Over 90% of the execution time is now spent in code outside the convolution kernel. Considering Amdahl's law I think we are done. 2/6
Me to customer: The code now runs at ~2k iter per seconds, a 150x speedup compared to the version without CUDA. 1/6
-
-
-
Costumer reply: We ran your code in nvvp. It says your code is only 40% efficient. Come back when you are at ~4k iter/sec. 3/6
- Show replies
New conversation -
-
-
Recently at work, I've been trying to explain to managers and marketing that my feature isn't as fast as they think it is.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.