Intel Plans for Dual-Core Prescott CPUs in 2005

scapermoya writes "X-Bit Labs is reporting that Intel is planning to step up their introduction of dual-core processors, with the first chips to hit the market in late 2005. Intel announced this plan at the Technology for Business Today seminar, held in Washington, D.C. Looks like NetBurst is sticking around, despite what we have heard lately about a move toward the 'M' architecture. Supposedly, thanks to HyperThreading, the OS will see 4 installed processors. Snazzy."

Comments

Windows Licencing

[Neophytus]

The sticker on the bottom of this here laptop says Microsoft Windows XP Professional 1-2 CPUs. Will this mean that Microsoft will have to reconsider their licencing policy for CPUs if people are going to have "four" from one chip? I've never needed to run more than two (through hyperthreading) so if someone could shed some light on what happens if you give a "2-licence" four processors it would be appreciated.

Re:Windows Licencing

[afidel]

XP Pro understands hyperthreading fine, and in fact will work with a machine with dual hyperthreading CPU's. What won't work is 2k Pro. Also note that 2k3 server standard reduced the max number of physical CPU's to two for standard edition whereas it had been 4 in previous iterations of NT. Btw don't use hyperthreading on win2k even if you have enough processor licenses because it will balance evenly across the hyperthreading CPU's not realizing that they are really there for spare capacity and hence trash the instruction cache.

Re:Windows Licencing

[slash-tard]

Currently they license based on physical cpus, not the logical number. They had to make some minor changes when Intel introduced hyperthreading.

My guess is initially it will count as 2 cpus and as it becomes more widespread they will revert to it being counted as 1. A true dual system (4 cores, 8 logical) would be counted as 2.

Of course in another few years they will be feeling more pressure from linux so they may change the licensing a lot.

Re:Windows Licencing

[SamiousHaze]

Windows XP handels Hyperthreading, so you can have 2 physical processors and 4 logical processor according to microsofts hyperthreading document.

Re:Windows Licencing

[Short Circuit]

The same arguments were abound when Intel introduced HT. And Microsoft didn't touch the licensing.

My guess is XP Home will continue to only use one logical processor, while XP Pro will use two. (Now, whether the "second" logical processor is HT on the first core, or primary use on the second core, remains to be seen.)

Re:Windows Licencing

[qodfathr]

Licensing is for physical processors only. So, even today, if you have 2 physical CPUs with hyperthreading, you are compliant. Task Manager will show 4 CPUs, but the OS can determine that only 2 are physical.

I suspect, however, that a dual-core CPU will be treated as 2 physical cpus...(+2 virtual CPUs)

Re:Windows Licencing

[AliasTheRoot]

Weren't MS recently touting recommended specs of Longhorn to include dual core chips? I somehow can't imagine them insisting on people buying a dual CPU license to run it - i'd imagine they will stick to the amount of physical cpu's on a motherboard.

Hell, if it weren't so complicated to deal with, they'd probably go for something based on the overall performance of the CPU(s) in the system, as Oracle did (do they still do this? haven't dealt with the licensing in a while)

Re:Windows Licencing

[AliasTheRoot]

XP 2003, aka server supports more - if multi cpu setups were at all common in the desktop workstation market then this would be rolled up into XP Pro, but it isn't.

Pretty much the same thing has applied throughout the NT Lifecycle, with Workstation sticking to single / dual setups and Server handling 1/2/4/8 setups.

Re:Windows Licencing

[RMH101]

2k server does, mind. i've got six DL380's in front of me with dual HT xeons, and task manager shows them as 4 CPUs.

Re:Windows Licencing

[Asprin]

My six dual-Zeon IBM xServer 225's correctly see four processors as well. One of them also has MSSQL2K and it also understands the two of the processors are hyperthreaded non-physical CPUs and does not complain that we only have two processor licenses installed.

It was a surprise to me when I installed Windows on them a couple of months back because I didn't even think Win2K supported hyperthreading. w00t!

Perhaps any hyperthreading-related issues that may have existed with Win2K were patched in a service pack?

Re:Windows Licencing

[Kris_J]

Why do they care? What possible legitimate reason would Microsoft have for reducing the number of CPUs that XP can run on, if it's based on code perfectly capable of running on more? This is the same crap you have to fight if you want concurrent sessions. MS's licensing is by far the worse thing about their products.

Re:Windows Licencing

[Morgahastu]

If only my typo check + preview before submitting was redundant.

Re:Windows Licencing

[AliasTheRoot]

I think they will stick to physical CPU's tbh, the low hanging fruit for chip makers in terms of performance seem to be multiple cores rather than increasing clockspeed, i just dont see MS saying to Joe Consumer "well your single CPU has 8 cores so you need to buy 8 copies of Windows."

Licensing may be different for Server installs, but for consumer/desktop Windows I doubt it.

Re:Windows Licencing

[afidel]

Yes, but if you had a 4 ways system with HT enabled you would really only be using the first two CPU's and their HT units, the other two physical processors would be doing nothing. Not to mention that the OS and older apps will naively balance across all 4 'CPU's' evenly despite the fact that the HT units are not fully capable processors and in fact can easily degrade performance by trashing the contents of the instruction and more importantly data cache. MS claims that windows 2000 will properly use the physical processors before the logical ones if the system is written to Intel BIOS specs but my real world experience says that there aren't a lot of correctly configured systems because I've had to restage 4 way servers from Dell, IBM, and HP and 2 way workstations from HP and Dell.

Re:Windows Licencing

[afidel]

That's cool but it's not true for all four way systems. Like I said there are a good number of systems that don't report their CPU's correctly and hence you end up with the situation I described. Trust me I've worked as a tech for some of the world's largest companies including IBM Global Services so I see a LOT more systems then your typical IT person.

Re:Windows Licencing

[sufehmi]

I don't see a need for hyper threading one dual core chips. It's redundant.

Well, a few weeks ago I had to install 3 test servers - but turned out we have only 1 available server.

I almost freaked out - but then I realised that it's a dual-Xeon box ---> 4 CPUs (to the OS). Using an obscure switch in VMware config file, I manage to assign each CPU to a (virtual) Server. The fourth CPU is assigned to the host OS.

Everyone was very happy with the result, and looking forward to utilise more of this feature (and this kind of CPU) in the future.

Re:Windows Licencing

[RMH101]

that's how it works. 2k server doesn't support over 4 CPUs, advanced server does - it's one of the reasons it's called advanced...

Re:Windows Licencing

[pedrop357]

XP Home also sees it as two processors. I used a Sony computer at Sam's Club that had a P4 3.06 HT. I was really surprised when I saw it was XP Home. Task Manager did show two processors and allowed me "set affinity" and everything.

Re:Windows Licencing

[Anonymous Coward]

Win2k will work with hyperthreading, it just doesn't work well. You'll probably see better performance if you turn it off, since Win2k doesn't realize there's only one physical CPU and thus doesn't balance the loads properly.

Re:Windows Licencing

[pedrop357]

I should have read the post a little better.
I was referring only to HT capable single CPU setups, not dual CPUs or a single CPUs with dual cores.
Of course, Intel/MS could do something to make XP Home treat one dual core CPU the same way it treats a HT CPU now.

Re:Windows Licencing

[TheRealMindChild]

Rediculous. The POINT of hyperthreading is to LOOK like 2 CPUs... be USED like 2 physical CPUs. The operating system SHOULD NOT even CARE whether its hyperthreaded single cpu or, in fact, 2 CPUs. Then you would be defeating the point (except for the per-cpu licensing, which I feel is silly anyway).

Re:Windows Licencing

[Sloppy]

Who needs 4 processors in a non server environment anyway?

I heard that nearly same question a few years back, but it was, "Who needs a 386 in a non server environment anyway?"

Re:Windows Licencing

[afidel]

No, the point of hyperthreading is to expose functional units that are not being used by the main process('s) running on the physical CPU. If you naively schedule everything across both 'CPU's' you will end up with stupid things like running parallel versions of a tight integer loop which is already maxing out the integer calculation units thus polluting the data cache and stalling the pipe MUCH more often which is a BAD thing on the P4. For more info on scheduler tweaks to accomodate HT I suggest you see this LK post by one of the Linux scheduling gods =)

Re:Windows Licencing

[pla]

What won't work is 2k Pro.

Sure it will! It just does a check at startup. It compares that check to a simple registry key. So, while you can't install with more than two CPUs active, you can install on two (to make sure it uses the multiCPU kernel), tweak the registry (try Google, I don't remember the key off the top of my head), then enable the rest of the CPUs. Not even a real hack, more like turning on LargeSystemCache. "Max CPUs? Why, I think I'll take 32, please!".

Gack. The thought of needing to "upgrade" to XP to run more than two CPUs makes me want to stay with only two, as nice as four (or eight, for two dual-core HTs) virtual CPUs sounds... No way will I put up with XP for that. Such a requirement might well give me the incentive to deal with Linux as my primary desktop machine, moving away from Windows entirely. ;-)

Re:Windows Licencing

[NanoGator]

"Rediculous. The POINT of hyperthreading is to LOOK like 2 CPUs... be USED like 2 physical CPUs. The operating system SHOULD NOT even CARE whether its hyperthreaded single cpu or, in fact, 2 CPUs."

Wrong. HT is not as fast as having a dual, not by a long shot. It needs to have some smarts in there or you end up with a machine that is often slower with HT.

Re:Windows Licencing

[Shippy]

That refers to 1-2 physical processors. You're allowed as many logical processors as you want. That only applies to WinXP and later, though. Back in the Win2k days, they counted logical processors toward your limit. Sure, there's ways of getting around the Win2k limit (as others have said), but out of the box, that's how processor count behaves.

Re:Windows Licencing

[IPFreely]

I always thought there should be something in between HT and multicore.

How about a single core that had a lot more units, more integer units, more floating units, more execution units, more register sets, more of everything. Then you could configure it to support one to eight threads similar to hyper threading. If you want raw speed, configure it for one or two threads. Each thread will have plenty of resources. If you want throughput, configure it for eight way, and keep as many pipelines full as you can.

The real bitch will be the scheduler. That's probably the main reason something like that hasn't been done.

Re:Windows Licencing

[Quikah]

Actually you should test your application in 2k with HT on to see if there is a problem. I have run a boatload of apps in 2k with HT with improved performance.

Re:Windows Licencing

[IPFreely]

I wasn't too far off the mark. But it looks like Andy Glew thought of it first.

It's not even real?!

[Anonymous Coward]

UPDATE: A representative for Intel Corporation told X-bit labs the company had never released any precise details in regards the dual-core strategy. The information published herein should not be considered as based on official statements.

WTF?

Damn

[Anonymous Coward]

Two core dumps for each segmentation fault.

Re:Damn

[DigiShaman]

But with HT enabled, that would be a Quad dump..damage...errr....

Too much Quake guys. Sorry. My mind is blown.

Re:Damn

[ongeboren]

correct me if i'm wrong but a code dump is make from the memory.. you got 2 processors, not 2 memories..

Heat?

[Sir dies alot]

Isn't it a llittle soon to claim to put 4 processors in one chip by 2005, especially sinse last I had heard, one processor was causing a heat concern. Have they fixed this or is this Intel making predictions and setting dates that will only get pushed back anyway?

Re:Heat?

[mobiux]

Intel...

Fighting the next Ice Age since 1968.

Re:Heat?

[funkytwig]

1 processor is causing heat consernes in thin core. The point is as the core gets thinner the power required to stop lekage across a ever thinner insulation layer increses. A couple of jumps thinner and we would have chips that require the power of a houshold iron. Multi-core is a solution to this problem, maybe Intel are not using very thin core technology to reduce heat in there multi core processors. There was a very interesting article about this in New Scientist but I dont think it was one they put on the web for free ;-(.

Re:Heat?

[Fweeky]

2; you only get 4 by a bit of extra hardware to virtualize one as two. Hyperthreading is just exploiting the already superscalar architecture a little more.

HTT has a transistor count overhead of ~5%; dual core is over 100% :)

Re:Heat?

[AliasTheRoot]

I'm oversimplifying for sure, but aren't the heat issues (and other more difficult quantum effects) primarily due to the ever increasing demand for clock speed?

As a layman it kind of makes sense to put 2 lower speed cores on a die rather than one faster one, and get lower power consumption and more importantly less heat production, and let the software deal with utilising it?

Anyone that actually knows about this care to comment?

Re:Heat?

[Kris_J]

Intel: You down with Entropy? (Yeah, you know me.)

Re:Heat?

[jadel]

These chips are going to be based on the PIII-m core which is already a low power design.
A quick google reveals a typical heat dissipation of 22W for a PIII 1.2GHz, doubling that for two cores would still be half of the 100W that a high end PIV would dissipate.

Re:Heat?

[LordKazan]

Yes

Modder processors use MOSFET transitors in CMOS arrays (Complimentary PMOS and NMOS networks in each gate)

This which means the only time power is expended (And therefore heat created) is when a gate transisions (if a gate stays the same across multiple clocks heat is only produced at the transition into that state).

So the more clock cycles you have the more often it's probably going to be switching states - each gate creates miniscule heat and power dissipation, but there are a lot of gates.

It is true that if you have two cores you have twice the heat - but that heat is also spread out across twice the area, less concentrated heate is easier to cool. Why do you think heatsinks spread the heat out all over the place.

A dual core processor running at 2Ghz would be roughly equivilient to a single core at 4Ghz - so the demand on clock rate increase is reduced and heat created/time is lower.

Re:Heat?

[marcosdumay]

Another important (maybe, because I don't understand the subject very well either) is if there aren't something better to do with the additional transistors used than an entire second core. Microprocessors today are so greatly optimized that there is no need for additional transistors in their core? And yeld is so low that is cheapper to achieve a certain performance level with a single huge silicon chip than with two smaller single core chips?

Re:Heat?

[Anonymous Coward]

I'm oversimplifying for sure, but aren't the heat issues (and other more difficult quantum effects) primarily due to the ever increasing demand for clock speed?

For the most part, yes. Power density is a problem. Increasing L2 cache sizes is an easy way to increase performance without adding much in terms of high power density... it's mostly static power consumption. But, this is changing too as leakage power increases with shrinking dimensions.

By going dual core, you save power with pads and you can share the L2. There's also the reduced cost of motherboards by not having extra sockets, pins, VRMs, caps, etc., and increased computing density, yadda yadda.

As a layman it kind of makes sense to put 2 lower speed cores on a die rather than one faster one, and get lower power consumption and more importantly less heat production, and let the software deal with utilising it?

Very good point. One thing to keep in mind, though, is Amdahl's Law... single-threaded performance still matters even for parallizable code, and parallel programming is a pain in the ass.

But, as Power 4 and 5 decided to put two complex cores onto a single die, there was an old project by Compaq/Alpha called Piranha that used 8 simple cores on a die (I believe they were Alpha 21164s, when 21264s were the hot thing). It was optimized for database/transaction code, since there's a ton of parallelism in db code and you can get a lot of performance improvements by reducing communication misses (and communication miss latency) and increasing L2/L3 caches (which reduces communication misses). Here's the link to the paper.

Longhorn won't see 4

[frs_rbl]

Thanks to patented HyperSlowing(TM) technology, Longhorn won't see 4 processors but 1.

''And not a very fast one'' a company exec was quoted saying

Re:Longhorn won't see 4

[Loconut1389]

Who needs those applications to slow your machine down for playing old 286/386/486 games? Just install XP!

Re:Longhorn won't see 4

[TheRaven64]

Actually, this would be more in keeping with Intel's strategy with the x86 line. x86 chips take superscalar architectures to extremes, and the idea behind a superscalar chip it to implement parallelism without forcing the compiler to know about it. A logical extension of this would be to make a number of physical processors appear as one logical processor.

The Itanium goes in the opposite direction, requiring the compiler to explicitly parallelise instructions. While this approach has the potential to allow much more scalability, it requires a much cleverer compiler. The combination of EPIC/VLIW chips with JIT compilation and run-time profiling could create some very high speed applications.

Ars

[ViceClown]

Ars is covering this too. Ken Fisher makes it a point to mention that the person who made the claims is in marketing. He also speculates, quite logically, that bringing out dual core Prescotts in '05 would be a feat even for Intel. Worth reading for a more sobering take on the situation.

Re:Ars

[paitre]

Exactly.
And -PRESCOTT- cores?
What, do they think we're nutty enough to have a desktop system that needs to dissipate 200+ watts of heat?
Please. I don't think so.

Re:Ars

[gr8_phk]

"bringing out dual core Prescotts in '05 would be a feat even for Intel."

Did he point out that it's easy for AMD? The K8 architecture has had dual cores built into the design from the start. Apparently they actually chopped one off for the first couple years. I've read that they have them running in simulation with both cores and I'd speculate they've even made sample chips with two. I've been wanting to know if AMD will go dual core at 90nm or wait for 65nm. I suppose it all depends on what Intel does. IMHO AMD needs to start acting like a leader instead of a follower - their 130nm parts are actually competetive with Intels 90nm ones in terms of performance.

Re:Ars

[Cleveland Steamer]

It'll probably happen when it is most cost effective, not when Intel decides to do it.

IMHO AMD needs to start acting like a leader instead of a follower...

AMD is acting like a leader instead of a follower. They announced their plans to support dual cores at day one. They introduced the AMD64 technology long before Intel followed them. They were the first to design an x86 processor with an on-die memory controller. They introduced Hypertransport, a technology for directly connecting processors together.

Re:Ars

[gr8_phk]

AMD is acting like a leader instead of a follower. They announced their plans to support dual cores at day one. They introduced the AMD64 technology long before Intel followed them. They were the first to design an x86 processor with an on-die memory controller. They introduced Hypertransport, a technology for directly connecting processors together. And they use a processor labeling scheme that rates performance in terms of Intel-equivalent clock speed. That holds Intel as the gold standard against which everything should be measured - not exactly the behavior of a leader is it? They have been quite innovative, and produce great products but they are only starting to see what it means to be out front. I expect them to have problems with speed ratings on the new 90nm parts, as there is no intel equivalent yet.

Re:Ars

[peterxyz]

"What, do they think we're nutty enough to have a desktop system that needs to dissipate 200+ watts of heat?"

It sounds like a great idea - combined CPU and espresso machine.

Start saving now, kids!

[dynoman7]

I'm sure they are going to be eSpensive

Re:Start saving now, kids!

[BlackHawk-666]

Lot's of their old ones, all the ones I needed, are free and available in PDF format.

Re:Start saving now, kids!

[10101001 10101001]

Don't you mean eXorbitantly eXtreme & eXpensive CPU?

Prescott?

[lachlan76]

The article doesn't actually say that Prescott will be a very promising architecture to use for a dual core configuration...imagine 200W of heat coming from a single dual-core processor.

Having multiple cores will make the already-present high heat requirements increase, while the processors in laptops get faster and faster, but not necessarily much hotter. The P6 architecture is the way to go, I think.

Re:Prescott?

[Jeff DeMaagd]

This may just be a hold-over while Intel readies the P6 version of a dual core architecture. This particular iteration might only show up in workstations. I think this may have been one of those projects that were far elong enough to justify keeping it so that there isn't a big gap in the CPU line-up when compared to AMD.

Re:Prescott?

[aunitt]

or
"Prescott" as in fat bastard who deserves a good slap?

Overweight and generates lots of hot air, I think that's much more appropriate.

Re:Prescott?

[shizzle]

I don't know for a fact, but my guess is that it's Prescott, OR... not too far from Hillsboro where the Prescott CPU was designed.

Dancing Techs

[Anonymous Coward]

What happened to the days of the intel dudes dancing around in bunny suits?

4 CPU's

[Deliveranc3]

I remember when Hyperthreading came out there were serious performance problems in many apps. This lead many reviews and I must assume educated consumers to disable this feature.

Other than tollerance for spyware does this have any real advantages?

Didn't we hear some rumblings on this count from AMD? When does their roadmap state this stuff'll be ready to go?

Re:4 CPU's

[Kris_J]

There's a utility out there somewhere that lets you leave hyperthreading on for the OS, but only show physical CPUs to the applications. Can find a link though -- didn't realise when I found it before how important it might be. (I use AMD myself, but mostly because I like the nForce2 chipset.)

Re:4 CPU's

[TheRaven64]

SMT (HyperThreading) provides the OS with two CPUs whose capabilities vary (the first one can do anything, and the second one can do anything the first one is not doing). If you use a classical SMP scheduling algorithm on these virtual CPUs then you are likely to get a performance hit, since it may schedule (for example) two integer intensive threads to one physical CPU, which will generate resource conflicts. If the scheduling algorithm is SMT aware then this problem goes away.

Re:4 CPU's

[mrm677]

Actually I believe you are wrong here. Intel split the reorder-buffer in half, split the physical register file in half, and fetches from each thread every other cycle. If I am not mistaken, the latter point means that each thread gets equal execution resources as once instructions have been fetched and decoded, dispatch and execute don't care which thread an instruction belongs too. Like I said, I think this is the case but am not sure.

From my friends in the architecture community, Intel's SMT implementation is sort of half-assed.

On the other hand, IBM's Power5 also fetches from each thread every other cycle, however it shares a reorder buffer physically (but of course not logically).

Re:4 CPU's

[Kupek]

It's actually Symmetric Multithreading. HyperThreading is Intel's marketing term.

Re:4 CPU's

[Kupek]

D'oh. Make that Simultaneous.

Cool :)

[execom]

I mean, Hot !

Will it come with the Prescott Survival kit ?

IBM chugging along

[Anonymous Coward]

I've heard that IBM already has ridiculous 8-core POWER5 prototypes. You'd think by late 2005 they'd have knocked out an Altivec-enhanced, dual-core POWER5 derivative for Apple. Though, given their troubles at 90nm on the PPC970, maybe we should be waiting until early to mid 2006 to see that.

Re:IBM chugging along

[GlobalMind]

Yes, meanwhile IBM is on their second generation dual core chip, POWER5 -- now available in eServer i5 systems -- shipping TODAY.

The way POWER4 was packaged for the higher end boxes, you have what they call a Multi-Chip Module (MCM) with 4 POWER4 processors on-board. This means each MCM was an 8-way.

Now, for POWER5, they have added the Dual-Chip Module or DCM. With the i5 model 570, you can get a 1/2 way or 2/4 way box. If you buy the 1/2 way, you have one DCM installed...and if you buy the 2/4 then you get two DCMs.

POWER5 has what IBM calls Simultaneous Multithreading -- SMT, which is the same type of idea as Hyperthreading. Essentially if the application supports multithreading, it will functionally see twice the processors...but this is a logical thing...a 4 way is still a 4 way...not an 8 way.

Now, having said all that....never underestimate IBM development labs. I hear POWER6, 7 and maybe 8 are already out in development.

TGM.

Re:IBM chugging along

[CommieOverlord]

And Sun already has dual-core UltraSparcs. Ramping up to 32-core chips in a couple years.

Re:Linux support?

[G-funk]

Worst. Trol. Evar.

Because intel's always so reticent in giving out the instruction sets and specs for their processors. That'd be a really smart business move, and the #1 way to attract developers.

But you should see the heatsinks

[Groo Wanderer]

I caught a pic of the heatsink for this beast at Computex, so it must be real.

http://www.theinquirer.net/?article=16426

-Charlie

(for the humor impared, think humor - haha, not humor - I don't get it)

Re:But you should see the heatsinks

[Groo Wanderer]

No, it is clearly for 1U servers you troll. The laptop one is almost a cm shorter. There is also a new line of fitted luggage to go with the thin and light versions.

-Charlie

200 Watts?

[ahfoo]

Well clearly this demonstrates that Intel really does get the best smoke on the market today. That shit has got to be pricey, because the whole joint is stoned out of their heads.

Let's do some math for them. If we leave our PCs on all day --and that is why we have 24/7 broadband connections isn't it-- that's 5KW/Hrs a day.
At 15cents KW/Hr it now costs seventy five cents a day to have an Intel CPU. That's twenty bucks a month.
But do you get 15cents per KW/Hr lately? Check your bill, you might be closer to twenty cents. A buck a day. Hey, I running the Intel PC costs almost as much as broadband. Perhaps they should include free broadband connections with these things.

Re:200 Watts?

[gunix]

Power consumption is a very important thing, and I'm not only talking about the heat. 200W, and all you make are a few word documents, send a few emails. And the machine is turned on 24 hours per day.

But on the other hand, they claim making love consumes over 300W (per individual), so it might acctually be a good power saving thing to surf for porn instead.

Re:200 Watts?

[NubKnacker]

You seem to be forgetting the fact that your CPU takes that much power only when it's actually working, and not when it's sitting idle. So unless you got SETI running, your power bill shouldn't be that high. Ofcourse if you have SETI running, then it could also be the aliens drawing power from your computer.

Re:200 Watts?

[Wedge1212]

that silly halt command throws everyone for a loop :)

Re:200 Watts?

[timeOday]

Ah, but only for half the year. In the winter my PC is just taking some of the load off my central heating.

Then again, in the summer I have to pay once to power the PC, then again for the AC to pump out the waste heat. That sucks.

Re:200 Watts?

[Reziac]

Methinks you meant "halt and catch fire" ;)

Re:200 Watts?

[Reziac]

Someone I know, who lives in the chilly Seattle area, actually heats one room of her house with nothing but an AMD-based box, I kid you not. It puts out enough heat that the room is warm even in winter.

Compared to an electric space heater, it's probably pretty cost-effective!

Stare of play in CPU design.

[funkytwig]

1 processor core technology is causing heat consernes in thin core. The point is as the core gets thinner the power required to stop lekage across a ever thinner insulation layer increses. A couple of jumps thinner and we would have chips that require the power of a houshold iron. Multi-core is a solution to this problem, maybe Intel are not using very thin core technology to reduce heat in there multi-core processors. There was a very interesting article about this in New Scientist but I dont think it was one they put on the web for free ;-(. (sorry I posted this as a reply to someone elses article but am hanging it of the original post as it seems relevent).

Re:Stare of play in CPU design.

[leperkuhn]

The extra power is because of wasted electrons that leak.

Moderators: WTF?

[trezor]

Troll or joke, that's really a subjective matter, but the guy who modded this "Informative" should give up moderating priveleges ASAP.

Multi-Processor boards

[Apocalypse111]

I am curious what the maximum number of processors that WinXP supports is, or for that matter Longhorn. The reason I ask is because, with this technology, seeing 1 proc as 4, then logically it would see 2 proc's as 8. 8 processors? Sounds a little rediculous, but how exactly would XP handle it? Will we have to wait until SP3 or something? Would it see only 4? Would you have to turn of hyperthreading? Will there be boards built to support multiple processors with this new chip type? The more and more I think about this, the more I think this is a distributed computing project in the works. And please, no jokes about Beowulf clusters...

Re:Linux support?

[Short Circuit]

I don't think active support will be required.

With hyperthreading, the OS recognizes one physical processor as two CPUs, with nothing more than standard SMP behavior.

In fact, the Linux scheduler had to be tweaked in order to weaken restrictions on passing processes from one logical CPU to another, if both were on the same chip. Otherwise, hyperthreading actually caused a performance loss when treated as an SMP system.

Dual Core vs. Single core

[totoanihilation]

I'd take a DUAL dual-core 2.5GHz G5 powermac over a dual single-core 3GHz anytime.
It would probably be less of a technical challenge as well, and would follow the "GHz doesn't matter" philosophy the POWER(tm) manager said a few days ago.
The 90nm process encounters problems at high clock speeds. So, bring on more efficiency at lower clock speeds!

Re: Superscalar

[Short Circuit]

I prefer Ars Technica's Understanding Pipelining and Superscalar Execution...

Besides, I feel HT exploits the fact that the processor is pipelined more than its superscalar nature.

Re:To meet Longhorn

[Eu4ria]

Why is the parent a troll ? The supposed minimum spec for longhorn suggested needed a dual core processor - story was slashdot a few weeks back - too early to find the link i havent had my coffee yet.

NOT snazzy

[JamesR2]

Be careful with HT. These fake CPU's can actually drag down overall throughput due to the fact that they can't do everything a "real" CPU can. My theory is that I/O is one of them. Caution databases servers.

Re:NOT snazzy

[JamesR2]

Ok point taken; I should have said SOME scenarios are no better. From your link: "These tests also show that the performance improvement of Hyper Threading is not always predictable. In the Web and Ad DB tests, the performance change varied from an increase of 3% to a decrease of 5%.". What I am saying is caution - don't assume automatically that HT is better.

Everyone missed the most important news...

[dark404]

Intel reportedly said that with the HyperThreading technology enabled operating systems will report availability of four microprocessors into the system when a single dual-core Prescott is installed. Representatives also confirmed that future Prescott products will feature 64-bit capability.

Didn't intel say previously they weren't going to make 64-bit desktop chips?

Re:Everyone missed the most important news...

[Surt]

It's unlikely anyone officially speaking for intel would say this, as the potential need for 64bit desktops has been clear for long enough that it would look pretty stupid for them to make such a claim. Not that it rules out the possibility, but it would be exceptionally dumb.

The need for 64bit desktops to deal with multigigabyte multimedia is clearly in our near future. It may be 4-5 years before most desktop users feel the need for a 64 bit system, but it seems unlikely to be much longer than that.

You may have heard intel claims that they saw no immediate need for 64bit on the desktop. By the time they release 64bit on the desktop, those claims will have been nearly a Moore's Law old.

Re:heat issues

[AliasTheRoot]

But thats a Software / OS issue right? It it strikes me if MS realises this with Intel prodding them then in a few years it wont be such an issue.

Strained SIlicon issue

[charnov]

Actually, the big jump in heat for the Prescott cores is from Intel use of only starined silicon in manufacturing. By creating a strained lattice for the silicon, you increase the likelyhood of current leakage (hence more heat). This is why AMD and IBM went with silicon on insulator and added strained silicon later (the SOI process helps to mitigate the leakage in strained silicon).

Here's a simple primer

Yeah...the power chip is $20,000

[charnov]

The chip all by itself is $20,000 and is about the size of ones hand. It is 4 physical, 8 logical with 144 MB (that's right...megabytes) of cache.

Re:Yeah...the power chip is $20,000

[Wesley Felter]

I think you're talking about the POWER5 multi-chip module (MCM); it actually has four processor chips, each with two cores. It also has four cache chips. Check out a picture.

Re:Linux support?

[Anarke_Incarnate]

Ask them for the specifics of their new EFI "BIOS" so you can work with it.

Mod Parent Up (+5 Slick)

[LordKazan]

Parent deserves to be moded up for slick moves on a multiprocessor system :D

Re:heat issues

[timeOday]

As for two slower cores instead of one faster one, that definitely can save power and increase speed. The problem is that most consumer software is single threaded, so will run slower than the single high speed core alternative.

But thats a Software / OS issue right? It it strikes me if MS realises this with Intel prodding them then in a few years it wont be such an issue.

It is not just an integration issue. Sure some things are easily parallelized (like a Photoshop filter that does the same thing to every pixel) but others are not (e.g. use multiple CPUs to make a browser or word processor "snappier"). Anyways, introducing parallelism just to utilitize the hardware complicates the software, increases the prevalance of hard-to-catch timing bugs, and adds the overhead of synchronizing processors and memory.

As a developer, it's the more difficult programming model that bothers me. I don't have phobia about multithreading - just the opposite, I've done it enough to know that writing correct software is much harder when you have to worry about concurrency.

Besides correctness, there is performance. Writing software to fully utilize two processors is MUCH harder than to fully utilize one. For instance, you might write a multi-cpu aware game by doing the physics in one thread and the graphics in a second thread. But unless those tasks happen to require exactly the same CPU power, you will not achieve full utilization. So you resort to partitioning the functionality in some unnatural way to make it balance.

When I discuss this with my office mate he says: "Great! That will keep us in the job." And I guess he's right. I think there will be a lot of opportunities for new language features (or languages) to exploit all this parallel hardware. And of course a lot of education for software developers. (Not that parallel processing is brand new, but networking wasn't brand new when the Internet hit either - broad acceptance changes things).

So what's the upshot to the consumer? Simple: the whole parallel computer is less than the sum of its parts. If it were easy to make anything faster by throwing more processors at it, multiple processors would have become ubiquitous years ago.

Not what ars says....

[222]

I read this on Ars last night, so i would take it with a grain of salt....

Ars Technica: The PC enthusiast's resource: "Now, your guess is as good as mine, but it sounds like this 'Intel employee,' whom the report identifies as a marketing manager, was talking out the rear, as we say in Beantown. HyperThreading, for what it's worth, might 'take off' in the future but right now what's taking off is the competition. Now, Intel may have some mojo up its sleeve that hasn't made its way through my sources, but I'll be rather surprised to see dual core Prescotts in a year's time unless Intel has managed to patent a dry ice freezer for cooling purposes. The future is quite clearly the Pentium M, unless Intel has solved power leakage problems and not told anyone about it (which is possible, but unlikely). My best guess with the information at hand is that this is Intel marketing speaking, and Intel marketing isn't going to tell you that Prescott doesn't have a future. Designing a dual-core, HT-enabled CPU that won't scale just doesn't make sense, and I can't imagine Intel doing it. "

Re:To meet Longhorn

[cens0r]

Those weren't minimum specs. Those were what Microsoft believed the average computer was going to be.

Prescott?

[jpellino]

So is this
"Prescott" as in "our only president's alleged nazi-financier grandfather"
or
"Prescott" as in "if we keep naming chips and OSs after stuff in the southwest then the moldy bunch in rain-soaked Washington will keep writing whatever we tell them to on the lure of actually getting dry socks"?

Just wondering.

They have to do something...

[Anonymous Coward]

IBM's PPC is going to crush them and there's no denying the continued growth of AMD in the market. Intel is on shaking ground and will be for quite some time if they do not react in an intelligent and proactive way.

The future is now

[TimeZone]

IBM is now selling Power5 boxes. Each Power5 chip is dual-cored and dual-threaded meaning 4 CPUs per chip. They will (soon?) be selling machines with Power5 MCMs, which are 4 Power5 chips on a single module. That's right, a sixteen-way in the palm of your hand. (well, hands, really, they're big. Look at the picture.)

End of Advertisement

TZ

Oh no...

[null-sRc]

...thanks to HyperThreading...

yeah thanks hyperthreading for making my games unstable.

has anyone else encountered this? i've seen two computers with lots of strange bugs and system instabilities; untill you disable hyperthreading and then blam! no crashes since... :| :| :| !!!

(two different motherboards, one asus, and one intel)

Re:Oh no...

[fredhero]

Sounds like the game might not be particularly thread safe and was probably never tested on a real dual cpu box.

mplayer, however

[bcboy]

... will still only use one cpu. ;p

Not really a happy ending

[sufehmi]

Heh, thanks; very interesting mod tag :)

Unfortunately, it doesn't really have a happy ending - originally I installed VMware (and the virtual servers) on Linux. However, turned out our FM [1] (Facility Management provider) doesn't support Linux.

So I had to reinstall the OS and VMware and run the whole bunch on Windows [2] instead :(

[1] FM provider is kinda like a co-lo company - you give your servers to them, and they'll take care of them for you. Except that they're much more expensive...

[2] From my experience, it seems that VMware 4.x workstation is more stable on Linux than on Windows. I've experienced several quirks when running VMware on Windows, so I try to avoid Windows for this whenever possible.

A side note: Contrary to some believe/comments, VMware 4.x workstation is quite powerful.
I've heard people saying that it won't run on more than 1 GB of RAM / multiple processors - well, one of our server is a dual-Xeon and have 3 GB of RAM, and VMware 4.x is running happily on it.

Re:Not really a happy ending

[SmittyTheBold]

The RAM limitation was fixed in 4.5, which is a relatively recent release.

Cool.

[IPFreely]

Pretty soon, they'll start making chips with water ducts built in. You put the processor on the MB, attach water pipes directly to the in and out nozzels on the processor itself, and start up the water cooling pump *BEFORE* you turn on the CPU power.

IBM is already there...

[AIXmaster]

IBM got POWER5 with two processors cores on one
physical chip; with each core SMT enabled.
So POWER5 have four "processors" on one chip.

Re:heat issues

[drinkypoo]

some things are easily parallelized (like a Photoshop filter that does the same thing to every pixel) but others are not (e.g. use multiple CPUs to make a browser or word processor "snappier"

Using a multithreaded design makes everything slower (from slightly to dramatically) on uniprocessor systems and should in theory make nearly everything faster on multiprocessor systems. If your gui and your rendering library are in separate threads and you have multiple processors you're going to see a speed improvement.

Unfortunately multithreading seems to still be difficult to pin down. It seems like every OS has a dramatically different method of handling threads, further complicating the problem...

Dual Car Prescott

[pommiekiwifruit]

So sorry, but it had to be said...

(so is an individual core a Jaguar?)

Re:Linux support?

[Anarke_Incarnate]

No, I mean the source code without an NDA. I occasionally talk to a developer with the Linux BIOS group and he is the actual person who has been trying to get that information. I don't believe he has gotten it. One would hope they would love to have more people able to use their CPUs, but they want the control more.

1 running, 3 idle

[octogen]

That makes 4 'virtual' CPUs. Anyway, as most PC users are running NT (or one of its successors), and as locking is anything but fine-grained on NT, while one thread running on one CPU will call some NT kernel function which locks a CRITICAL_SECTION, three other threads running on the remaining three CPUs, who call some other kernel function, will wait for the same CRITICAL_SECTION; effectively, only one CPU gets any real work done.

That's why some other OSs may perform similar on uniprocessor hardware, but may substantially outperform NT on multiprocessor hardware. For example, if you take a look at the SunOS kernel, you'll find much more efficient locking techniques (faster mutexes, lots of rw-locks, more fine-grained structures == better parallellization).

Multiprocessor hardware performs well with technologically sophisticated software only; I mean: it doesn't make sense to increase the number of processors if you don't use all of them simultaneously.

Re:1 running, 3 idle

[drsmithy]

That makes 4 'virtual' CPUs. Anyway, as most PC users are running NT (or one of its successors), and as locking is anything but fine-grained on NT, while one thread running on one CPU will call some NT kernel function which locks a CRITICAL_SECTION, three other threads running on the remaining three CPUs, who call some other kernel function, will wait for the same CRITICAL_SECTION; effectively, only one CPU gets any real work done.

What makes you think locking on NT isn't fine-grained ? It was designed from the start for SMP and has had ~15 years of development since. Nothing I've ever read about its design or development would suggest it doesn't have excellent fine-grained locking.

Re:heat issues

[RzUpAnmsCwrds]

"For instance, you might write a multi-cpu aware game by doing the physics in one thread and the graphics in a second thread."

Actually, most games are already written like this, but they are not multithreaded. Most games rely heavily on the GPU for - well - graphical processing. More and more of the work has been offloaded to the GPU in recent years. The CPU is for physics and other game functionality.