1
edit
Changes
→Assignment 3
=== Assignment 3 ===
Now I realized I did not use shared memory for assignment 2. So I changed that. In my assignment 3 I am using shared memory. As for assignment 2 I am doing 1000 times image negate. Last lectures for this course helped a lot to learn to optimize. Specially coalesced access. It is amazing! In assignment 2 my negate function calculations became faster almost twice just because I was using the kernel. I was thinking there is nothing to be done more. And now my negate function calculations became faster more than twice than in assignment 2 thanks to the optimization. This is my profile for assignment 3:
[[Image:Profiler_scr.png]]
