Thoughts about CUDA
I just figured out that my Geforce GT210 has about 2.3 GFLOPS. Well, it isn't much I guess compared to now's standards, like the RADEON's really fast Graphic card that reaches over 3 TFLOPS, but I am thinking about how I can apply 2.3GFLOPS of processing power to compute a ranking system for my search engine.
According to FLOPS article on Wikipedia, Nintendo 3DS reaches 1.73GFLOPS, and I may be wrong about the GT210, (I will check again later), but still, 2.3 GFLOPS is probably more than enough to calculate a decent ranking system.
I am thinking of using either OPEN CL, OPEN GL (?), or CUDA for this. I think CUDA will be the best choice for now.
I need to start studying matrix system more, I think, although I got a bite of it at my high school.
According to FLOPS article on Wikipedia, Nintendo 3DS reaches 1.73GFLOPS, and I may be wrong about the GT210, (I will check again later), but still, 2.3 GFLOPS is probably more than enough to calculate a decent ranking system.
I am thinking of using either OPEN CL, OPEN GL (?), or CUDA for this. I think CUDA will be the best choice for now.
I need to start studying matrix system more, I think, although I got a bite of it at my high school.
Comments