News for hackers. Stuff that matters.
Re:Benchmarks for Blaze, A high-performance C++ math library by fizixer on Tuesday August 25, 2015 @02:03PM (#10118392)
In the first plot, why do all libraries slow down at the n=1000 mark? something to do with cache?
Re:Benchmarks for Blaze, A high-performance C++ math library by acconsta on Tuesday August 25, 2015 @02:23PM (#10118537)
Yep. A 32KB L1 cache can hold at most 4000 doubles. 4000/(2 input vectors + 1 result) =1333.

http://danluu.com/3c-conflict/

Re:Benchmarks for Blaze, A high-performance C++ math library by wall_words on Tuesday August 25, 2015 @02:12PM (#10118463)
I'm guessing that's the point at which the working set exceeds the L1 cache size. You can see a few more subtle dips in the performance graph at later points; these correspond to working set spilling out of the L2 and L3 caches.