Home > News > How Intel Xeon Phi Got its Codenames

How Intel Xeon Phi Got its Codenames

Intel Knights Corner Die

One of questions often asked by our readers is how projects get their codenames. In our discussions with Intel representatives during the event held at Texas Advanced Computing Center, we learned how "Larrabee-reborn" got its name.

First off, MIC (Many Integrated Core) project represents unification of several projects inside Intel's R&D group. The teams that worked on the project came from numerous different projects with TeraScale (remember the Polaris die with 80 "Atom" look-a-like cores glued together?), Larrabee and Cloud Computer. The slide below shows you a path to the introduction of Xeon Phi.

Just like any other project, choosing a codename for three parts which complete the circle of Xeon Phi development wasn't easy. In our conversations with the Intel executives and engineers, we learned that giving codenames is probably the toughest part of the project, since it will never be used as an actual product – but it still needs to be clear of any copyright issues which might appear.

The name for the three successors of Auburn Isle (codename for Larrabee silicon) are as follows:

  • Knights Ferry (1st silicon)
  • Knights Corner (pre-production silicon, used in TACC and other installations)
  • Knights Landing (shipping silicon, available 2013)

The codenames come from Folsom lake in Northern California, where Intel has quite sizeable research facility. In fact, if you just happen to find yourself in Folsom, you can take a Ferry, and after a Corner there will be Landing. Knights Ferry and Knights Landing are also locations in California, while Corner was added as a bridge between the two.

That in a nutshell, is a story of how strengths of a failed project (Larrabee / Auburn Isle) got combined with different internal well-thought projects and became an actual product line – Xeon Phi. Time will tell just how well the Knights Landing will… land. A lot more details should become available during the upcoming SC 12 supercomputing conference which will be held from November 10-16 in Salt Lake City, UT.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Read previous post:
NVIDIA’s 18K+ Tesla K20 GPUs claims ‘world’s fastest open-science supercomputer’

NVIDA has just announced the completion of ‘the world’s fastest open-science supercomputer’ at the Oak Ridge National Laboratory in Tennessee...