About
I am a French-born, Canadian-living developper that want to create a Chess Proof-of-Concept using nVidia’s CUDA technology.
I began with Basic language & assembly code in the 70’s (8080, Z80, 6800, 6809, 68000, 80×86 …), switched to C on the 80’s, Java on the 90’s, and PHP on the millenium. CUDA is for me an opportunity to code close to the GPU architecture, to do massively-threaded software & libraries, and experiment new way to organize inter-thread communications, at the Block-level, GPU-level, computer-level and Cluster-level!
CUDA is challenging, with the ability to deliver incredibly high performances when correctly used and to be the worse architecture if you don’t plan to organize your code and algorithms to fit it’s needs.
Naturally I also consider OpenCL, albeit the idea of running the SAME CODE on different architectures (x86 CPU, nVidia’s CUDA GPU, ATI Radeon GPU, Cell), may lead to random performances.
