Intel demonstrates 80-core processor

Now that the Megahertz race has faded into the distance (we hear it was a myth), Intel is well and truly kicking off the start of a multi-core war with the demonstration of an 80-core research processor in San Francisco last week. It's not the first multi-core processor to reach double figures -- a company called ClearSpeed put 96 cores onto one of its CPUs -- but it's the first to be accompanied by the aim of making it generally available; an aim that Intel hopes to realize within a five year timeframe. The long time frame is required because current operating systems and software don't take full advantage of the benefits of multi-core processors. In order for Intel to successfully market processors with CPUs that have more than say, 4 cores, there needs to be an equal effort from software programmers, which is why producing an 80-core processor is only half the battle. On paper, 80-cores sounds impressive, but when the software isn't doing anything imaginative with them it's actually rather disappointing: during a demonstration, Intel could only manage to get 1 Teraflop out of the chip, a figure which many medium- to high-end graphics cards are easily capable of. The multi-core war may have begun, but the battle will be fought with software, although that's not to say that the hardware side has already been won: apparently the test chip is much larger than equivalent chips -- 275 mm squared, versus a typical Core 2 Duo's 143 mm squared -- and Intel currently has no way to hook up memory to the chip. Hopefully half a decade should be long enough to sort out these "issues."

[Thanks, Michael]

[js @ Feb 11th 2007 7:30PM]

Yeah, but can it play Doom?

[Dog Guy @ Feb 11th 2007 7:38PM]

Sounds really good so far but wouldn't multi-cores help with video processing speed? To import High Def video takes a alot of processor power. This should take advantage over the multi-core processors. I'd like to see how much more effiecient it is. Rendering is also processor intensive. What about BUS speed? How will that play out in multi-core processors?

I am getting tired of buying a computer and it being outdated in weeks but I guess if software makers aren't taking advantage of the updated processors it really is insignificant. It makes sense to buy a computer currently using available software not what software may or may not be available.

[E @ Feb 12th 2007 6:09PM]

Dog Guy @ Feb 11th 2007 7:38PM
"Sounds really good so far but wouldn't multi-cores help with video processing speed? To import High Def video takes a alot of processor power. This should take advantage over the multi-core processors. I'd like to see how much more effiecient it is. Rendering is also processor intensive. What about BUS speed? How will that play out in multi-core processors?"

Somewhat, but your mbiggets problem is throughput, not CPU speed. Getting the frames into the box, then shuffling them around is a killer. Sure CPUs help a lot, but you need to have really fast RAM and a raid array.

You think you have problems, try rendering 3D and compositing in HD or film res!

[Kenban @ Feb 11th 2007 7:45PM]

Not quite sure what the point to this chip is right now but I have a feeling on what something like it could be used for in the future. As it says in the article "The computing elements are very basic and do not use the x86 instruction set used by Intel and Advanced Micro Devices' chips, which means Windows Vista can't be run on the research chip. Instead, the chip uses a VLIW (very long instruction word) architecture, a simpler approach to computing than the x86 instruction set."

When I see chips like this and the Cell two things come to mind supercomputers and graphics. With graphics cards and Direct X moving to being programmable I expect we are going to start seeing GPU's which use a lot of cores instead of the current setup. Supercomputers don't need to have a huge instruction set or anything really complex normally they just need the ability to do a lot of calculations quickly simple cores like this would be perfect for the job.

[Grandalf @ Feb 11th 2007 8:14PM]

Mmmm yes ... concurrency programming is such a joy.... in erlang atleast... please go away C# concurrency :p

[cmonkey @ Feb 11th 2007 8:03PM]

What I'd like to see is on-chip or even on-die FPGAs. Or, more realistically, a socketed FPGA that would go along with AMDs plans to allow coprocessors on dual boards. Since most tasks don't see much improvement from parallelization, being able customize a single fast FPGA to the tast could be more useful. Imagine an FPS reprogramming the FPGA to run their own particular brand of AI on it, etc.

[Pete @ Feb 11th 2007 8:36PM]

"Imagine an FPS reprogramming the FPGA to run their own particular brand of AI on it, etc."

Not likely -- remember that pretty much all current OS's are multitasking. What happens when the FPS finishes its execution time and another application executes for a time -- does that second app get to reprogram the FPGA every time too? And back again?

Constant reconfiguration doesn't seem too efficient to me, not to mention the added software complexity.

For the story itself:

Five years is not going to be enough for this. We don't even have adequate tools to write such massively parallel programs with ease and designing/building such tools could well take that long (or longer). Even if we end up with compilers capable of automating much of the hassle (unlikely -- I'm expecting some app-level changes, maybe even specialist programming languages), debugging a highly parallel system can be an absolute nightmare.

You can't rely on the OS to take advantage of that many cores automatically either, as (at least at present) I'm sure they just move processes/threads onto different cores. If you just have one or two running applications this will not work so well unless the apps are designed specially, etc... You see this now with some older (faster) single-core processors still outperforming their dual-core brethren at some tasks.

[ShortFuse @ Feb 11th 2007 8:04PM]

yeah, i think the chip is too big =P

[Mark @ Feb 12th 2007 12:40AM]

Conrad, that's 275 square mm, not 275 mm squared unless you write it: 275 (mm squared).

[pate @ Feb 22nd 2007 4:05PM]

"Too big????".. due the icture is the wafer and not the chip! moron!

[Bman21212 @ Feb 11th 2007 8:20PM]

Yes, some graphics card (8800 anyone?) can go over a teraflop now. But look at the size of those things , vs the size of a processor. Even at 275mm squared, it is a fraction of the size of a graphics processor

[helloUser @ Feb 11th 2007 8:22PM]

ugh, still no news if penryn processors will remain socket 775 and 771 for xeons....

[Macchar @ Feb 11th 2007 8:31PM]

Tough going for AMD!

[MikeTLive @ Feb 11th 2007 8:41PM]

Take a look at the processors from SunMicrosystems.
http://www.sun.com/processors/index.html

Sun Microsystems has their "Niagara" chip with 8 cores executing 4 threads = 32way.
the Niagara2 chip doubles that to 8 cores executing 8 threads = 64way.
http://www.sun.com/processors/niagara/


the "Victoria Falls" systems are dual processor = up to 128way.
the "Rock" chip due in 2008 will have 16 cores and an undetermined number of threads however expect it to be no less than 8 = at least 128way per chip.

theres a ton more articles about them on http://www.sun.com and on cnet

[Matt @ Feb 11th 2007 9:11PM]

Two of these in a macpro = maximum pwnage!!!

[Keaton @ Feb 11th 2007 11:31PM]

Yeah... 160 cores and no memory... Total pwnage for sure...

*rolls eyes at the absurdity of mac fanboy's*

[Grant @ Feb 11th 2007 10:22PM]

Sony has found their next uber-chip for the PS4. watch out Microsoft!
lol

[Giltronic @ Feb 12th 2007 11:38AM]

thats pretty fukin funny, id think we you right, that is if sony actually realeases a ps4...

[Kenban @ Feb 11th 2007 11:26PM]

Sony already has their uber chip for the PS4 and Microsoft has nothing to worry about. The whole idea behind the design of the cell is that they can just continue adding more cores. The Xenon in the 360 actually used 3 of the PPC cores which are in the cell (at least thats the theory since all the specs match up). I expect Microsoft will do something similar with its next generation xbox by just using more PPC cores while Sony will add both SPE's and PPC cores. I think Microsofts approach is actually better for developers since all the cores are the same.

[Ryan @ Feb 12th 2007 12:24AM]

More techy details and the likes here: http://www.pcper.com/article.php?aid=363

[allaun @ Feb 12th 2007 1:15AM]

Now you can run vista with ONLY 80% load on idle! :p

[Tim @ Feb 12th 2007 1:39AM]

Will it blend? ;)

[Stranger @ Feb 12th 2007 2:21AM]

I guess NSA and many other security/intelligence services will be more than happy to support this project. A look at Deep Crack will show us what multi-core these chips can be used for - and it's not Quake X or Unreal Tournament 2015. Time to increase that bit-length used in cryptos maybe?

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

[Kizul Emeraldfire @ Feb 12th 2007 3:55AM]

JS Wrote:
‘Yeah, but can it play Doom?’

The question NOT, "Can it play Doom?" It most certainly can play Doom. The question IS: Can it play Battletoads?

[Brendon @ Feb 12th 2007 12:27PM]

Good thing the I/O bottleneck will keep 70+ of these cores idle.

[rektide @ Feb 12th 2007 7:11PM]

nvidia has 128 processor chips already. running half the speed, but yes, they play doom, and yes you can buy them in stores nows.

[Bryce @ Feb 12th 2007 8:19PM]

This is the same company that told us Netburst (Pentium 4) would scale to 10GHz w/in 5 years. This is just hype.

[JustElite @ Feb 19th 2007 4:36AM]

Here is the video: http://pcnews.tv/147/prototype-80-core-chip/

[eliot1785 @ May 2nd 2007 1:08AM]

Saying it only got 1 TFLOP is somewhat misleading. Processors always have lower TFLOPS than GPU's, because they do other kinds of calculations. A 1 TFLOP processor is a lot more impressive than a 1 TFLOP GPU.