AMD Demos Dual-Core Athlon 64

DigitumDei writes "Dual core chips came closer to reality as AMD demonstrated their Athlon64 dual-core offering. The 90nm technology chip will use the same 939-pin infrastructure and cooling solutions as the current Athlon 64 chips, meaning that upgrading to a dual-core chip from your current AMD64 will require little more than a BIOS update. Available in the second half of this year, the chip will be added to AMD's current line (Athlon64, Athlon FX, Sempron)."

Comments

Am I Missing Something?

[Dragoon412]

I don't understand the hype about dual core CPUs.

As I understand it, they work almost identically to a SMP setup, meaning they don't offer much of any performance benefit in most apps (particularly games). They draw more power, they run at higher temperatures, etc.

Is there something I'm missing? Or is this whole dual-core mess really just SMP on one CPU? Because from what I've read on the likes of Extremetech, Anandtech, and so on, I'm not finding any reason to be impressed.

Re:Am I Missing Something?

[DrXym]

they don't offer much of any performance benefit in most apps

And how many apps & other processes is your system running at the moment? Mine's running 58 with 518 threads.

Re:Am I Missing Something?

[teg]

I don't understand the hype about dual core CPUs.

As I understand it, they work almost identically to a SMP setup, meaning they don't offer much of any performance benefit in most apps (particularly games). They draw more power, they run at higher temperatures, etc.

SMP without the mess (extra CPUs, cooling, expensive/complicated motherboards) and cost is definitely something to be impressed about.

It should give a big performance boost to a multi app and multi thread environment.

Re:Am I Missing Something?

[forkazoo]

Yup, pretty much. But, some of us do more than play games. Also, as multiprocessing hardware becomes more common, game makers will begin to take advantage of the benefits. For me personally, when I want to use my box for general-purpose stuff, and it is running the mythtv backand and transcoding some files into MPEG4, and I am rendering a 3D animation, and so on... Well, having SMP sure isn't a bad thing!

Re:Am I Missing Something?

[drwtsn32]

Sure, benchmarking a single app on an SMP system often makes little to no performance difference, but SMP is fantastic if you are a heavy multitasker and work with several apps at once.

My first SMP system was a dual Pentium 133Mhz box. After that I never went back to a single proc... until the Pentium 4 came out. It's disappointing that this chip does not support SMP (except for the Xeon line). P4 hyperthreading helped bring back some SMP goodness, but it's still not as good as two real chips.

Personally I can't wait for dual core CPUs!

Re:Am I Missing Something?

[i41Overlord]

As I understand it, they work almost identically to a SMP setup, meaning they don't offer much of any performance benefit in most apps (particularly games). They draw more power, they run at higher temperatures, etc.

The reason most games don't get a performance boost from dual CPU's is because they aren't programmed to take advantage of the other CPU. How many end-users' home systems have dual CPU's? Hardly any of them. There was no reason for game makers to go through the effort programming for something that 99.99% of their customers can't use.

With the new dual core chips, technically it isn't anything groundbreaking but it will ensure that there's much more widespread adoption of multiprocessor systems. With more of the userbase using dual core CPU's, game makers will have a reason to program to take advantage of it, and you'll begin to see games that do see a performance increase when using dual CPU's (or dual cpu cores).

Re:Am I Missing Something?

[Junta]

The big deal is that it is a way to get *cheap* SMP, motherboard sockets x2, the downsides being cooling such densities (read: reduced clock per core), and sharing a memory controller per two cores (which is what Intel SMP has been doing forever, AMD used memory controllers per processor in a NUMA fashion full time, and Hyper Transport to access memory not associated with the current processor).

Theoretically, the dual core clocks will add up to more cycles overall than a single core, but the single core will have more clocks per individual thread, so unless a game leverages threading very nicely in the processor intensive segments, a multi-core may be slower than a single-core for the high-end gaming scene, however for workstation/server/HPC fields, it is very exciting.

Re:Am I Missing Something?

[Shalda]

Effectively it's SMP on one CPU. Something I very much look forward to as I always build my desktops SMP. It also incorporates some of the overhead of SMP into the CPU driving down the system price a little bit.

While you're right that SMP offers little performance to most apps, I tend to run a lot of CPU hogs at the same time. Watch a DVD while waiting for a project to finish compiling or whatnot. It can also help keep runaway processes from sabatoging your system. I used to have a program that set its priority to 'AboveNormal' and would from time to time it would hang up in a loop. Since it was running at a higher priority, you sometimes couldn't bring up Task Manager to kill it off as the higher priorty thread always took precedant. And if all else fails, you can set up 2 SETI @ Home clients and process twice as many packets. But do your self a favor and set their priorty to 'Low'.

Also, you've got a chicken and egg problem. The reason there's so few programs that benifit from SMP is that there are so few computers that are SMP. When I was running some computer labs at a Big 10 university a few years back I was insisting on SMP workstations so that the CS students could learn to program multithreaded apps and see the benifits of it when it ran on 1 vs. 2 processors.

Lastly, my basement is very poorly insulated and gets a bit chilly in the winter. Anything to help warm it up and keep my fingers working properly is a good thing(tm)!

Re:Am I Missing Something?

[magarity]

they don't offer much of any performance benefit in most apps (particularly games)

Ah, rejoice irresponsible youth! Right now a $25,000 8 CPU machine that can no longer keep up with a decent sized corporate database needs to be replaced with a $60,000 16-CPU machine. After dual core hits the market, it can be upgraded for the price of 8 new dual-core CPUs and a BIOS flash. Less money for hardware == more money for bonuses... W00T! W00T!

Re:Am I Missing Something?

[SpongeBobLinuxPants]

meaning they don't offer much of any performance benefit in most apps (particularly games). They draw more power, they run at higher temperatures, etc.

from TFA:

For example, a processor with dual 2.0-GHz cores can deliver performance not all that different from a single-core 3.5-GHz part. More important, such a dual-core part will hold down power dissipation to a figure closer to that of a standalone 2.0-GHz CPU, allowing processing throughput to effectively double for not much more power.

and

At such speeds, single-CPU processors can often dissipate more than 150 W.
The dual-core Athlon 64 runs at a clock-speed of 2.4 GHz and has a maximum power dissipation of 100 W.

Re:Am I Missing Something?

[lsmeg]

Dual core is basically SMP as far as I know, so you're not missing anything, except maybe some of the benefits of running SMP. From my experience, SMP can go a long way to making a system feel less "sluggish". Tasks that will essentially lock up a single processor system can run in the background without making a system useless on SMP.

Now this doesn't usually add up enough to warrant a normal user to spend the extra money on SMP, but if dual cores become the replacement for the desktop line, they should (eventually?) be more affordable than SMP. Plus, if dual cores become ubiquitous, more and more software will no doubt be written to take advantage of SMP.

I can't see any downside to the push for dual cores, as we seem to be slowing down in how far we can push a single processor in clockspeed and performance.

Re:Am I Missing Something?

[dsginter]

And how many apps & other processes is your system running at the moment? Mine's running 58 with 518 threads.

But what's the processor utilization? On most systems, its usually less than 10 percent. So when a user does something, the bottleneck is usually not the processor. Its usually the hard drive.

Money would be better spent on RAID, rather than dual core or dual processor.

Re:Am I Missing Something?

[roach2002]

Some instructions on the x86 hardware have a delay; let's say I'm asking for something from ram or cache. This takes more than one cycle (3 from cache if I remember correctly). On a normal, non hyperthreading processor, the processor sits idle until that memory value comes back.

On a hyperthreading processor, the processor can do an instruction on another process in that time. So a second process can come in and do a couple ADDLs.

It may not make your game run any faster, but if you have something in the background that can complete what it's doing more quickly.

SMP with n processors can theoretically run n instructions in one cycle. Hyperthreading still limits you to one instruction per cycle per processor, but it reduces the number of cycles where the processor does nothing because it's waiting on other hardware.

So for you, it probably isn't worth it. But for some workloads, ie: some servers that are constantly delaying on memory, it can speed secondary processes up

Re:Am I Missing Something?

[oldmanmtn]

As I understand it, they work almost identically to a SMP setup, meaning they don't offer much of any performance benefit in most apps (particularly games).

I suppose if your idea of "most apps" is games, then this probably isn't an area that would be of interest to you

If you have any multithreaded app that is even remotely competently written, then it will benefit from dual cores or (possibly) hyperthreading. If your multithreaded app is full of "big locks", then dual cores won't help, and the application designer is a failure.

If you have a workload that has multiple processes running simultaneously, then it is also likely to benefit from dual core. It gets more interesting with business/server workloads, but "home users" can benefit two. Even something as simple as running xmms and gcc at the same time should go faster. Or running two instances of lame.

The real win with dual core comes from increased throughput. A single job/application/process isn't likely to go any faster, but a full workload of multiple, reasonably parallelizeable, tasks will be faster.

Re:Am I Missing Something?

[Waffle Iron]

And how many apps & other processes is your system running at the moment? Mine's running 58 with 518 threads.

Typically 517 of those threads are asleep waiting for IO or a signal, and the one piece of information that you are currently waiting for is being processed in the single remaining active thread.

Re:Am I Missing Something?

[drinkypoo]

You can use an affinity tool to put your process on a single processor alone, and everything else on the other. This all but eliminates context switches in your program's context, while all other processes can continue to run on the other core(s). The only thing that consumes a lot of CPU that is not typically multithreaded is game software. Eventually, this too will be multithreaded, as we see more multi-core systems out there.

Re:Am I Missing Something?

[TheRaven64]

I only do two things that really tax my CPU. Compiling and video editing. Compiling is embarrassingly parallel, and make programs (including GNU make) have been able to take advantage of this for ages. Generally, the best performance can be achieved by running make with number of CPUs + 1 way parallelism. Video editing is similarly parallel, since most CPU intensive things are effects that need to be applied to a large number of frames, making it trivial to split the workload. I would certainly see a large performance benefit from SMP.

Before I abandoned the desktop in favour of the laptop, I had an SMP system, and it was nicer to use than my faster UP system, since single-threaded computationally expensive things could be run on one CPU leaving the other one free for UI-related tasks.

Re:Am I Missing Something?

[c.r.o.c.o]

SMP without the mess (extra CPUs, cooling, expensive/complicated motherboards) and cost is definitely something to be impressed about.

The motherboards supporting dual core CPUs should be identical to those running single core CPUs. I guess this is where having the memory controller integrated into the CPU really pays off for AMD since it further simplifies mb design. But in the past SMP motherboards weren't THAT much more expensive (at the most $100 extra) than similar single CPU motherboards. The main cost associated with SMP setups were the very expensive SMP CPUs, which were anywhere from 1.5 times or more expensive than regular CPUs. The pricing of dual core CPUs remains to be seen, but I think it'll still be cheaper than 2 separate SMP-enabled CPUs.

However I completely agree with the rest of your post. Not having separate heatsinks, large motherboards, etc is a definite advantage. Just because of that the market acceptance will increase very rapidly. I wouldn't be surprised if games and other CPU intensive apps started supporting dual core CPUs soon.

Re:Am I Missing Something?

[stecoop]

Does the software you're running take advantage of the dual cores for multithreading though>

So Intel is doing something crazy by making dual core processors because application haven't had to think about multiprocessors right?

Well, change is inevitable and developers can't stick their head into sand and stay that way forever. Intel recognizes that change has to occur in the development community to enhance performance their product line. What better way then introducing dual cores? This will force programmers to start thinking about programming their application with multi - user, threading, layer, etc' thus over time, the application will be better utilized for the future.

Re:Am I Missing Something?

[Anonymous Coward]

Yes, you are missing something. Many of us use computers for more than games. Java applications for instance are nearly always threaded and benefit greatly from multiple CPUs (at least on Linux, I can't speak for Windows). Or how about working on a PowerPoint presentation while your spread sheet recalc's. Most any app that has a GUI will be more responsive if other activities are going on as well.

Whether the price difference is worth it is another question. Also you can expect many threaded apps to break. If you haven't debugged a threaded app on a multiple CPU system, then you haven't debugged the threaded app. And how many developers have multiple cpu systems laying around?

Re:Am I Missing Something?

[drinkypoo]

I trust you are aware that you could have gone dual AMD? And that in many cases HT actually results in poorer performance?

Re:Am I Missing Something?

[Vaystrem]

I've had the pleasure of using a number of friends Dual CPU systems across the years.

The most relevant one to this discussion is my friends Abit BP6, Dual Celeron, setup. At the time he was running Dual Celeron 300s. Not that impressive right? Except that he was able to host our Unreal Tournament Server - and then join it with no lag for any of the players. Running Unreal Tournament by himself showed a 50% load on each processor. He was able encode MP3s, burn CDs, and play games simultaneously. Something I was not able to do, and wanted to for the sake of time saving. I did not want to choose one activity over another - a 'leisure' productivity issue if you will.

Fast forward to now. A nice dual cpu system would allow me to play games, encode movies / my audio files, simultaneously while running Distributed.net. Are dual core's absolutely necessary? No, but when you are doing some very intensive applications you can't do anything else. As well even though not all games are multi-threaded, various aspects of the game may be, the networking code/the sound code etc, meaning that the game may run somewhat better on a dual cpu setup.

Re:Am I Missing Something?

[plague3106]

You know, I've been hearing this alot on this thread, and I don't understand the thinking. You do NOT need threaded apps to take advantage of SMP. A SINGLE app will run faster (potentially) on SMP if it is threaded, but running a SINGLE application isn't the big benefit SMP gets you.

Its that Firefox can be run on one processor, while your MP3 player is running at the same time on the other. This in turn will speed up BOTH applications, since Firefox does not ever have to yield to the player, and visa versa.

Since there's more going on then just those two apps (various system process, etc), your machine should be faster as each process now only has to worry about HALF the number of processes it did before.

I don't understand why a largely tech audience misses that point. We're not in DOS anymore; the OSes we are using all run more then one process at a time.

Re:Am I Missing Something?

[ivan256]

You can't really take TFA seriously, since it's InformationWeek, and not a reputable source. Just look at your first quote. There are plenty of 2.0Ghz single core chips out there that can deliver beter performance than a 3.5Ghz P4...

When they talk about single core chips using 150 watts, they're talking about intel chips (almost certainly the Itanium). 100 Watts for a dual core 2.4 Ghz Opteron is a 60% increase in power usage over the single core version.

Of course the parent to your post is just plain wrong about not offering a performance benefit for most apps. Now that developers are starting to try and use Hyperthreading to their advantage, those apps will see some serious benefit from multi-core chips; benefit that users could only wish for with a P4HT.

Re:Am I Missing Something?

[MindStalker]

Also with dual core memory controller on chip has the distint advantage of being able to share that on chip memory. Meaning core to core communication is significantly faster than chip to chip communication.

Re:Am I Missing Something?

[Donny Smith]

>Money would be better spent on RAID, rather than dual core or dual processor.

You're right about that.

Unlike CPUs which become worthless in less than 2 years, RAID h/w last a bit longer.

Some five years ago I bought an Ultra2Wide SCSI 320 card and a (at the time big) 8GB HDD - I paid $400 for the card and $250 for the HDD.

I still use the card - I haven't checked but it should be as fast as SATA II I guess - and the SCSI disk works too (although it's quite useless - I use it as dedicated swap disk).

In the meantime I went thru 3-4 generations of motherboards and CPUs (consecutive 100% wipeouts) and my RAID stuff still rocks...

By year's end I'll go not for a dual core CPU system but for what's today top of the line nForce4 system. Screw the hype.

Re:Am I Missing Something?

[Thagg]

Multiprocessing, both discrete and multicore, will accelerate all compute-bound applications in the future. Right now we haven't reached a critical mass yet -- where programmers feel it's worth the effort to multi-thread all of their applications -- but we will get there soon. It's not an easy change, and there are a whole world of problems that programmers haven't had experience with yet, but either these programmers will learn or they will not be competetive any longer.

Face it, the days of increasing clock speeds is over. It's done. Finished. Kaput. The low hanging fruit has all been eaten.

On the other hand, the multiprocessing benefits are huge and practically untapped. There is every reason to expect that in ten years we have 64 or 256 processors on a chip. People who hope to be working in ten years better learn how to write for these systems.

Compare the P4 to the Cell. The P4 goes to unbelievable lengths (even literally, in pipeline lengths!) to run at a high clock speed. Its contribution to global warming is substantial. It's expensive. And, it's an absolute dead end. Intel has already abandoned it.

The Cell has eight much-simpler processors along with its Power core. It can, and will, compute 10 times as fast as a P4, if programmed correctly. The game programmers are going to be pulling their hair out for the next couple of years, but they are going to be the high-demand programmers of the next decade as they are the first over the wall of significant multiprocessing.

Thad Beier

Re:Am I Missing Something?

[flocculent]

Money would be better spent on RAID, rather than dual core or dual processor.

Actually RAID doesn't noticeably speed most things up in a desktop environment in my experience.

http://faq.storagereview.com/tiki-index.php?page=S ingleDriveVsRaid0 has this to say:
Dont assume RAID 0 offers increased performance for all or even most applications... and dont assume that transfer rates reflect application-level performance.

Perhaps the money is better spent on more RAM? :)

Re:Am I Missing Something?

[Just Some Guy]

But what's the processor utilization? On most systems, its usually less than 10 percent. So when a user does something, the bottleneck is usually not the processor. Its usually the hard drive.

That's most systems, but certainly not all. I wrote a web application in Zope that acts as a portal to our scanned document warehouse. Whenever a customer wants to access some of the data we're storing for them, we fetch a few TIFFs from a Samba filesystem, convert them into a PDF with ImageMagick, and send them out. A RAID wouldn't make a bit of difference to our setup, since even the comparatively slow network file retrieval is much faster than the image processing which is the real bottleneck.

Our system is idle probably 95% of the time. In fact, it currently has a 5-minute load average of 0.04. But in that other 5% of the time, we want it to respond NOW and not 30 seconds from now. This is a pretty common situation for server machines - relatively long periods of inactivity punctuated by short periods of frantic scurrying - and it seems reasonable that AMD is offering their 64-bit server chip with this feature.

Re:Am I Missing Something?

[ColdGrits]

Absolutely correct.

Indeed, you can do that across all Sun's UltraSPARC-III - based SunFire range, replace the single-core US3 with dual core US4 CPUs. Something Sun deliberately decided to ensure was possible right from the outset.

Re:Am I Missing Something?

[That's Unpossible!]

But what's the processor utilization? On most systems, its usually less than 10 percent.

Regardless of the fact that you pulled that number out of your ass ;-) there are many applications that are processor, not data, intensive. Also, in the case of servers that run multiple services, if one of your services has a problem and pegs out a CPU, your site is not completely crippled, and it is also much easier to remotely connect to the server to fix it. Ever try fixing a machine with one CPU that is completely pegged out due to one broken app?

Re:Am I Missing Something?

[magarity]

What makes you think the database is CPU-bound instead of I/O-bound?

CPU load versus disk queue length versus memory page accesses. But no worries about my off the cuff example being the only case; a very popular move these days is to run VMware (or similar) on an 8 or 16 way machine in order to provide 20, 30, or more "servers" doing light to modest loads so introducing dual core CPUs will allow even more virtual servers hosted on a single physical box. The cost saving potential is TREMENDOUS.

And what makes you think the manufacturer of the two machines will allow this software patch

It only takes one manufacturer to advertise "Buy our 8 way single core now and upgrade to dual core on the cheap later!". You can buy an 8-way now with only 2 CPUs and add later; there's no reason why you won't be able to buy 8-way dual core capable with only a few single cores and upgrade/add CPUs for years. Big corporate servers costing serious money are upgraded and/or assigned to different roles for a long time compared to desktop PCs. It's a whole different world. One company *finally* decommissioned a quad PPro server and donated to a nonprofit I know. That's been in service 24/7 for what, 14 years? It was probably over $50K when new and they have to get their money's worth out of it. Anyway, the manufacturers know that most IT budgets are not unlimited and customers always like less expensive alternatives.

Re:Am I Missing Something?

[sgant]

Waiting for someone to come along now and quote some article about how RAM doesn't noticeably speed things up in a desktop environment.

Pretty soon we'll talk ourselves into a corner where nothing speeds anything up and we were better off with a C-64.

(btw, I'm joking)

Re:Am I Missing Something?

[yorkpaddy]

SMP ready cpu's weren't always more expensive than there UP breatheren. The CPU's are almost exactly the same, but the manufacturers just disable the SMP ability and sell them for more.

Re:Am I Missing Something?

[default luser]

Where's the surprise in this? They havn't pulled any amazing numbers out of their hat, 55w is roughly %70 more power than a single core Winchester. The other %30 is probably accounted for by a new more efficient core revision. The .09 micron Athlon 64 Winchester parts already sip power AT FULL LOAD (27w for 3000, 30w for 3200+, ~33w for 3500+).

The design spec for the Winchester core says 63w TDP, but that's just so designers can eventually drop in a 2.4-2.6GHz+ Winchester core. These higher frequencies will most likely require higher voltage, and thus have substantially higher power consumption than the 3000-3500 Winchesters.

Or perhaps you could just, say, drop in a 2.2GHz dual core. So many wonderful options...

Re:Am I Missing Something?

[twiddlingbits]

RAID's whole purpose was NOT speed it was REDUNDANCY. Most RAID controllers are just an upgrade over the standard ones in most PCs, that's where the performance upgrade comes in. They move data faster but if your program can't process the data as fast as it comes in you really don't see a benefit. And even with RAID the disk size makes a difference, smaller is better in many cases.

Re:Am I Missing Something?

[pclminion]

Its that Firefox can be run on one processor, while your MP3 player is running at the same time on the other. This in turn will speed up BOTH applications, since Firefox does not ever have to yield to the player, and visa versa.

Only if those applications were maxing the CPU to begin with. An MP3 player on a modern processor only utilizes around 1% of its capacity. Firefox a similar amount. They can easily share a CPU with 98% of its capacity to spare. They might run imperceptibly faster due to better cache utilization, but the reality is that almost every application spends 99% of its time waiting for something slower, like disk or network.

The only sort of application that a typical user (i.e. a non-developer) uses that's actually capable of maxing the CPU is, say, video editing, or a high-performance game.

Re:Am I Missing Something?

[pclminion]

If you have any multithreaded app that is even remotely competently written, then it will benefit from dual cores or (possibly) hyperthreading. If your multithreaded app is full of "big locks", then dual cores won't help, and the application designer is a failure.

You assume everything can be trivially multithreaded. That's utterly wrong.

First of all, multithreading is not always done for performance. Sometimes it makes the code easier to write. It's nice to be able to compartmentalize tasks into threads and control them with a boss.

Second, in the field of parallel programming there is a term called "embarrassingly parallel." This means that it's obvious how to multithread the task, in such a way that you get maximum benefit from each CPU. However, real-world tasks are rarely "embarassingly parallel." It's absolutely unfair to call the application designer a "failure" because the application can't be effectively multithreaded. Most things can't be.

There are many real-world super parallel applications (the kinds of things that run on 10,000 CPU clusters) that have properties like, to double the execution speed you must have 8 times the number of CPUs. This doesn't mean the application designer is an idiot, it means the problem is hard.

It's easy for people who've never done parallel programming to run their mouth off, however.

Re:Am I Missing Something?

[router]

This reminds me of the SCSI-IDE debate; if you have used multi processor boxes as your primary workstation you will always want one from that point on. Same with SCSI back in the day, doesn't make sense until you use it and get used to how much better it is, going back is like getting your hair pulled. Netscape could crash and I wouldn't know it, I could encode mp3s without worrying if playing another one at the same time was going to get hokey, CD burning was not failure prone, etc. Just makes the experience better if you are actively doing more than one thing at a time. If you don't agree, don't buy one, I don't care.... I love duals tho, if I used a desktop instead of a laptop now it would be a dual anything (PIII 500ish+) with scsi.

andy

Re:Am I Missing Something?

[MindStalker]

Also from a single app standpoint you may not see a significant increase in speed. But when running several apps like most desktops do nowadays, you should see a good increase.

Re:Am I Missing Something?

[ckaminski]

Um, the PPro came into production use in ~1995, so no more than 10 years.

Re:Am I Missing Something?

[RzUpAnmsCwrds]

"2.6GHz Athlon FX57 with 1MB cache. The latter has a 95W power draw already, but it is 130nm"

There is no Athlon 64 FX-57. The fastest available Athlon 64 is the Athlon 64 FX-55, at 2.6GHz / 1M / 130nm.

All 130nm Athlon 64 CPUs, excluding notebook parts, have a Vcore of 1.5v and thermal design power of 89W. Measurements of the Athlon 64 FX-55 actually place it at around 70-80W.

All current Opteron processors (excluding the HE and EE variants) have Vcore of 1.5V and TDP of 89W as well.

Winchester-based (90nm) Athlon 64 cpus have Vcore of 1.4V and TDP of 63W. As with earlier K8-based CPUs, the TDP is for the entire line, regardless of clock. Low-end parts consume substantially less power than indicated by the TDP.

Intel, on the other hand, understates their TDP. Their TDP is meant to reflect "typical" usage patterns. Actual P4 Prescott CPUs have been measured to draw as much as 130W - 150W when under heavy load. Intel is relying on the thermal protection built into the CPU to slow down the system if necessary and prevent overheating.

Re:Am I Missing Something?

[pclminion]

You're saying all this to a person who has used exclusively SMP systems since 2000. I have five years of experience telling me it doesn't make any fucking difference for everyday use. It's not "theoretical" in the slightest bit.

Maybe it's your operating system that sucks.

How much power is "reasonable"?

[garcia]

They talk a lot about this being the savior of power-consumption but:

They are seen as the solution to power-consumption problems that have come to the fore as clock-speeds have increased beyond 3.0 GHz. At such speeds, single-CPU processors can often dissipate more than 150 W. In contrast, dual-core parts can reduce power consumption to more reasonable levels. For example, a processor with dual 2.0-GHz cores can deliver performance not all that different from a single-core 3.5-GHz part. More important, such a dual-core part will hold down power dissipation to a figure closer to that of a standalone 2.0-GHz CPU, allowing processing throughput to effectively double for not much more power.

Yeah, great, so it reduces power-consumption to "more reasonable levels" yet in every article I have read on this no one really mentions much more than that. What's reasonable? Telling me twice the speed for not much more power doesn't mean anything to me (other than marketing doublespeak).

What I want to know is how much money these processors will save in power consumption compared to how much more they will cost over their single core cousins... No one has said anything about that yet.

Now, also, how many OSs (and applications) are prepared for dual-core support? Are there any available systems that are stable and do that?

Re:How much power is "reasonable"?

[oldmanmtn]

Now, also, how many OSs (and applications) are prepared for dual-core support? Are there any available systems that are stable and do that?

Solaris supports dual cores on both SPARC and x86. The UltraSPARC IV processors are dual core.

Any application should be "prepared for dual-core support". If the application even has to be aware that it's running on a dual core or hyperthreaded CPU, then the OS is broken.

Re:How much power is "reasonable"?

[Jugalator]

Now, also, how many OSs (and applications) are prepared for dual-core support?

I won't go into detail of applications since I have no idea which apps you're interested in, but Windows XP Pro supports dual cores (it runs its multi-core kernel even if you just have a Pentium 4 with hyperthreading).

Windows XP Home will not suffice though, which is a bit amusing since this might be the most common OS sitting in homes of gamers which are often the early adopters of this kind of tech nowadays. Unless they just pirated Windows XP Pro with a volume license key of course. :-)

Windows XP Media Center Edition, and Windows XP Tablet PC will support multiple cores though, probably in the same fashion as Pro.

Another one may give details about the common Linux distros, but I'd be very surprised if this support isn't in by far most modern distros, or can be enabled fairly easily.

Re:How much power is "reasonable"?

[bbrack]

The power increase between a single core processor and dual core processor is probably less than you think: I would not be surprised if it is in the 20% range on average.

Simply running the clock (and not performing any operations) on most processors will draw ~60-70% of the parts max power, which suggests that the loading on the part determines how much of that extra 30-40% is being dissipated.

Working under this assumption, worst case, a dual core processor would draw 40% more power than a single core processor, while effectively doubling your maximum throughput.

As for cost, I would be surprised if these processors debuted at more than 50% more than their single core cousins, and the prices would probably drop fairly rapidly

Re:How much power is "reasonable"?

[SirCyn]

Now, also, how many OSs (and applications) are prepared for dual-core support? Are there any available systems that are stable and do that?

Microsoft Windows 2000 and XP support 2.
Apple OSX supports 2.
FreeBSD support 4 (or more?). NetBSD supports 2 (or more?). OpenBSD is working on it (last I knew).
Linux 2.4.x and 2.6.x support 2+.
Sun Solaris has support 2+ for as long as I know.
AIX, HPUX, SCO Unix and all those support 2+.

Did I miss any?
Almost all OSes for the last several years have supported multiple processors natively. At worst these OSes would need a patch to update their SMP awareness.

Applications on the other hand, well they've been slower to change to a multithreaded moddel. Many server grade programs are ready. Most common desktop programs are not.

I have used a dual Athlon MP system for a long time now. The biggest difference I can tell you between dual 1.6GHz and single 3.2Ghz is that one process can not take over the processor. Even with modern preemption I can tell the difference when I have a second CPU processing my clicks and keystrokes. All I can say is "try one for a while, you'll get hooked".

Re:How much power is "reasonable"?

[TheRaven64]

Yeah, great, so it reduces power-consumption to "more reasonable levels" yet in every article I have read on this no one really mentions much more than that.

Power consumption is roughly proportional to the square of the clock speed (everything else being equal). Doubling the clock speed of a CPU will increase the power usage by a factor of 4. On the other hand, simply adding an extra core will only double the power requirements of the chip.

Re:How much power is "reasonable"?

[Malor]

If XP works the same way 2000 did, when you upgrade to a dual-core from a single-core, you will have to reinstall the OS for the second core to be activated.

2000 had two entirely separate sets of system files, one each for uni- and multi-processor. Even if you added a second CPU, if you didn't have the multiproc HAL to begin with, it simply wouldn't work.

Because XP is just 2000 with a facelift, I suspect this won't have changed. You are correct that if your initial install was on a P4, which 'looks like' two physical processors, XP would have installed its multi-cpu core.

If, however, you are installing a dual-core Athlon in, chances are quite high that you didn't do your initial install on a P4. So you won't have the multiproc system files, and you'll probably have to reinstall to get the second proc going. (A 'repair' installation may be adequate, and would be much less painful.)

Linux works somewhat similarly, but fortunately you can replace just the kernel, rather than the entire OS.

Re:How much power is "reasonable"?

[Chris Colohan]

Good question. If you look at the recent history of CPU design (say, for the past 20 years), you see that the primary concern of architects have changed with the technology over time. In the 80's any design you came up with was limited by how many transistors you could squeeze onto a chip, and so everyone was worried about transistor count. By the 90's the transistors became abundant and small enough that a lack of transistors was no longer a primary concern of designers --- instead, they were much more concerned with the wires: how to pack them in and route the nicely, and how long it took for signals to get from one place to another.

These days, wires are abundant, transistors are abundant, and the main concern is power. Why?

a) GHz == Power. If you increase the clock rate, you increase the power drawn. So every time you want to make the machine "faster", you need to feed it more power.

b) More logic == more power. If you want to make the chip more complex or have more cache, it will need more power. But at the moment, it is cheaper (in terms of power) to add more logic than to add more GHz.

So why is power a problem? How much is too much? First, let's talk about total power: the total amount of electricity you have to feed into the chip to keep it running. This changes depending on how you utilize the chip, but designers have to design for the worst case.

There are two problems with total power: how do you get the electricity into the chip, and how do you get the heat it produces off of the chip. My understanding is that both problems are hard, but the heat one is currently harder. If you put too much power in, you need to get the heat out. More power == more heat. More heat == more expensive cooling solution. For example, if you had a low power chip you could get away with a simple and cheap heat sink. As it draws more power you have to add expensive (and noisy!) fans. Add more power, and you might eventually have to add a liquid cooling system, air conditioning unit, and a bunch of plumbing. Add more power, and perhaps you have to hire a team of workers to constantly feed dry ice into your machine, or some other creative cooling solution. So the fundamental problem is that once your chip draws enough power keeping it cool becomes too expensive. Right now it looks like 100 to 150W is the max power which can be kept cool by fan systems which are cheap and quiet enough that people will actually buy them. Nobody wants to buy a computer with plumbing for their home or office, liquid cooled machines are still only used for special business applications where they have little other choice.

Now total power is not the only problem. There is also localized power. Think "will a portion of my chip melt into a pool of molten silicon since I can't cool it off fast enough"? There is a nice story of how one of the first Alpha CPU prototypes cracked in half due to having too much power running up the middle of the chip... It turns out that this localized heat is a major limitation these days, and is a primary reason why Intel won't sell you a 5GHz chip. They don't know how to avoid this problem in a reasonable way. So, instead, they are going to improve performance by adding cores.

If you have one core, you have one central location in the chip that is constantly working hard, and so it is generating a lot of heat. Getting the heat away from that tiny point is hard. If you build a chip with two cores, then you divide the work over two points, and so each point generates half of the heat. Keeping that cool is much easier. With four cores....etc.

A research paper at last year's ASPLOS suggested that a good way to work this problem is to have a program run on one core until it gets too hot, then move execution to the next core, and repeat. This way you always have one core doing work, and the others are cooling off. We may end up having to do this.

Re:How much power is "reasonable"?

[un4given]

2000 had two entirely separate sets of system files, one each for uni- and multi-processor. Even if you added a second CPU, if you didn't have the multiproc HAL to begin with, it simply wouldn't work.

Fortunately, this is incorrect. As described in Microsoft's knowledge base, a HAL change is all that is required to take advantage of the second processor. Windows NT 4.0, 2000, XP, and 2003 all are capable of this, although NT needed a utility uptomp.exe to accomplish this feat.

It is very common to have to do this on dual-processor capable servers when installing a second processor.

Re:If this is anything like

[Junta]

HT is a way of letting one processing unit work with mulitple threads at once. Multi-core technology is identical to SMP, meaning more physical processors actually doing work, so it isn't token.

However, expect lower clockspeeds, two cores in that proximity causes a severe power/heat problem that would mandate reduced clock over single processor solution.

Why not two different clock speeds?

[Gates82]

Rather than using two identical chips, is there any difficulty in putting say a 2.4 Ghz chip with a 500 Mhz one? I would love to have the latest and greatest chip for gaming, and crunching through video and then have a low powered second chip to play my mp3's and surf the web, while the high-end chip is crunching through numbers in the background. Guess that's why I have a laptop, play on that while the desktop is doing its thing.

--
So who is hotter? Ali or Ali's sister?

Re:Why not two different clock speeds?

[bbrack]

The main problem with this is that the processors share a clock tree and arbitration logic - if the clock multiplier is contained in the arbitration logic, then having one core at one speed and another at a different speed would be impossible.

If the clock multiplier is contained separately in each core, it would be possible - however, having different clock ratios on each core would considerably complicate the arbitration logic, since it would have to deal with different setup and hold timings when sending data to one core vs. the other - this would probably greatly increase your chances of inducing a processor error.

Trying to do this could also require a great deal more design difference between the two cores, which might cause many problems. It also would make it much more difficult to sell single core versions of dual core chips (i.e. one core fails, the other core is good - blow a few fuses to get the chip to look like a single core chip, and sell at as a single core)

Re:Why not two different clock speeds?

[bbrack]

While there may be separate clock domains per core, the clock multiplier circuitry may be contained in the arbitration logic, as this could probably improve the arbiter's efficiency.

I think that the design specifics of the arbitration logic (specifically, what speed it communicates with the cores, and how core-to-core communication is handled) is the determining factor in how feasible multi-frequency dual core circuits are.

I'm poor!

[shamowfski]

Does dual core mean dual price? With current fx-55's costing around a grand, what can we expect these to cost? 1,500-2,000? If AMD wants to remain competitive with Intel they are going to have to work on that. Who ever guessed AMD would be the one who had to lower prices to compete??

Re:I'm poor!

[ThaReetLad]

According to AMD, no. IIRC they said recently that dual core Opterons will be inline, price wise, with the top end Opterons currently available, while offering much improved performance.

Re:I'm poor!

[Hoser McMoose]

More specifically, the chips cost $1,299 when they were first released.

Many people seem to forget that this was not considered abnormally expensive for a processor not all that long ago.

Check your licensing agreements first

[bigtallmofo]

Before you buy one of these dual-core processors for your server, make sure that your software vendor isn't going to double your price on you.

Oracle and others have announced plans to increase their revenue by charging people for multiple cores in their single processor.

Re:Check your licensing agreements first

[Anonymous Coward]

What's gonna happen when you try to install the Intel Dual Core "Extreme Edition" on a Windows XP box?

XP sees the single core + HT chips as 2 CPUs, but if the extreme edition dual cores are HT enabled, will XP allow "4" CPU's??

Re:Check your licensing agreements first

[Jugalator]

That's an interesting question. :-)

First, I think Microsoft has previously said they'll only charge one license per physical processor. However, the problem is that Windows XP Pro supports only 1 or 2 logical CPU's. I wonder how it'll react. It seems like you'd not make optimal use of your system. And in case it does restrict itself upon working on two logical CPU's, you should hope Windows XP restrict itself upon the two cores instead of one of the cores and the HT feature on it!

Re:Check your licensing agreements first

[ThaReetLad]

Buying Oracle is like buying a Ferrari. If you have to ask how much it costs, you can't afford it.

Awesome! 939 Huzzah!

[Anonymous Coward]

I actually just purchased a socket 939 board for this exact reason. I'm extremely pleased with AMD for not forcing yet another motherboard upgrade on us based on chip advancement. I got a cheap Athlon 64 3000+, but two or three years from now I can go dual-core without getting a new motherboard, memory, etc. and I like that.

I understand that sometimes it's necessary to upgrade motherboards instead of just chips (FSB adn so forth), but for those of us who can't afford top-of-the-line, bleeding-edge stuff, it's nice to see upgradability for more than just a few months into the future.

Free Sony PSP from Gratis

Re:Xmas list material

[Anonymous Coward]

MORE POWER ARH ARH ARH!

Poor Tim Taylor... His claim to fame was nothing more than a strange barking while being laughed at by a fat man in a flannel.

If only he realized how hot his wife was. I'd never be in my god damn garage with her running around!

Similarly, no one should be putzing around with more power in their dual-cores when there are women to be had!

Oh wait, sorry, I forgot this is Slashdot (the REAL "tool time") ;)

Re:If this is anything like

[Kiriwas]

You should probably get your terms right before you comment on it. HT is simply Intel's name for SMT (simultaneous multithreading). They didn't choose an optimal implementation and people shouldn't expect the same performance from it as you would from dual processors. SMT is simply an extension of the superscalar idea. Disconnect the dispatch mechanisms from the execution mechanisms and you can run an out of order processor a lot faster than an in order. Make multiple execution units and multiple fetches per cycle and you now have an n-way superscalar. A few more additions (mostly replication of units in the processor) and you can grab instructions from multiple threads instead of from the same thread (it is difficult to get lots of instructions per cycle from the same thread because of the high frequencies of branches in the code stream - and branch prediction isn't perfect). Dual core is completely different, they simply put two processors on the same chip. Dual core has the problem that it cannot share the same resources between the two threads. The resources (execution units, queues, etc.) are partitioned x amount for thread 1 and x amount for thread 2. The designs are really very different, depending on the use, sometimes dual core is better, sometimes SMT is better. AMD's planning on bring out Dual Core SMT where each core will have 2 threads running through it for a total of 4 thread running simultaneous. If you want more information about this "throughput computing" google for Sun's Niagara chip.

Most games are multi-threaded

[wowbagger]

Most games have several potential threads running at once:

• Graphics rendering
• Sound rendering (compositing the various sounds together, and playing music)
• Game logic (monster AI, object movement, physics model)
• User input monitoring
• Network processing

An SMP system can greatly benefit a game designed to be truly multithreaded.

Even if the game is NOT designed to be multithreaded, there is the fact that one core can be running the game, while the other core handles interrupts, operating system processing, and other tasks.

The days of your computer doing only one thing at a time are long gone.

Re:Oracle and Dual Core CPUs...

[Shalda]

The answer of course, is don't use Oracle if you can help it. :) They're just as evil as Microsoft, but they don't give away nearly as many free T-Shirts.

Servers envorement, duh.

[KZigurs]

It's simple. Everyone knows that you can scale better in heavly multithreaded envorement by adding more CPU's instead of increasing speed of single CPU - it consumes less time on each CPU spent for managing tasks and context switches than on single, since each cpu has to work with smaller amount of threads.

And since currently even a desktop computer starts to approach point where there are hundreds of threads running (check in task manager or top) - this makes quite a lot of sense.

Also, a lot of people mistakenly believes that Hyperthreading in their intel CPU's brings similar benefits as SMP. No, it is just a nifty trick to keep the long p4 pipeline filled with as much data as it can. But multicore chips are in fact two CPU's in the same casing. (think - real hyperthreading :P).

Also there is the issue of simplified motherboard design (less traces for the same amount of cores), reduced packaging costs and higher computing density. All three being quite considerable points. As they say - computing today is all about integration. And multicore CPU's are one of the answers to allow simpler integration and allow greater flexibility (same costs to produce MB that supports single or multicore cpu. But the performance benefits - quite significant).

Hrmmmmm

[REDSECTOR1]

MAKEOPTS="-j3"

Horray

Complexity

[Efialtis]

They have been expirimenting with multi-layered parallel processing for a long time, and I think this is the "realized results" of those expiriments.
We will see newer dual and multi-core processors come out in the future, and tha ability to parallel process with multiple chips on one board...
Should be exciting...

Sweet...

[sapgau]

Intel wake up!! See how easy it is to upgrade, no new socket layouts, no new motherboards.

Besides trying to determine what model is the Pentium dual core gives me headaches.

/owns AMD, trying very hard to repress fanboy attitudes.

So buy a cheaper processor

[mapmaker]

With current fx-55's costing around a grand, what can we expect these to cost? If AMD wants to remain competitive with Intel they are going to have to work on that.

If you need a car but you're poor, you buy the Chevy Cavalier, not the Chevy Corvette.

If you need a processor but you're poor, you buy an AMD Sempron, not the AMD FX-55.

Complaining that AMD needs to lower the price on their top processors is like complaining that Chevrolet needs to lower the price of Corvettes.

Re:If this is anything like

[masterofsw]

Actually, HT is not "simultaneous multithreading". Intel added extra hardware registers, etc, to make switching threads faster. HT still is a single threaded processor.

Lockless techniques to the rescue

[bsmoor01]

I think this will be overcome with more atomic instructions from cpu vendors. Lockless techniques generally give much higher performance, and can often acheive the same goals as the old 'lock to enforce synchronization' paradigm you mention.

I use lockless counters heavily in the code I work on in order to reference count objects. Very handy, and much faster than lock-based counters.

The pain with lockless coding is that there aren't many portable primitives. So I have to maintain my own abstractions for every platform I work on, which is a pain since I'm in embedded systems. It would be awesome to have a standard (AND portable) lockless utility library. One day perhaps...

Re:If this is anything like

[Kiriwas]

Actually I beg to differ. http://www.intel.com/technology/hyperthread/ says explicitly that HT is a form of SMT (simultaneous multithreading). The processor contains multiple PC registers which allow it to actually follow multiple threads simultaneusly -- which means grabbing instructions from multiple threads simultaneously. Again, this is really just an extension of superscalar, which could only grab instructions from a single thread.

Closer to reality?

[cmclean]

Dual-core chips are already a reality, Sun's UltraSPARC IV uses 2 UltraSPARC-III pipelines.

Perhaps the author means "x86 dual-core chips"?

Re:Dual Processor, Dual Core?

[speculatrix]

there's nothing in the article to suggest that a dual-core amd64 can be used in a dual-cpu motherboard?

I would think that AMD will have to make multi-cpu versions of the dual-core chips, to support the inter-cpu comms, just as there are 1xx, 2xx, 4xx and 8xx versions of the regular opterons. But it'd be cool to cheaply upgrade a dual mobo to quad!

We have some dual-cpu opterons tyan motherboard machines here, and they are awesomely quick - we use FreeBSD, amd64 where possible but too often i386 mode because java and other things aren't supported in 64 bit mode at all yet.

rc5-72 rating is about 9M keys per second, twice that of a P4 2.8GHz machine; however, they are slaughtered by a dual Apple G5 system which achieves nearly FOUR TIMES the performance.

Hope they make cheap ones

[springbox]

I hope that the processor manufactures make some economy version of the dual core CPUs like Intel has done with the Celerons, since anything new coming out from that industry seems to be generally pretty expensive.

Mmm dual core Celeron..

not sure about RAID-0 but RAID-1

[MarcoAtWork]

is very noticeably faster on my system, for a bit I was running a single HD (due to other issues) and as soon as I added the 2nd and mirrored it my XP boot time has gone down by around 40% as well as Half Life 2 load times.

I know I am paying a penalty in write speed but the doubled read speed more than makes up for it. With HDs as cheap as they are now (I have 2x200G Seagate SATA) and RAID controllers integrated in most mobos (I have a Silicon Image in my a8n-sli board, which I prefer to the nvidia chipset raid) I think it's stupid *not* to go the RAID route, as with a very modest cash outlay (for a 2nd HD, or for a 3rd and 4th if you plan to run RAID0+1) you'll see a noticeable speedup, not to mention that if one of your HDs packs it in you won't be SOL.