SuperPi on GPU, were going CUDA

[saaya]

And ATI has something like CUDA, called CTM (close to metal), to bad you have to program in assembler with CTM, in CUDA you can program in C & C++

i thought that too, but its not like that... you cant just run c and c++ code on cuda. its similar to c and c++ afaik, but you still need to basically program pretty much in machine langauge for cuda.

cuda and ctm are very similar afaik, both are way too complex and way NOT user friendly, which i think is the main reason why we dont see any gpgpu based apps so far.

the directx gpgpu approach seems to be very promising, a united api that drivers are already optimized to work with and that is already somewhat familiar to developers makes more sense than cuda or ctm imo. Plus its going to be ONE api, maintained by ONE party, which means coders dont have to decide for either ati or nvidia, and actually ANY gpu can be used for this, even sis IGP or whatnot