Correction: The Microsoft and Intel press conference is scheduled for Tuesday.
Microsoft and Intel on Tuesday are expected to launch a joint research initiative to tackle programming for multicore processors.
The two PC industry giants sent out a media alert saying that they will host a teleconference to announce the research venture.
The need for more research stems from the emergence of processors with two or more processing units, or cores, which have now become mainstream. With multiple cores, chip designers can boost a machine's processing muscle in a more energy-efficient way than by increasing the processor's clock speed.
But multicore technology poses significant challenge to both hardware and software providers. Without writing programs to be optimized for multicore processors, applications will not benefit from the added chip power, or could run slower than previous chips.
Both Microsoft and Intel--as well other IT companies--have made programming tools for multicore processing a high priority in terms of product development and research.
Andrew Chien, the director of Intel Research, and Tony Hey, corporate vice president of external research at Microsoft Research, are scheduled to host the media announcement.
The amount of funding for the research, which several universities bid on, will be $2 million annually for five years, according to the Journal.
A report from EE Times said that about 14 faculty members at the University of California at Berkeley lab started work on the project in late January. The researcher will focus on creating development frameworks that make it easier for programmers to parse out computing jobs so that they can be done in parallel by processors with several cores, according to the report.
Essentially, the lab is aiming to define a way to compose parallel programs based on flexible sets of standard modules in a way similar to how serial programs are written today. The challenge in the parallel world is finding a dynamic and flexible approach to schedule parallel tasks from these modules across available hardware in complex heterogeneous multi-core CPUs.
The group believes developers could create a set of perhaps a dozen frameworks that understand the intricacies of the hardware. The frameworks could be used to write modules that handle specific tasks such as solving a matrix. New run time environments could dynamically schedule the modules across available cores of various types.
More details to follow after the 10 a.m. PDT press conference on Tuesday.
Advanced Micro Devices on Wednesday released the source code aimed at making it easier to build media applications for x86 multicore processors.
The AMD Performance Library, which is now called Framewave, is available on SourceForge.
The framework is a collection of routines that optimize performance of media-oriented applications on multicore processors. For example, a developer could link to a library that optimizes video display on a specific graphics card.
By starting a project at SourceForge.net, the chipmaker will allow programmers to customize and further optimize the software, which will work with compilers for Linux, Windows, and Solaris, said Margaret Lewis, AMD's director of commercial solutions and software strategy.
Experts say creating tools for developers to optimize their work for multicore processors is one of the biggest challenges the semiconductor industry faces.
Multimedia applications, in particular, get a relatively big performance boost by optimizing the code, Lewis said.
"Those particular applications are demanding and very complex, and a lot of the time, those routines are used repetitively," Lewis said.
AMD will continue contributing to the software and take outside contributions through the Framewave project, she added.
These days, most new PCs have dual-core central processors (CPU). That's one chip with two complete microprocessors on it, both sharing one path to memory and peripherals.
If you have a high-end gaming PC or a workstation, you might have one or two processor chips with four cores each. An eight-core PC is a very powerful machine--in real terms, up to eight times faster than the best desktop PCs you could get in 2004. For many years, PC performance doubled roughly every 18 months; multicore technology has produced annual doubling for three years now.
But that's not really so impressive when you look at the 15-year history of 3D graphics on PCs. The companies making graphics processors (GPUs) have delivered a doubling of performance every 6 months or so for this whole time. That means today's graphics chips are faster than 1992 products by a ratio of 2 to the 30th power, or about a billion to one.
CPU progress is slow by comparison. Over the last 30 years of the microprocessor, performance on integer operations has improved by about a million to one. Floating-point performance looks much better, since early CPUs had to emulate floating-point operations in software. With hardware floating-point units (FPUs), today's processors run FP-intensive code about 100 million times faster than those of 1977.
Floating-point performance is the key to the rapid progress of graphics chips, too. Most of the math required to display the special effects in a game like Company of Heroes is done using floating-point numbers. This math has to be done over and over for each pixel on the screen, but the work for many pixels can be done in parallel.
That's the loophole here, the trick that graphics chips exploit to boost performance so much with each new generation of 3D chips. It's all done in parallel.
So if you think eight processor cores is great, think about an Nvidia GeForce 8800 GPU with 128-thread processors running at 1.35GHz...or an AMD Radeon HD 2900 with 320-stream processors running at 743MHz. These processors are very simple by comparison with the cores in a CPU, but there sure are a lot of them. (There's no easy way to make direct comparisons between these numbers, so don't worry about it...just let the numbers flow over you.)
Then think about this: either one of those chips could, in principle, run a word-processing program all by itself. But such a program would probably run on only a few of those thread or stream processors, and inefficiently at that, so the program would probably run more slowly overall than it would on some old Windows 98 machine. For some things, CPUs are still much better than GPUs.
And this brings me back to the subject of yesterday's blog (here). Nobody's really sure how to evolve a CPU to the point that it could replace a GPU without losing what makes it a good CPU. Or vice-versa. CPUs and GPUs are likely to have distinct designs for a long time to come.
But they won't necessarily stay on separate chips. I'll explain why later this week.
Intel on Tuesday is scheduled to release the source code to a development tool for writing applications to run on multicore chips.
The company released Threading Building Blocks last August, a C++ template designed to simplify the job of writing applications that take advantage of processors with multiple cores, or processing units.
During the last year, Intel found that customers and potential customers wanted greater platform support and assurances that the toolset would be around for a long time, said James Reinders, the director of Intel's software development products.
To address these concerns, Intel has decided to release the tool under the General Public License version 2 with runtime exception. With the runtime exception, commercial software can choose to embed it in their own closed-source products, Reinders said.
Intel will continue to sell support for Threading Building Blocks for $299 a year.
Software vendors and chip manufacturers have for the last few years urged programmers to retool their applications for dual or multicore chips, which are becoming commonplace. But for the most part, there are a few applications that take full advantage of dual-core chips on desktop PCs, for example.
"There is definitely a lot of untapped potential staring us in the face," said Reinders. But "you can't really expect a lot of programmers to do something about parallelism if it is an extreme distraction from their job."
Threading Building Blocks works with Windows, Mac OS X, and popular Linux distributions on x86 chips. Alpha versions of the tool on Solaris 10, FreeBSD and MacOS on G5 chips are also available.
Ever since Intel showed off its 80-core prototype processor, people have asked "Why 80 cores?"
There's actually nothing magic about the number, according to Jerry Bautista, co-director of the Tera-scale Computing Research Program at Intel, and others. Intel wanted to make a chip that could perform 1 trillion floating-point operations per second, known as a teraflop. And 80 cores did the trick. The chip does not contain x86 processing cores, the kind of cores inside Intel's PC chips, but cores optimized for floating-point (or decimal) math.
Other sources at the company pointed out that 80 cores also allowed Intel to maximize the room inside the reticle, the mask employed to direct light from a lithography machine to a photo-resistant silicon wafer. Light shining through the reticle creates a pattern on the wafer and the pattern then serves as a blueprint for the circuits of a chip. More cores, and Intel would have needed a larger reticle.
Intel is now trying to take what it learned with the 80-core prototype and make commercially available chips. Read more here.
CORONADO, Calif.--Future versions of Windows will have to be "fundamentally different" in order to take advantage of multicore processors, according to Ty Carlson of Microsoft.
"You're going to see in excess of 8, 16, 64 and beyond processors on your client computer," said Carlson, director of technical strategy at Microsoft, during a panel discussion at the Future in Review conference. Windows Vista, on the other hand, is "designed to run on 1, 2, maybe 4 processors," he said, referring to the fact that quad-core processors are now available from Intel and are on the way from Advanced Micro Devices.
The problem, as has been noted on many occasions, is that loads of PC applications were programmed with serial processing in mind, meaning that the performance of those applications increased as a chip's clock speed increased. That's not how it works anymore. The chip industy has decided that multiple cores are the best way to keep increasing performance, and that means applications now have to be designed with parallel processing in mind.
Intel and AMD have not confirmed processor plans beyond eight cores, and only in theory at that. Intel has demonstrated an 80-core processor, but that's just a research project that can't run conventional code. But Carlson appears convinced that he and other software developers should start getting ready for that world.
"In 10 to 15 years' time we're going to have incredible computing power. The challenge will be bringing that ecosystem up that knows how to write programs," Carlson said. Windows Vista is designed to take advantage of multiple processing threads, but not 16 threads. And application developers are even further behind in making the transition to the multicore world.
- prev
- 1
- next
