste, morestandardized component generations. "This software runs on generation 4 and 5 computing components"

Of course gaming is to blame for this constant demand, So I figure nothing will ever change until we hit that mythical peak in which everyone says "Good enough" and the race to real time photorealism is over.

(And yes, I too am very happy on my 6 year old system, I just wish I could give it a ram upgrade, The only problem being... I mistakenly thought RD-Ram was the future, Check the prices on a gig of that and have a laugh at my expense, In conclusion, sorry for the rambling, Slashdot is my outlet for insomnia)

Re:from intel's point of view

[Lonewolf666]

Depends a lot on the applications.

For surfing the web and running an office suite, I agree that an older machine does fine. Games will happily use all power your machine can provide, and so does compiling large programs, depending on language/compiler. My next upgrade will probably be triggered by something from that field. Right now, my 2.4 Ghz P4 is barely adequate to handle DOD:Source.
Video editing is supposed to be even more power hungry (I have no experience of my own with that).

Re:from intel's point of view

[joto]

Processor makers need to really work on energy efficiency of all their desktops, these speeds were achieved through sheer increases in heat and power consumption, and its really flatly unacceptable

Didn't you pay attention in class? All the processor makers have started doing this. For the last one or two years, the mantra has been "computing per watt", pretty universally, no matter which company you were from. And by the way, if you compare your 386 and a modern computer, I'll bet the modern computer gives more "computing per watt", no matter how you decide to calculate, so all improvements can't have been "achieved through sheer increases in heat and power consumption".

Maybe if every 3-5 years there was a responsible and substantial leap in computing power people would upgrade in regular phases [SNIP] Of course gaming is to blame for this constant demand

Ha ha ha. Hi hi hi. Ho ho ho. Hilarious! So you think if it wasn't for the gaming industry, we would somehow magically have every manufacturer on the planet agreeing to wait releasing their improved products untill someone (who?) said it was time for a new "generation"?

Apart from the fact that such a thing has never happened in other industries (or would you care to give a counterexample?), that it is uncompatible with the idea of a free market, and that it is bad for consumers as well as producers, please explain why you think that would happen AND why you think that would be a good thing.

So I figure nothing will ever change until we hit that mythical peak in which everyone says "Good enough" and the race to real time photorealism is over.

Yeah, because when graphics are suddenly "good enough", nobody needs computers for other purposes. People don't need plain old office computers, workstations, databases, web-servers, B2B-applications, industrial control software, or anything else. Because the only thing driving the computing industry forward is gaming. And when photorealism in gaming has been achieved, all computing innovation will stop. Right!

And besides, there will never be a "good enough" for graphics. In my view, audio has been "good enough" since long before the invention of the CD. That doesn't mean that there aren't people still working on creating better and/or cheaper audio components.

Re:from intel's point of view

[keesh]

Speaking as a software developer, I need the extra cores. Not for customers to run the software. For me, to compile things.

My previous development box was an Athlon XP 1700+. It did a full compile and test run of my current project in about sixteen minutes. I've just been upgraded to a Core Duo 2, which does it in three (make has a parallelise option, so it can use both cores). Give me a box that's twice as fast (which twice as many cores is, for compiling) and the compiler will finally be able to keep up with the coding, which means no time wasted sitting around.

Is it worth the money? For some people, yes. For others, clearly not. I couldn't justify getting a super expensive IBM pSeries box as a development system, but a cheapo x86 desktop doesn't even show up on the budget...

Re:from intel's point of view

[Tony Hoyle]

Except few compilers will actually use the cores. I have vs.net currently compiling on an X2 and it's not apprecialbly faster than my previous machine... bacause one core is basically unused. With a 4 core 3 cores would be unused, and it'd still be no faster. Under linux you can do things like make -j3 of course.

2 cores/CPUs is damn near essential for compiling as the compiler (especially on windows) can lock the UI completely on a single core by maxing out the CPU. 4 cores I can't really see the advantage right now.

Re:from intel's point of view

[keesh]

Not the compiler. Compilers are complicated enough that making them threaded is just a recipe for disaster. The build system is what should be doing the work -- in a project with lots of files, a decent build system (like, say, make) will be able to use all cores quite happily. Heck, make -j can use all 32 CPUs on our big release box without any difficulties.

If your build system doesn't support parallelisation, you should look into switching build systems. Developer time is not cheap.

Re:from intel's point of view

[noyren]

> The last couple of years I am finding it extremely hard justifying upgrades.

Yeah I've had that problem as well, I've decided virtualization is going to be my next excuse ;)

Re:from intel's point of view

[WuphonsReach]

For usability of said web/business applications, when dual-core only adds ~$70 to the price of a $800 system, why not add dual-core for reponsiveness?

That's been our decision. The price breaks on AMD X2 chips back in July 2006 made it an easy one ($300 chip now sells for $150, only $50 more then the $100 CPUs were were buying). For such a minor price increase, we get another few years of longevity due to the 2nd core.

Should be pretty easy for that system to save the user $50 over the course of a year because they're not waiting on all sorts of minor delays. There are even some energy-efficient X2 3800+ chips that just hit the street that could eliminate even the excess power consumption argument. Not that modern CPUs are that bad; they'll ramp their speeds back down when the power isn't needed.

Re:from intel's point of view

[squidfood]

Heck, make -j can use all 32 CPUs on our big release box without any difficulties.

Hey, thanks! Got a duo-duo machine for scientific computing last week. Task is nicely parallel so it's already 4x faster (clocked and confirmed), but just tried make -j (I'm a scientist not a programmer dammit I don't know all these tricks) and it's 4x, too. Whups, just finished, back to it. Less compile-time slashdot reading. Thanks I think.

Re:from intel's point of view

[rthille]

Of course if you've got your dependencies screwed up in the slightest you'll have issues. Not that you shouldn't be fixing those anyway, but as someone who went thru a build-system rewrite on a large project, that can be a lot of man hours itself.

Re:from intel's point of view

[jtwronski]

I am trying to see where having some webpage open 5 milliseconds faster or something like that is any sort of huge advantage that I should pay hundreds of dollars for.

If all you do is surf the web and watch movies, well then you'll be fine for some time. I edit movies quite often at work, and at 720x480x2000kb/s, I'm screaming for as much horsepower as I can get my hands on. I'm currently on a P2.4 with 1Gb of memory, and i'd like at least 4 times as much power as that. Cost be damnned. It'll pay for itself in saved time even before it becomes obsolete.

Re:from intel's point of view

[feepness]

My previous development box was an Athlon XP 1700+. It did a full compile and test run of my current project in about sixteen minutes. I've just been upgraded to a Core Duo 2, which does it in three (make has a parallelise option, so it can use both cores). Give me a box that's twice as fast (which twice as many cores is, for compiling) and the compiler will finally be able to keep up with the coding, which means no time wasted sitting around.

We just got an IncrediBuild distributed compile license. We have around a dozen machines on the link (single core). Our compiles went from 20 minutes... to... wait for it... just over 1.

No, I don't work for them, just a happy user. Now I have to get back to work, the damn compile is done already.

PS: This doesn't mean I think quad-core is useless... quite the opposite... but just thought I'd share.

Re:from intel's point of view

[Fulcrum of Evil]

Back in the dark ages when dual processors were rare, I found that I got a larger speedup by using faster disks than adding a cpu.

Re:from intel's point of view

[lazarusdishwasher]

If all you are worried about is compile time, the samba team have come up with distcc and ccache for distributed compiling and a compiler cache respectively. Using this set up a network with four computers with four cores would be able to use 16 cores with make -j 16 for a faster first time compile. The recompile would be faster because the compiler would only have to compile the parts that would actually change.

I don't code for a living, but it does seem to speed up gentoo when I am playing with emerge.

http://distcc.samba.org/
http://ccache.samba.org/

Re:from intel's point of view

[StikyPad]

The only reason I bought a Core 2 Duo was because my last system was on the fritz. It's somehow managed to screw up 4 DIMMs in 3 years, along with a video card. Beyond that, it's AGP, ATA, and DDR. As those standards age, the price of the components is actually increasing due to lower supply. Since I would need to buy a new motherboard (and that's not even guaranteed to solve whatever the fsck is frying my components), I went ahead and bought a socket 775 motherboard. Since I was doing that, I went ahead and got a Core 2 Duo, a shiney new graphics card, a couple gigs of DDR2, an SATA HD, etc. Not the top of the line models mind you, but one step down. Comparable performance for half the price. I agree that I could have kept using my 3 year old system had it not suffered catastrophic failures, and I don't plan on any upgrades for another 3-5 years, barring some bizzare new market entry like a quantum computer for $300. I was an early adopter-type in my early-mid 20s, but that era has passed and now I'm content with what I have. I've learned that great graphics != great games, and the graphics eventually get tuned out anyway after a few hours of play.

That said, my new system does feel faster, and I don't regret having purchased it. It's both a vast improvement, and it (hopefully) solved my problem of randomly frying components (although I also bought a voltage regulator/UPS to protect my "investment" [insomuch as you can call something that's guaranteed to lose value an investment] since I'm pretty sure the power here is at least a contributory factor).

But really, this is nothing new in the computer world: If your system is good enough to meet your needs and/or wants, then don't upgrade. I'm sure there are more than a few people here still running P3-era systems, if not P2 and earlier. But if you want more power, the Core 2 Duo's really do provide it -- especially if your system is a 3+ year-old P4 like mine was. I still run an Athlon XP 2200 for my HTPC, and it works fine for that purpose, but I don't mind the extra processing power on my main rig at all.

Re:from intel's point of view

[CuttingEdge]

Give me a box that's twice as fast (which twice as many cores is, for compiling) and the compiler will finally be able to keep up with the coding, which means no time wasted sitting around.

Just use Smalltalk which has an incremental compiler built in so that compiles take a fraction of a second. In a Smalltalk IDE system you never wait for the compiler during development.

More cores = better systems as long as programmers implement their multi-threaded applications using actual concurrency controls instead of half hearted hacks (as all hacks are).

Re:from intel's point of view

[Iron Condor]

computers hit a "good enough" level a few years ago now it seems,

Yeah - 640k are enough for everybody.

Problem with that test

[Sycraft-fu]

They didn't test Intel's Core chips. Not Core 2 but Core. Core 2 is the highest performance, most features chip, but the Core is the most efficient. The fastest Core Duos (2.33GHz) have a TDP of 31 watts, the slower ones (1.66GHz), 15 watts. Then there's the Core Core Solo which on the low end are 5.5watts TDP. Remember that the thermal dissipation spec is the max amount of heat they are ever going to put out (actually slightly more). It is the "If you design your cooling to dissipate this amount it'll never overheat," figure.

You can get Core Solos and Duos for desktop/server boards (browse Newegg).

I lack the hardware to run the numbers on a desktop system, but based on Laptop results I bet you find they are unbeatable in the performance/watt category at the present time. The Core Solo is especially attractive if you don't do multi-threaded work. I mean it's as fast as a reasonable speed P4 and 5.5watts? You've got to be kidding me I think I have power bricks that give off more than that idle. If I ever actually build a media centre unit, I'll check one out. I figure combined with a new budget card (like a 7300) that can accelerate video decoding it should get the job done and need no active cooling.

Re:Problem with that test

[Gr8Apes]

With all that said, why are the iMac, MacBook and MB Pro running hot then?

Re:Problem with that test

[wgaryhas]

I'm not sure where you are getting those TDP numbers, but the core solo T1300 has a 27W TDP and all core duos (T2300 to T2600) have a 31W TDP.

Re:Problem with that test

[Gr8Apes]

That was my point exactly. The TDP of the CPU is only part of the story. Comparing the TDP of an Intel chip to an AMD chip is apples and oranges, unless you include the NB chip as well.

Bring on the more complicated architectures!

[Rodyland]

As a sofware developer, I can't help but think the move to multiple cores is a good thing. In my mind, anything that makes software development MORE complex can only improve my employability.

Re:Bring on the more complicated architectures!

[Plutonite]

Wouldn't it be fun if you were right, if the main diffrence between the prestige of software.eng and other types was just this: complexity? You see more complex means greater need to design, and that is what engineering is..

Re:Bring on the more complicated architectures!

[niceone]

The biggest advantage to developers is that they can write sloppier code and it will still run at an almost bearable speed.

Bring one 8 cares!

Re:Bring on the more complicated architectures!

[Tower]

Device driver writers should already be comfortable with a multi-core type of environment... when it comes down to it, a single core system behaves like a multi-core for any bus hardware, since the processor and the bus HW can be accessing the same memory at the same time - throw in the ever-present interrupt or non-blocking requests from multiple threads/processes/tasks (your choice) and you already have more than enough elements to create all of the interesting situations... especially if your processor is weakly consistent, so that you don't necessarily see the latest data DMA'd by your adapter.

Re:Bring on the more complicated architectures!

[Rodyland]

Funny? Funny!? I wasn't trying to be funny - I thought I was being serious!!

Re:why do I need something more than my five-year-

[Yokaze]

5 years ago? The system you describe sound as if it was already 5 years in 2001. Pentium-MMX have been released in 1997.

I am pretty sure my current system (Athlon Thunderbird 900MHz) is from that era.

FSB

[Dersaidin]

They should have made the FSB 4 Mhz faster. That'd just be too cool. Fight back AMD! Don't take that shit!

Re:FSB

[Technician]

They should have made the FSB 4 Mhz faster.

Right now the bottleneck at home is the Internet connection. I wish it could get cranked up from it's 0.001 Gig speeds. It would be nice to get a connection speed somewhere near the CPU speeds. Then sending e-mail while downloading music might be possible. ;-)

Re:FSB

[Dersaidin]

Its just that
1333 + 4 = 1337

:/

Re:FSB

[Anonymous Coward]

or rename it to the roy core
so we got a 1333-roy :D

Re:FSB

[Technician]

Anonymous Coward on Tuesday September 12, @04:08PM (#16091296)
You should get a Pentium !!!. I hear it makes your Internet faster.


I already have one. But thanks for recommending one anyway. I think that may be part of my internet bandwidth problem. I have a bunch of machines hanging off one router which is hanging off a slow cable modem connection.

One thing they didn't benchmark on it....

[gsasha]

Is Linux Kernel compilation. It should rock there, that's an inherently parallelizable task.

As a programmer, I want one. No, I want two :)

Re:One thing they didn't benchmark on it....

[the_humeister]

They also didn't test it with povray, which would have been great since 3.7beta supports multithreading.

Us coders are delaying the Singularity!

[PaulModz]

I'm in the middle of Kurzweil's latest book. ;)

Finally, multi-proc hardware is heading toward ubiquity on the desktop, but every CPU review I've read in the last year or so points out the fact that lots of software needs to catch up.

Why is it so hard to get developers to write decent multi-threaded code? It's not that hard, and using threads properly can almost always improve performance and/or responsiveness on single proc/core machines to boot.

Re:Us coders are delaying the Singularity!

[savuporo]

It isnt, if you use the right tools.
http://www.openmp.org/
OpenMP to the rescue !

Re:Us coders are delaying the Singularity!

[joto]

Why is it so hard to get developers to write decent multi-threaded code? It's not that hard,

Let me put it this way. If all the developers in the world were as smart as you think you are, it would not be that hard. As it is, however, coming up with scalable, manageable, efficient ways of writing multi-threaded code, in a way that is future-proof, as opposed to simply optimized for todays generation of hardware, is hard. Very hard. Not as in research-subject hard, but as in continuing-research-for decades-has-still-not-brought-us-much-closer-to-a- solution hard!

, and using threads properly can almost always improve performance and/or responsiveness on single proc/core machines to boot.

Let me rephrase part of the above sentence: "using threads properly can...". Did you notice which word I emphasized? Can you guess why?

Any idiot can use threads. The difficulty is to find the right granularity of threads (which is related to what kind of hardware you've got), which tasks are parallelizable, which parallelizable tasks should (or should not) get parallelized because of communication overhead and other factors (which is also related to what kind of hardware you've got), and so on.

It is also important to note, that few existing programs are designed from scratch today. In fact, almost all existing programs were designed in the past! In the past they didn't have affordable multicore (or multi-CPU) computers. And thus, those old designs didn't take that into account.

Re:Us coders are delaying the Singularity!

[ardor]

Why is it so hard to get developers to write decent multi-threaded code? It's not that hard, and using threads properly can almost always improve performance and/or responsiveness on single proc/core machines to boot.

Because it IS harder. It introduces new pitfalls (deadlocks, livelocks, race conditions), debugging is harder (gdb with multithreaded programs.. brrr), old paradigma have to be thrown overboard (and new ones introduced, such as task- or stream-based processing). Also, threads NEVER improve performance on a single-core machine. They do help with responsiveness, however. If you want performance boosts, use a multicore machine.

Re:Us coders are delaying the Singularity!

[PaulModz]

Threads CAN and DO improve performance on single-core machines in today's connected world. If you need to query 10 SQL servers or download 10 files from the Internet, multiple conncurrent threads will always be faster than doing things one at a time in a single thread.

Granted, the performance improvement comes from harnessing the power of remote processors, and you are right that threads would NEVER improve performance on a single core box that isn't connected to a LAN or WAN.

Re:Us coders are delaying the Singularity!

[mpcooke3]

Also, threads NEVER improve performance on a single-core machine.

If performance is a measure of how much work you can get done in a time period and a program is bottlenecking on anything other than CPU usage then multithreading can vastly improve performance on a single core machine.

Re:Us coders are delaying the Singularity!

[Kjella]

Also, threads NEVER improve performance on a single-core machine.

Unless you call blocking functions (like IO), where one thread'll block and the other(s) will keep going just fine. But yes, if you're 100% CPU bound then making it 2x50% won't help at all.

Re:Us coders are delaying the Singularity!

[mpcooke3]

Because threads (with a shared memory model) is a very poor way to deal with parallelism. It's incredibly easy for even experienced programmers to make mistakes and sometimes the deadlock/livelock situation can only occur very infrequently makeing debugging difficult. Sometimes the bugs never occur when the debugger is running due to the difference in speed.
In most OO languages threads are orthogonal to the object model rather than integrated into it so you may not even be aware when working on a large system what the threading implications are.

There are other more reliable ways to write code that runs in parallel such as using a message passing system (Communicating Sequential Processes) but no major language supports this natively (that i am aware of) and you still end up having to interface with built in librarys that use the threading model.

Re:Us coders are delaying the Singularity!

[ardor]

Well, thats usually not the definition of performance when it comes to stuff like multithreading.

How many 8x8 DCT blocks the CPU can inverse-transform in a second - this is performance.
Multithreading won't help here. (But this task is easily parallelizable, so multi-core CPUs are a big win here!)

Re:Us coders are delaying the Singularity!

[ardor]

Here we see the problems with measuring performance and distinguishing it from responsiveness - sometimes it is very hard to do. The blocking IO almost always harms responsiveness, not performance. A file manager scanning thousands of files for thumbnails generation has responsiveness problems if this is done without multithreading.

In fact, I would see responsiveness as the bigger problem nowadays. It affects the user directly (for example, the file manager not reacting to any kind of input while scanning a large directory) and is ignored too often.

Re:Us coders are delaying the Singularity!

[joss]

Are you being deliberately stupid or what ?

Some things are easy to parallelise, a lot of things arent. Processing an image.. fine.. give each processor a chunk.. but wait.. you get edge artifacts since each pixel needs neighbour pixel information.. that has to be shared.. at what point does it take *longer* using multithreads.. doing a large matrix inversion that takes hours.. ok, parallelise it.. but wait.. standard algorithms that you can see in numerical recipes in C assume single thread.. u need a completely new algorithm, and u need a completely new mechanism for stabilisation bla bla bla.. and this is before u even get into resource contention etc etc.. These are not compiler issues, they are design issues, some of which are at least as hard to solve as writing the entire single thread program. Until computers are smarter than humans, they are not going to get solved automatically.

Re:Us coders are delaying the Singularity!

[Kjella]

Why is it so hard to get developers to write decent multi-threaded code? It's not that hard

Trivial use of threads is easy. Complex use of threads where a workload is spread over several CPUs is hard. I use threads in my software but they mostly go like "main thread" -> "work thread w/queue" -> "notify main thread". It keeps the application responsive, but it's trivial - the CPU load is 1% main thread, 99% work thread. Now imagine instead I want to introduce a number of worker threads. If they are all truly independent (no common resources, no dependencies) it is still easy.

The moment you start introducing dependencies (beyond a queue) and shared resources, things get ugly really really fast. It is so easy to lock yourself into a state where none of them can't be met, where locks aren't released. Queues are simple because you do "lock queue, add to queue, unlock queue" and "lock queue, copy item to private data, unlock queue". No locks need to be held beyond the scope of those two trivial functions. In fact, I use a mutex that'll unlock when it goes out of scope, so it's not possible to screw up (apart from forgetting the mutex altogether). But once you start dealing with lock states.... well it's *NOT* trivial. And that's what you need to really distribute any work that's not trivially parallel.

Re:Us coders are delaying the Singularity!

[joto]

Dude, have you written any multi-threaded code? It really isn't that hard, regardless of how smart I may or may not be.

Dude, have you ever written some multi-threaded code that

1. Wasn't slower on a single-processor than properly optimized single-threaded code would be?
2. Wasn't more complicated to understand than when implemented straight-forward as a single-threaded program?
3. Wasn't obviously parallelizable, at a level that fits present-days multithreaded processors?

The JIT compilers in Java and .Net/Mono take care of many future hardware problems, since the JIT can optimize your code for processors that didn't even exist when you wrote it.

So can any other compiler or interpreter. They are a way of abstracting the details of actual computer hardware into more abstract concepts such as if-statements or member functions. In that respect, it isn't anything fundamentally different between a JIT and any of the other common ways of implenting computer languages, they all "take care of future hardware problems".

The reason this works, however, is that while computing hardware varies, these variations has always been minor variations of more or less the same architecture. This architecture is called a von Neumann architecture. Now, a parallel architecture is by definition not a von Neumann architecture. So, once you decide to write parallel computer programs, you must also decide on a parallel architecture to run your programs on!

And while you may think that should be multicore cpus, other people think differently, and want to programs SIMD computers, transputers, scientific clusters, database clusters, cell processors, programmable chips, etc... No matter how great you think java is, it won't help make programs efficient across these different architectures, unless you can abstract it all away into one common architecture, like the von Neumann architecture has done for normal computers.

Re:Us coders are delaying the Singularity!

[rufty_tufty]

Easy there tiger!
I always used to argue this same point myself! Still do many times. I write massivly parallel code all day (I'm an ASIC engineer coding in verilog). I never saw what these softies were complaining about when they said parallel code was hard.
Then a friend told me about the problems he was having in visual studio writing multithreaded code, so I tried it out myself. Simply put, the current state of tools sucks!

Let me explain: Suppose I'm daft enough to use visual studio. Suppose I'm writing a multithreaded application and I've got a data structure being corrupted. In my verilog simulator this is easy to debug - there are very strict rules in the language about how variables get driven and the simulator is able to trace the multiple drivers of any net.
In visual studio however, anyone can write to a data structure (if you get things wrong). Not only that, but the debugger isn;t fully aware of the concept of parallelism, so won't break all of the threads properly - they may even keep runnign while you're debugging them.
So how do i trace who is stomping on my data structure if it's changing while I'm debigging it?
If the language doesn't support proper isolation between processes, or clean message passing?
Yes for small applications this is less of an issue, but I can imagine in a cruft infested bit of code, without the tools to help you debug, a language that is designed to be inherenltly parallel (and not to just have it bolted on like C or java) then I don't see how the development can occur except with the grossest levels or parralisation.

Speaking as someone looking from the outside a few things need to happen to software to make paralleism take off:

1) A language with the ubiquity of C, but with parallism handles similar to occam needs become mainstream.
2) Operating systems need to be updated with the new reality
3) Development tools need supporting to support all the above
4) Developer mindset needs to be changed, particularly with an eye to software architecture.

Only when all of the above has occured do I think we'll see what you say in your post.
TBH I think with a multicore CPU on every desk we may see an equivalent to the dotcom boom again - Not sure if I want that though...

Re:Us coders are delaying the Singularity!

[Tsagadai]

I must say you hit the nail on the head. With current levels of knowledge in parrellelism this is just going to be another dead end for speed gain. What we really need is alot more thought and research into this topic (again) and alot more training for programmers.

Re:Us coders are delaying the Singularity!

[zhiwenchong]

Incidentally, what do you think of microthreads, ie. found in Stackless Python and possibly concurrent languages like Erlang?
Quote: "It allows programmers to reap the benefits of thread-based programming without the performance and complexity problems associated with conventional threads" As I understand it, microthreads are more lightweight than threads and far more manageable.

Re:Us coders are delaying the Singularity!

[TheSunborn]

The solution is to use multiple processes insted of multiple threads. Then you have to use explicit message passing to exchange data. (If you are working on large dataset, you might want to allocate the memmory in shared memmory, and then just use messages to cordination and communication.

Martin

Re:Us coders are delaying the Singularity!

[mpcooke3]

That's just one way of defining performance and it's not particularly useful for most real world applications.

When most people i know (developers and customers) talk about performance they are not referring to how many 8x8 DCT blocks the CPU can inverse-transform in a second. A lot of real world tasks involve IO operations and in some cases using multiple threads on a single CPU/Core can result in completing the task faster or improve the number of tasks per second that can be completed.

Re:Us coders are delaying the Singularity!

[ardor]

The person's work performance improves because of multithreading, yes. But the person is usually not included in this equation. For the person, the computer is a black box. This thread is about the black box, not about the things outside using it. And in the blackbox, you improved responsiveness, not performance. The file manager reacts even though its scanning thousands of files? Improved responsiveness.

Re:Us coders are delaying the Singularity!

[PaulModz]

I'm not trying to be difficult, guys. I've been writing multi-threaded code 12 hours a day in C# and Visual Studio for the last few years, so its hard for me see what all the fuss is about.

C# has some nice language features like the lock and volatile keywords, and the .Net framework includes classes like the ThreadPool, Monitor and ReaderWriterLock which go a long way towards elminating most of the issues ruffy mentioned.

As far as debugging the code, I haven't seen a debugger that handles multi-threaded code really well. The best way to debug MT code is good logging, along with a GUI that can manage and filter the logs.

The project I'm currently working on is a game built on top of a Solar System simulator. The rendering code and the physics code each run in seperate threads, which allows me to adjust the speed of each independant of the other. This is helpful because adjusting the size of timesteps in the physics code is dicey, and the best way to control the sim rate is to adjust the rate at which the physics thread runs. Maintaining a constant physics rate is more important than a constant render rate, so the render thread checks CPU usage and adjusts its speed accordingly. Yes, this code runs a little slower on single procs, but I track the lock contention and Context Switch rates in perfmon and it isn't bad, the flexibilty is more important.

Here's a link to a demo of the prototype if anyone wants to see it http://video.google.com/videoplay?docid=-789042178 3974814380&hl=en (I have a high res version on my own server if anyone cares.) Unlike Celestia, all the orbits are calculated in real time, and I still had enough bandwidth leftover to write @ 25 jpgs per second to the hard drive to create the demo. Besides showing off, I just wanted to point out that I'm not delusional or clueless (for the most part)

Re:Us coders are delaying the Singularity!

[addaon]

So how does your code take advantage of a quad core processor if it has two threads? How about a 32-hardware-thread system, hierarchical, with separate processors, cores with shared L2, and hyperthreads with shared L1?

Re:Us coders are delaying the Singularity!

[Sebastopol]

The problem with threads:

http://www.computer.org/portal/site/computer/menui tem.5d61c1d591162e4b0ef1bd108bcd45f3/index.jsp?&pN ame=computer_level1_article&TheCat=1005&path=compu ter/homepage/0506&file=cover.xml&xsl=article.xsl

Basically, even the simplest tasks require significant armor plating to run correctly.

OTOH: Multi-PROCESS programming is far simpler than multi-THREADED programming.

Re:Us coders are delaying the Singularity!

[PaulModz]

Good point.

One of my current tasks is moving all the civilization calculations (this is a 4X game similar to GalCiv2 but in realtime) into a third thread instead of piggybacking on every 10th physics frame. If I had enough CPUs to prevent context switching, each AI civilization could have its own thread, but that's only 8 more threads tops. Hmm, 22 processors left ;)

Right now my game can only process one Solar System at a time, so each one is more like a new level that you access by moving on the galactic map once you have gathered sufficent tech/resources. If I had 15 or 20 extra CPUs sitting around, I could run a different Solar System on each one, making the game much more open and dynamic.

This nice thing about this is that the speed of light (speed of information) allows me to limit the amount of interaction between systems. If we really are living in some sort of Matrix, god the programmer must have added the speed of light as an optimization for our multiprocessing substrate! ;)

Re:Us coders are delaying the Singularity!

[mpcooke3]

I feel I am repeating myself somewhat here, but anyway...

The performance of a single thread on single core/CPU system can sometimes be poor due to IO wait, this includes threads used for batch jobs and for applications with no interative GUI. Frequently for example a thread will be blocking on Disk or Network IO. Relatively few tasks have no IO and normally the IO operations are slow compared to the CPU speed. Therefore, as the original comment stated the performance of some applications can be improved by making the application multithreaded even on single core/CPU.

Of course threads are often used in GUI code to improve responsiveness too.

Re:Us coders are delaying the Singularity!

[joto]

Incidentally, what do you think of microthreads, ie. found in Stackless Python

As far as I'm aware, stackless python "threads" aren't threads, but is really what the rest of us calls coroutines. I like coroutines, as they are predictable and easy to understand, and gives a lot of (useful) power to the programmer. Now, it might be true that later versions (such as the current) has added some scheduling beyond manual yield(), in which case you would call them userspace threads. Personally, I prefer system threads, coroutines, or both, and think of userspace threads as a bastardization of the concept. But if people find user threads useful, I'm not going to argue with them. To each his own...

and possibly concurrent languages like Erlang?

I have no experience with Erlang. From what I understand, erlang threads are userspace threads and are not used for efficiency, they are used for ease of modelling. And message-passing is the way these threads communicate. This is completely unfamiliar territory for me, but intelligent people I respect like Erlang, and projects in Erlang has a good track record. As a result Erlang exists somewhere on my got-to-learn-that-sometime-in-the-future list. Unfortunately nobody has time for every interesting thing out there.

As I understand it, microthreads are more lightweight than threads and far more manageable.

As I understand it, "micro"-threads are really user-space threads. This will not give any benefits for parallel processing on machines with more than one execution unit. Even though you have structured your program for multithreading, it will still execute on just one of your CPUs or cores (Erlang might be an exception to that, as I believe Erlang is also good for distributed programs, meaning objects/threads/whatever can live in different address spaces.) On the other hand, since the majorithy of computers used to have just one execution unit (cpu core), this also used to mean that threads would be more lightweight and there could be less overhead.

Re:Us coders are delaying the Singularity!

[zhiwenchong]

Thanks for your thoughts!

Re:Intel Fails

[DDLKermit007]

Intel is at 65nm manufacturing and has it in the works to go smaller. While AMD is still at 90nm manufacturing and still trieing to get to 65nm. I prefer AMD (gota root for the underdog), but this is a clear situation Intel is releasing better product. Intel's Core Duo stuff is coming at just the right time. Hell they got the Apple contract at the perfect time. Neither IBM or AMD are anywhere near the heat & power consumption Intel is at right now with those chips.

... Next page....

[KillzoneNET]

After

... Next page ...

getting up

... Next page "Interested in..." ...

to the first

... Next page "No please..." ...

page of

... Next page "Buy our offer..." ...

the actual

... Next page "We are desperate..." ...

charts,

... Next page "Buy now! Limited time offer!" ...

I lost what

... Next page "SPECIAL OFFER!" ...

the art

... Next page "Brought to you by Fixed Random Ads Inc" ...

icle was about...

WAIT!! for quad core to mature

[atarione]

these things use too much power... idle the quad core uses more power than a core 2 duo at idle...

i say wait a gen or two for more energy effiecent quadcores... (i'm asking cause i do not know...) but is it feasable for a multicore cpu to be designed to shut off extra cores when idle as to minimilize it's engergy consumption?

honestly i keep waiting for something to really "blow me away" and maybe it is the things i've been using the pc for ...but my A64 3200XP still seems just fine... for the momment.

Never ask that question again.

[Anonymous Coward]

"The question is, do we need this much processing grunt?"

OK, a lot of people have asked the same thing and I'm not picking on you. Hell, we all think it and wonder it but the answer is so crystal clear no one should ever ask it again.

I wish my 1 MHz Commodore 64 was a little bit faster. I would have paid good money for an extra MHz or two. If you would have told me then that one day computers would be 100 times faster I would have been shaken but not stirred. It'd be hard to imagine computers being that much faster than my 64 and it would sort of seem wasteful, but we're talking future here so sure, why not.

If you'd told me that computers in 20 years would be 3000 times faster than my good old 64, well I imagine I would have had a hard time wrapping the brain around that one. I'm sure I'd be asking what would be the point? How could you possibly use 3000 MHz?

Of course we all know how now. So, if you were to tell me today that next year computers are going to be 10 million times faster than today, I'd say cool.

Exactly

[joss]

I grew up programming transputer clusters cos I figured Moore's law wqould have to slow down sometime and then we would have to move to multiprocessor systems. Efficiently using more than a couple of cores is *not* easy.. and it opens up a whole realm of interesting algorithmic work where basic problems with established solutions suddenly become open again.

Its about fucking time...

Coolies

[tomstdenis]

Still a hack, not complete, but coolies.

Of course, AMD quad-core will actually "fully work" when shipped ... hehehee /me shakes fists.

Tom

Don't worry too much about CoD2

[spyrochaete]

For those who love Call of Duty 2, don't be too apprehensive about stepping up to multicore CPUs. The game didn't work in Tom's benchmark roundup, and it gave me trouble as well with my Athlon 64 X2. The problems were odd too - the game looks and sounds just fine, and everything works great on the surface. However, the grenade tutorial in the first mission cannot be completed, and you don't heal at all throughout a level! The latest patch fixes these problems and the game plays beautifully.

Re:Don't worry too much about CoD2

[denjin]

Plus, that bug was just with AMD Dual Core CPUs. It doesn't exhibit these problems on Intel ones (I know, since a Pentium D 830 system I played the game on was fine).

Re:Don't worry too much about CoD2

[Mistah Blue]

Concur with the latest patch for CoD 2. I run CoD 2 without any issue on a Lenovo Thinkpad T60p (which is dual core). I also run Falcon 4.0 Allied Force (latest patch) on this same laptop with no issue.

How long will it be ...

[LaughingCoder]

before Microsoft starts charging more for multi-core installations? Seriously, if quad core means fewer boxes in the rack, it means fewer licenses.

Re:How long will it be ...

[merlin_jim]

before Microsoft starts charging more for multi-core installations? Seriously, if quad core means fewer boxes in the rack, it means fewer licenses.

Ummm try a few years ago? We have a fair size data center here and most of our Microsoft server software beyond the O/S itself is licensed per core, not per computer. 4-core box = run on one proc or buy four licenses.

Re:How long will it be ...

[suggsjc]

Probably about the time the adoption rate of these pick up. They'll wait till people drink the kool-aid, then move in...

Re:How long will it be ...

[ohearn]

Right now Microsoft still needs to gain a lot more marketshare in the server market before they can really do that without driving away customers because of licensing issues. Now if M$ ever gets even half the marketshare with servers that they have in the desktop world, then just prepare to get raped by licensing issues.

Re:How long will it be ...

[LaughingCoder]

... most of our Microsoft server software beyond the O/S itself is licensed per core

Well, I guess the handwriting is on the wall then. It's probably only a matter of time before the O/S licensing is by core also.

Re:How long will it be ...

[LordEd]

What software is asking for extra licenses? Microsoft has explicitly said they are going on a per-processor model:

On October 19, 2004, Microsoft announced that its server software that is currently licensed on a per-processor model will continue to be licensed on a per-processor, and not on a per-core, model. This policy will allow customers to recognize more performance and power from Microsoft software on a multicore processor system without incurring additional software licensing fees.

Re:How long will it be ...

[merlin_jim]

I think I need to talk with some of our licensing folks - I think in some cases we've licensed per core and not per proc...

Re:How long will it be ...

[MobileTatsu-NJG]

"before Microsoft starts charging more for multi-core installations? Seriously, if quad core means fewer boxes in the rack, it means fewer licenses."

Microsoft had this problem before multi-core. Just having faster CPUs every year means fewer boxes on the rack. Every time they make a move like this they open the door wider for OSS. They know this.

Microsoft's Multicore licensing policy

[LordEd]

Google says...

On October 19, 2004, Microsoft announced that its server software that is currently licensed on a per-processor model will continue to be licensed on a per-processor, and not on a per-core, model. This policy will allow customers to recognize more performance and power from Microsoft software on a multicore processor system without incurring additional software licensing fees.

Mod parent up...

[Kjella]

...and depending on your sense of poetic justice, grandparent down

Carmack had no choice ...

[AHumbleOpinion]

Even John Carmack who used to be very much against multi-cores for gamins recently elaborated much on this area in his keynote. Practically any modern (lets call it nextgen :D) gaming platform is now multi-core.

Carmack had no choice, even on PCs. Both Intel and AMD have been telling Game Developer Conference attendees for at least two years that future performance improvements are going to be largely based upon multiple cores not clock rate. He could accept that his codebase was "obsolete"(*) and needed to be rewritten for threading/multi-core or his future products could be viewed as defficient when compared to others.

(*) In some ways Id's code obsoletes more quickly than others. Keep in mind that their business is more about licensing engines to other game companies than developing games themselves. This is why their games can be so bleeding edge and have such steep hardware requirements. They are not selling to the sweet spot of what today's gamers have, they are selling to developers that will not be releasing a game for a couple of years. By then the sweet spot has caught up. However they have to stay far ahead of it.

Re:Carmack had no choice ...

[tbradshaw]

That's a very common misnomer, about id Software being more about licensing engines than building games on them. Though, I can say first hand that it's patently false. It's more of a rationalization from gamers that were often id fanboys when they became disenchanted with id games but still impressed by their technology. (I'm not in that group, I've really enjoyed the id games for their "purity of vision" and still consider all id games among my very favorites.)

id Software focuses on their games with a narrow lense to do what they exactly what they want to do, the best that they can with the technology that John Carmack puts together. Their technology is so ahead of the curve because John Carmack is so ahead of the curve. Though, the id guys and Carmack specifically have taken a strong look at software development practices and with the "new game" (title and details unreleased), there's a very strong effort to not get ahead of the design process and have the technology steer the course--instead the game design will steer the course. (Often in previous games a sweet advancement would come along that would kind of be shoehorned into a release, because it was worth it. The restraint is new. Megatexturing is a new tech that is only going into the games that need it. However, the sweet wavy look from heat sources, that was actually added to Doom 3 at the last second because it was sweet.)

Even more importantly, if one is carefully following the company, one has to notice that lately id has chosen to form partnerships to continue development on their important franchises rather than simply licensing them away. Return to Castle Wolfenstein, Quake 4, and Enemy Territory: Quake Wars, these are all games that have been directed by id but developed by different teams (Splash Damage, Raven, etc).

The particular view that id Software is about licensing engines and not making games, it usually results in a +5 mod around here, but it's not true (and has nothing more than conjecture to back it up). The intentions of the guys at id are to make great games. The fact that Carmack's engines are so extremely valuable means licensing will always be a part of the business plan, but that doesn't supercede the desire of Carmack or the rest of the talented id team to make great games.

Re:Carmack had no choice ...

[AHumbleOpinion]

Those are some powerful arguments but wouldn't you need a revenue breakdown of licenses vs retail sales, not game development philosophy to say what the focus is? Sure programmers and artists are *thinking* games, not licenses, but what is paying the bills? Regarding the bit about id increasing it's focus on it franchises, could that be a strategy dictated by increased competition on the engine licensing side?

Re:Carmack had no choice ...

[tbradshaw]

I don't think that a revenue breakdown is necessary, since intent is determined by the developing team's personal intention rather than the bottom line. And the team has made it pretty clear that they put their games first, their licensing second.

Although, I think the pricing for the licensing makes things pretty definitive. A Q3 engine license was $150,000 plus 3% royalties (at peak). While a good chunk of change, few (if any) Q3 engine games so much more than Q3 that they could overcome the (only) 3% royalties. Older engines were licensed for less. The missing variable is that I don't know Activision's cut and/or id's cut per gross sales of id Software's games... but now that it's came up that would be an interesting question. I've just never brought it up in conversation before. Next time I talk with Todd or Marty I'll ask, I'm sort of curious now. (I wonder how it compares to like Artist revenue from CD sales and junk like that.)

Anyway, I would contest that developing team intention trumps revenues, at least in id Software's rare case... after all, they don't "need" the money like other development houses do. They have the fortunate situation of being independantly wealthy with exceptional talent/leverage. Of course, the freedom they enjoy is somewhat dependant on continued excellent sales, but it's certainly not the case that they need to sacrifice the game (what they truely enjoy doing) for the technology (which is a means to an end).

This fiscal independance is also the primary reason for continuing Mac and (especially) Linux support. John sees it as the Right Thing to do, and they are able to stand by that philosophy rather than going purely by the numbers. (Linux sales for q3 were an absolute flop, heh)

Re:Carmack had no choice ...

[AHumbleOpinion]

Regarding Linux. Many years ago I believe someone was quoted as saying it does not make business sense, that they just do it because they think it is cool to do so.

Getting By On 4x3GHz a Day

[Doc Ruby]

"If we take our benchmarks into consideration you can no longer get by without a quad-core processor."

I love HW geeks. This CPU isn't even released, many apps don't even work on it, chipset makers are still scrambling to release the mobos that will actually use it. But even though nearly no one can get one, no one can get by without one.

If HW geeks weren't so demanding, Intel and its competition wouldn't do so much work keeping them packed with the hottest toys.

Re:Of course FOSS ppl will not use those

[wirelessbuzzers]

Niagara is actually very nice if you have a massively parallel (~ 32 threads always on the CPU) task with no floating point operations. So it shines for enormous webservers, or servers combining every tier on one machine. Unfortunately, there's only one FPU on the whole chip, so floating point pretty much kills the performance.

Also, for anything running single-threaded, the performance is terrible. The processor runs at 1-1.2 GHz and isn't superscalar.

The combination of these factors means that not too many people will be buying SunFire T2000 servers at $13k apiece.

Is this not ment for servers?

[edxwelch]

They ask a very good question, do we really need 4 cores? Well, no at least not for games and the benchmarks confirm that one fairly good. The chip actually performs slower than the conroe extreem (discounting the overclocking results).
It would seem to me that this chip would better used in a server that would natually make more use of the extra threads.

I wonder if I can a Mactel Core 2 Duo

[Jabrwock]

And replace it with this?

Early?

[Tmack]

Nah, they are still playing catch-up. SUN has already been selling their Niagara (Ultra Sparc T1), which has 6-8 cores each, with an additional 4 threads per core, and the ability to have 4 in a box (8*4*4 = 128 threads!) blows Intel out of the water. Yes, I am perfectly aware that this chip is a SERVER cpu, but you mention servers and virtualization (vmware, parallels, etc), which Sun basically designed this chip to do. As a sysadmin, this cpu also get my attention over Intel for using very little power, something Intel isn't exactly known for. Less power == less heat generally, and with the number of cores the T1 has, we can run multiple VMware instances on one box instead of several 1u servers or blades. The drawbacks to the T1 are its relative lower speed (1.2Ghz, tho it is a Sparc based CPU), and a single shared FPU per cpu, but for general purpose servers, the T1 is it. For our heavyweight servers, the ones that need the most CPU horsepower regardless of heat/power consumption, we still pick AMD opteron based Suns (Sun Fire x4600). The memory bus to the Cores alone far surpasses Intel when it comes to throughput, which directly translates to faster operation with bigger datasets, not to mention that you can get an 8xdual core 3Ghz Opteron setup...

AMD beat Intel to the dual core CPU, and will be releasing a 4core opteron soon as well. IBM's Cell cpus are also multi-core, though in a different arrangement, and will be used in the ps3 and IBM's next megasupercomputer.... Multi-core itself isnt too special, it just takes multi-cpu and puts them in the same physical chip to save space/heat/power. The advantages for programmers/gamers/etc (neglecting the environmentals like heat/power) are basically the same as if each core were a different physical chip. Intel got behind, and is now rushing to one-up AMD. The same old CPU war rages on

tm