Can Nvidia Outsmart Intel in HPC?

Print Friendly, PDF & Email

harveyWhile we track the moves of HPC vendors on a daily basis here at insideHPC, it is always interesting to get the perspective from Wall Street. Over at the Motley Fool, Srdjan Bejakovic writes that Nvidia and Intel are battling it out in the HPC space.

Intel and NVIDIA are fighting it out in the high-performance computing processor market, with Intel appearing to be the aggressor and gaining share. However, NVIDIA’s accelerators still lead in terms of performance, and NVIDIA’s support for ARM-based chips might even undermine Intel’s position in HPC. While this wouldn’t be a huge blow in terms of revenue, Intel should worry that ARM-based chips are taking even more of its business away.

Before we go any farther, let me (Rich Brueckner) say that I think this is lousy advice. Investors should remember that thing radio announcer Paul Harvey used to say: “And now, here’s the rest of the story.”

Today, every accelerated supercomputer that ships results in revenue for Intel. Every. Single. One. You need x86 processors to run the thing and AMD has dropped off the planet in HPC since Titan shipped a few years ago.

Despite what you may hear elsewhere, a win for Nvidia in this space is also a win for Intel. They are partners. Intel sells thousands of Xeons to customers who choose to accelerate with Nvidia GPUs. And the combination works really well for a lot of customers, thank you very much. Sure, ARM and OpenPOWER may be viable alternatives someday for accelerated systems, but not today.

What I’m saying here is that the real winner to put your money on is accelerated computing. Sure, vendors will duke it out, but the lion’s share of the TOP500 uses x86 and that is simply not going to change any time soon.

So before you go betting on your horses, check out the jockey first. Chances are, he’s going to start looking very familiar.