Through Dark Clouds of Confusion Come Glimmering Rays of Hope

Print Friendly, PDF & Email
Over the past year, barrels of ink have been used to describe all the things being doing wrong in the quest to reach exascale-levels of computation.
Disagreement on what direction to take, the evolution vs. revolution argument, undoubtedly has impacted exascale progress – but thankfully has not stifled it completely. Today, we finally have a number of reasons to actually feel encouraged about exascale progress as the dark clouds of confusion and skepticism give way to glimmering rays of technological hope.


The first area of promise, a foundational step toward achieving exascale, is being researched and discussed by both Intel and Nvidia, but our feature article on this topic comes from Intel. We interviewed Intel Fellow, Vivek De, to discuss Near Threshold Voltage processing, or NTV. NTV processing is a research area that holds tremendous promise for more efficient power management, and is applicable to numerous future computing applications ranging from mobile applications to HPC, and is likely to be one of the critical technologies required to enable power-efficient exascale systems. This power management approach, which has been demonstrated with functioning research prototypes, including a solar-powered processor, still has a long way to go. But, in our opinion, NTV is indeed a glimmering ray of hope.



The other ray of sunshine comes from Nvidia with some recent announcements made at their GPU Technology Conference.We were impressed with two innovative components of the Keplar GPU called Hyper-Q and Dynamic Parallelism. We tend to refer to these two as one, as they work hand-in-hand to improve the Keplar GPU’s performance by a factor of 3 to 4x per watt from the previous generation of Fermi GPUs.However, as we’ve lectured for so many months, and while we are encouraged by this progress, there is much more to achieving a practical exascale system than developing the technology. We spoke with NCSA’s Merle Giles – a very passionate discussion on finding an exascale business model. As Giles points out in this interview, the failure of Blue Waters was not in the technology, but in the business model. These conversations in themselves should be seen as a ray of hope. As a community, we’re not having these conversations at the level we should. But that seems to be changing – thanks to evangelists like Giles. If you are wondering what he means by an exascale business model, be sure to listen to the audio podcast interview.

And finally, we have the announcement of Intel acquiring the interconnect hardware assets of Cray.

Intel was the first company to put a stake in the ground stating we would have an exascale system by the end of the decade, and, we would be able to run today’s applications on those exascale systems without rewriting all the code. While a number of people believe this is not much more than marketing hype, a growing number of others are convinced Intel has been right all along. They haven’t said the applications will run well or set any sort of performance milestones – only that they would run. There is plenty of time to make sure this will happen. The important thing here is their commitment. Every discussion we’ve had – from the product groups to researchers deep within Intel Labs – has convinced us Intel is fully committed to leading the community to exascale. The acquisition of Cray’s interconnect hardware assets is a confirmation of Intel’s determination to forge the path to exascale, while providing the stepping stones to more practical, widespread use of petascale systems. In this issue, we look at the Intel – Cray deal from Cray’s perspective (an interview with Cray CEO, Pete Ungaro), along with an analyst perspective on the deal from John Barr at the 451 Group. In next month’s issue, we’ll be speaking with Intel’s Raj Hazra to get some more insight into this transaction that took many by surprise.

As the future of exascale starts to unfold, the dark clouds are slowly beginning to disperse.

For related stories, visit The Exascale Report Archives.


  1. JuneGuest says