• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

100 Core Processor

Joined
Jul 2, 2009
Messages
398 (0.07/day)
Location
New Hampshire
System Name Small Package (Build in Progress)
Processor Phenom ii X4 965 Stock
Motherboard Biostar TA890GXE
Cooling Antec H20 620
Memory 4GB of G-Skill
Video Card(s) GTX 470( soon )
Storage ocz 30GB SSD, 500GB WD Cavier Green, 750GB WD Cavier Green
Case Nzxt Vulcan / open box
Audio Device(s) Onboard
Power Supply Sparkle 750
Software Windows 7 64 professional
4x faster then the Nehalem-Ex :twitch:


http://www.wired.com/gadgetlab/2009/10/tilera-100-cores/

"Forget dual-core and quad-core processors: A semiconductor company promises to pack 100 cores into a processor that can be used in applications that require hefty computing punch, like video conferencing, wireless base stations and networking. By comparison, Intel’s latest chips are expected to have just eight cores.

“This is a general-purpose chip that can run off-the-shelf programs almost unmodified,” says Anant Agarwal, chief technical officer of Tilera, the company that is making the 100-core chip. “And we can do that while offering at least four times the compute performance of an Intel Nehalem-Ex, while burning a third of the power as a Nehalem.”

The 100-core processor, fabricated using 40-nanometer technology, is expected to be available early next year.

In a bid to beat Moore’s law (which states number of transistors on a chip doubles every two years), chip makers are trying to either increase clock speed or add more cores to a processor. But cranking up the clock speed has its limitations, says Will Strauss, principal analyst with research and consulting firm Forward Concepts.

“You can’t just keep increasing the clock speed so the only way to expand processor power is to increase the number of cores, which is what everyone is trying to do now,” he says. “It’s the direction of the future.”

In fact, Intel’s research labs are already working on a similar idea. Last year, Intel showed a prototype of a 80-core processor. The company has promised to bring that to consumers in about five years.

Tilera, a start-up that was spun out of the Massachusetts Institute of Technology, started in 2007. It says its product will be available in the next few months, which means the company, if successful, will have gone from zero to shipping a powerful chip in just about three years — a very fast time frame in the semiconductor world. That’s because it has created a chip architecture that removes the challenges present in Intel’s x86 design.

As the number of cores on a chip multiplies, a major challenge is how to connect the chip to memory without choking up the processor. That’s why Agarwal says Tilera has used a mesh network architecture. It eliminates the “on-chip bus interconnect,” a central intersection found in most multi-core CPUs through which information must flow through to get between the cores of a chip. That central interconnect presents bandwidth issues of its own, and also forces engineers to limit the number of cores on a chip to avoid information gridlock.

Instead, Tilera places a communication switch on each processor and arranges them in grid-like fashion on the chip. Because the overall bandwidth is greater than that of a central bus, and because the distance between individual cores is smaller, Tilera says it can cram in as many as 100 cores on a processor without running into bus-bandwidth congestion.

Each core has a full-featured, general-purpose processor that includes L1 and L2 caches, and a distributed L3 cache. The cores are overlaid with the mesh network, which provides extremely low-latency, high-bandwidth communications between the cores, memory and the processor’s input and output.

“If you need huge computing power, say for instance to encode and decode multiple video streams, our processor can do it at much more efficiency than Intel chip or a digital signal processor,” Agarwal says.

And unlike GPU-based computing systems, programmers can recompile and run applications and programs designed for Intel’s x86 architecture on Tilera’s processor.

“Tilera has put forth a novel approach to massively parallel programming,” Strauss says. “The 100-core processor is closer to a generic processor than anything else we have seen before.”

Don’t expect it to run Windows 7 on it though. For that, consumers will have to wait for Intel’s version in a few years."
 
Son of a b!tch, I should've just used a Phenom for the next few years!
 
Wasn't this an MIT project from last year, or possibly the year before?

Edit, guess I should read the entire article before posting useless comments.
 
I wonder much $$$ this baby will cost?
 
100 cores and only 4 times faster? That means it takes 6.25 of its cores to match one of Nehalem's cores. It sounds like something that goes in routers and not computers/fully programmable environments.
 
100 cores and only 4 times faster? That means it takes 6.25 of its cores to match one of Nehalem's cores. It sounds like something that goes in routers and not computers/fully programmable environments.

But it's a third of the power consumption!
 
unless it is going for an IBM supercomputer, who could possibly use that much core, maybe the next Final Fantasy - spirits within 2 ;) (Please dont do it Sony, for the love of kittens)
 
But it's a third of the power consumption!
Because takes three times longer to do the same tasks! :roll: Oh, 45nm vs 40nm helps a little...
 
The real mind blowing thing is that they did this in 3 years
 
to bad it wont work on W7 :( but 100 cores 4x times intels seems odd must be like 100 core at like 0.8ghz with only .8v and 50watts load or something like that idk im just coming up with numbers but only if it was like 15+ times faster but still crazy 100 cores on 40nm
 
a 100-core computer processor that offers four times the power of Intel's Nehalem-Ex, while using a third of the power.This is a general purpose chip that can run off-the-shelf programs almost unmodified.
 
This is a general purpose chip that can run off-the-shelf programs almost unmodified.

Yes, apart from the fact that the news article I read yesterday mentioned they ported several programs themselves to work on their own Tile architecture. Mostly database/webserver related things.
100 slower cores will outperform an i7 by far when handling thousands of connections. Though you can't really compare chips like this to our general purpose home toys.

As for pricing, it says $400-$1000 per chip here.
 
But will it play Crysis at high res. ?
 
what mobo do they expect it to work with?
 
you guys realize that this is not x86 compatible, right ? scaling a 100 core design with a single cpu-typical load is difficult. if your problem parallelizes fine go use cuda or opencl.

if there is demand for it intel could build a 100 core, 1000 core, 1 million core cpu too. but for current x86 apps there is no way you will get any kind of benefit. many programs dont even use more than one core, imagine the drama if your great shiny new cpu runs only at a fraction of the performance and there is nothing you can do about it.

"can recompile and run applications and programs designed for Intel’s x86 architecture on Tilera’s processor."
that means they give you a compiler (probably buggy as hell) that you can use to port your c/c++ programs to tilera. but you cant recompile windows or everything closed source.. whether they can pull it off to add a new arch to the linux kernel we will see, just trying to recompile the linux kernel wont work either without proper code changes.

and who's gonna fabricate their new cpu ? only tsmc and amd come to mind.. we'll see how well that works.
 
I read somewhere, each core will be clocked in the ballpark of 1,5 GHz. This is more of a GPU/CPU hybrid than a full fledged CPU. It has the paralerism of a GPU but lacks the complex pipeline of a CPU.
 
I read somewhere, each core will be clocked in the ballpark of 1,5 GHz. This is more of a GPU/CPU hybrid than a full fledged CPU. It has the paralerism of a GPU but lacks the complex pipeline of a CPU.

I suppose it is an evolution of the Cell concept. Which in the end is similar to Fusion and Larrabee. One thing that I like about this kind of projects is that they completely depart from x86 or PPC. And it's a pity that they will probably fail on the market like many others. I would like to see a non x86 CPU go mainstream. Academic work has demostrated again and again that x86 is an archaic ISA for te time. Claims say that with a new one made from the scratch, with no need for back compatibility they could create CPUs that are 10x faster. It makes sense. We've been using the same ISA for 30 years, thre musy be something better, but we will never know...
 
Hey look, a non-x86 Larrabee!
 
Cell has a general processor with sub processors attached to it. Intel Larrabee is an x86 GPU. AMD Fusion is a lot like your current Core i5 processors but more refined.

This is like taking the Broadcom processor out of a router, stripping it down to barely anything, sticking 100 of them on a single die, tie them all together, and finally, program them to do something.

It more resembles an multi-core ARM processor than anything else.

It could do very well if they put it in a stand alone device (like a SMTP server, web server, or network switch) and market it to corporations to cheaply fill a need. If they try to market it similar to an x86 processor, it will fail.


AMD killed the potential for a new architecture with x86-64. The requirements of 64-bit computing would have made it an ideal time for a paradigm shift. People like backwards compatibility so much that the only way to change it is to have no alternative. AMD killed that possibility with the Athlon 64 (AMD64).
 
Well I hear 128 bit is right around the corner anyway :D
 
For network devices, yeah. IPv6 is a PITA to handle with x86 because of the 64-bit integer limit. :(
 
i feel like tilera got beat to the punch by GPGPU
 
Larrabee is better for the "consumer". There are millions of x86 developers in the world, and the compilers just need a few simple extensions to work. It is a better solution for consumer multimedia, or for gaming and physx-like code.

This Tilera thing is like clearspeed or the transputers of the 1980's. These products already exist in the market. So what has Tilera got that doesnt already exist? I dont think it will go much beyond the concept stage.

Useful. Specialist market. Not likely to make it into your home anytime soon.

Clearspeed focuses on floating point code. Whereas this thing is much more basic and has extremely simple instruction set, for very basic tasks, done in parallel 100x. Good for database, webserver, cracking security systems, or control applications. Not more.
 
I read somewhere, each core will be clocked in the ballpark of 1,5 GHz. This is more of a GPU/CPU hybrid than a full fledged CPU. It has the paralerism of a GPU but lacks the complex pipeline of a CPU.

I saw that on the Toms Hardware article about this same processor. They also have 16, 32, and <100 core processors as well. Pretty crazy they did it in 3 years though. Smart guys:nutkick:
 
Back
Top