NVIDIA CUDA Tutorial 9: Bank Conflicts

This tute we'll look at bank conflicts. Bank conflicts slow shared memory down, they occur when multiple values are requested from a shared memory bank are requested from a single warp. I've also introduced the cudaDaviceReset function which is used to write all the performance counters from the GPU for use with a profiler. This function should be at the end of any programs you wish to profile. The clock() function is used to record how long your code takes. It's extremely accurate and fine grained and reports in individual clock cycles. Sorry in advance about the flickering that happens on some of the slides. Also, I started a music channel the other day. Have a look if you like music, it's going to have theory and prac tutes as well as (probably) me and mine playing music and talking about it, the youtube for it is: https://www.youtube.com/channel/UC2nWRGqPL6O-Lm8mexWhISQ