Me to customer: The code now runs at ~2k iter per seconds, a 150x speedup compared to the version without CUDA. 1/6
-
All attempts at explaining to them the error in their thinking (or the exact meaning of that "efficiency" number) failed. 4/6
This was over a year ago. I just looked at old code from that project and was reminded of how it ended. 5/6
-
-
I think they honestly believed I tried to con them or something. Do you have any Dunning–Kruger customer story that you can share? 6/6
-
Yeah, all my customers that think that their code will be 2880 times faster because their card has 2880 "cores".
End of conversation
New conversation -
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.