Michael Abrash’s First Look at the Larrabee New Instructions (LRBni)

Dr. Dobb’s just published the first of a new series of detailed articles related to Larrabee hardware and software architecture written by Michael Abrash.

A First Look at the Larrabee New Instructions (LRBni)

Larrabee is an architecture, rather than a product, with three distinct aspects — many cores, many threads, and a new vector instruction set — that boost performance. This architecture will first be used in GPUs, and could be used in CPUs as well.

One Response to “Michael Abrash’s First Look at the Larrabee New Instructions (LRBni)”

  1. purpledog Says:

    “the vectorization process can be made automatic and compatible with existing code.”
    Out of context, that’s… Well… Pretty funny. But being honest, the next sentence is:
    “Of course, nothing’s ever that easy…”

    The gather/scatter instructions looks like the end of Cuda and its over-convoluted memory model. Amen.

    Also, it seems that nobody made the link between proto and Larrabee yet … Still, the mixture looks *really* explosive to me.
    http://www.boost.org/doc/libs/1_38_0/doc/html/proto.html


Leave a comment