Feeds

Intel begs for multi-threaded software aid

Give us your poor, huddled test suites

Top 5 reasons to deploy VMware with Tegile

Hot Chips Long used to hawking GHz above all else, Intel has found the transition to a multi-core world with flat clock rates painful. The multi-core shift has, in fact, become so difficult that Intel is looking to partners for help designing and evaluating its future products.

Intel's CTO Justin Rattner issued the public cry for help yesterday at the Hot Chips conference being held on Stanford's campus. The executive would like to see companies, universities and other organizations come together to create a suite of standard software workloads that can address the performance of multi-core chips. Intel believes that such a new suite is required since today's most popular benchmarks don't really reflect the workloads of tomorrow.

"Nobody really knows what applications will look like in five to ten years," Rattner said. "But we can't leave it to chance to determine what those applications might be, so we decided to create our own suite of test applications and to work with partners."

Intel has a diverse set of in-house test applications today that it uses to measure future processors. The suite includes typical high performance computing, business and consumer software. It, however, also has some "futuristic" applications such as software for complex bioinformatics calculations, super high-end graphics code and fierce financial analysis algorithms.

It's this latter type of software that Intel believes will be most representative of code which will run on future multi-core chips. Such processors can handle multi-threaded software that is far more demanding than most of today's applications. Software companies – Microsoft included – will need to rewrite their applications en masse to take advantage of these new chips that use multiple low-power cores instead of just increasing the speed of a single core chip to boost overall performance.

Myriad unknowns surround the shift to multi-core chips, according to Rattner. Processor vendors will come out with a wide array of different looking chips – each capable of handling specific sets of software well. Meanwhile, software vendors will likely tap into this new type of horsepower in as of yet undetermined ways.

So, Intel designed its fancy test suite as its best effort at a software crystal ball.

The problem, however, is that Intel is not willing to turn over all of its proprietary test applications in the suite nor can it make public the test applications of unwilling third parties. So, the company has agreed to hand over a couple of items - for say body tracking and ray tracing – and then asked universities and others to contribute more applications to create a "public" test suite.

"There are problems with taking the suite public," Ratter said. "A lot of these codes are Intel codes and others belong to a number of our partners who consider these codes part of their intellectual property.

"So, we have gone around informally of late talking to people about their willingness to contribute to this public suite."

The likes of the University of Pittsburgh, UC Berkeley and Stanford have already agreed to ship some of their more demanding codes off for use in the public suite.

Princeton computer science professor Kai Li joined Rattner on stage at the Hot Chips conference and vowed to shepherd the test suite effort. Those willing to help with the project are urged to contact him.

It's hard to see a real downside to Intel's push for the public test suite. Such an application collection could help standardize the measurement of multi-core processor performance, especially if other vendors, er, chip in with the project.

As far as its own multi-core efforts go, Intel showed off a four-core (Clovertown) processor based system during Rattner's speech.

"From everything we can see, there are certainly interesting things to do with tens of cores and hundreds of threads," Rattner said. "That's where we are targeted with our research."

Rattner urged chip makers not to get caught in a "core war" where they make myriad multi-core chips just because they can. Software and tools vendors have a lot of catching up to do before such chips will be useful in broad terms, Ratter said.

Showing the depth of his commitment to Intel, Rattner gave the Hot Chips lecture even though he was on holiday celebrating a wedding anniversary. His wife left before the speech started to go shopping at the nearby, pricey Stanford Mall.

"My wife is probably going to do some damage with my credit card," he said.

Sweet, sweet revenge.®

Intelligent flash storage arrays

More from The Register

next story
Fujitsu CTO: We'll be 3D-printing tech execs in 15 years
Fleshy techie disses network neutrality, helmet-less motorcyclists
Space Commanders rebel as Elite:Dangerous kills offline mode
Frontier cops an epic kicking in its own forums ahead of December revival
Intel's LAME DUCK mobile chips gobbled by CASH COW
Chipzilla won't have money-losing mobe unit to kick about anymore
First in line to order a Nexus 6? AT&T has a BRICK for you
Black Screen of Death plagues early Google-mobe batch
Ford's B-Max: Fiesta-based runaround that goes THUNK
... when you close the slidey doors, that is ...
prev story

Whitepapers

Choosing cloud Backup services
Demystify how you can address your data protection needs in your small- to medium-sized business and select the best online backup service to meet your needs.
Forging a new future with identity relationship management
Learn about ForgeRock's next generation IRM platform and how it is designed to empower CEOS's and enterprises to engage with consumers.
Designing and building an open ITOA architecture
Learn about a new IT data taxonomy defined by the four data sources of IT visibility: wire, machine, agent, and synthetic data sets.
10 threats to successful enterprise endpoint backup
10 threats to a successful backup including issues with BYOD, slow backups and ineffective security.
High Performance for All
While HPC is not new, it has traditionally been seen as a specialist area – is it now geared up to meet more mainstream requirements?