[hyperbovine]

I still don't understand why certain performance aspects of the CUDA platform are so poorly documented. Why is successfully pushing the hw to its performance envelope considered a novel research result? Shouldn't I be able to look this stuff up on the Nvidia website?

[amindiro]

One reason is clearly the fast past at which nvidia is evolving the hardware. I would consider cuda a very well documented platform in general. What they lack is low level tutorials, but this is where posts like this one can be a good resource