Me. I've been doing some CUDA C++ through Visual Studio 2012 and the results return so much faster than the same code in C++ only running via OpenMP (to get the 3770K fully threaded). I'm working on a new code hoping it'll save me a few months worth of numerical optimization if I can get it all automated for me. Then I get the advsior to buy me Titan.
6GB has also proven to be beneficial, as I can store arrays 2x or 3x larger than I otherwise could on a 3GB or 2GB VRAM card.
Edited by Creator - 10/25/13 at 7:21pm