It's that time of year again... the current year is almost at an end... a new year full of joy and wonder just around the corner. I can't help but spend a few minutes reflecting upon the past and wondering about the future. A few years ago, Pat Gelsinger from Intel stated that NVIDIA's CUDA technology was destined to be a "footnote" in the history of computing. When he made this prediction Nvidia's CUDA initiative was barely off the ground and Intel, the 800 pound gorilla, was going to squash CUDA with their soon to be released Larrabee device.

As it turned out Larrabee was, shall we say, a bit of a disappointment. Instead of supplanting Nvidia GPUs and CUDA as the acceleration technology of the future, Intel's unimpressive Larrabee chip ended up being "rebranded" as a parallel computing tool research platform. Perhaps this explains why Pat Gelsinger no longer works at Intel. After starting out as a glimmer in Ian Bucks eye... Nvidia's CUDA has penetrated the Oil and Gas industry, mechanical engineering industry, image processing industry, supercomputing industry, and the financial industry (just to name a few). While some people "report" on these accomplishments I am actually a participant.

As a System Architect at one of the best performing hedge funds in the world, I have been one of the beneficiaries of Nvidia's vision. I followed their progress through their single precision days and tried to convince our quants that single precision was "good enough". But Nvidia didn't leave me hanging... they quickly put forth a road map to double precision. While they have never seemed to meet their deadlines they always delivered in the end which is more than can be said for Intel or AMD (I'm still waiting Mr. Gelsinger... oh... I forgot... you threw in the towel and bailed to EMC).

Many will state that you can never achieve the 100 fold increases that are promised with CUDA so you should just stay the "CPU" course. I'm here to tell you that in the world of finance you don't need a 100 fold increase to justify this paradigm shift. A measly 26 fold increase can save you millions of dollars and, in the world of finance, make you millions more to boot. I would just like to thank Nvidia for providing a platform for those of us brave enough to "think different" to achieve the impossible.

Happy Holidays,

Dr. Zaius