Open
Description
https://en.algorithmica.org/hpc/cpu-cache/associativity/
The section does not specify that this slowdown only occurs when iterating the array multiple times, leading to confusion, and perhaps believing that a single iteration of an array also suffers of some unspecific cache thrashing slowdown? In my cpu it seems to only exhibit a difference between the two loops when I do a few full iterations of each, not if I allocate a fresh array for each loop.
I suggest changing the loops to the following, in order to prevent confusion.
for (int j = 0; j < K; j++)
for (int i = 0; i < N; i += {256,257})
a[i]++;
Thank you for the truly great resource that is this book!
Metadata
Metadata
Assignees
Labels
No labels