This forum has been archived. All content is frozen. Please use KDE Discuss instead.

SIMD vectorization for dynamic vs fixed size matricies

Tags: None
(comma "," separated)
richardson
Registered Member
Posts
1
Karma
0
From http://eigen.tuxfamily.org/dox/group__T ... Class.html

Finally, depending on circumstances, Eigen can also be more aggressive trying to vectorize (use SIMD instructions) when dynamic sizes are used, see Vectorization.”


The link to “Vectorization” said “write this dox page”.

I was wondering, why is Eigen more aggressive with SIMD with dynamic sizes than fixed sizes?
User avatar
ggael
Moderator
Posts
3447
Karma
19
OS
For example, with SSE and 5x1 fixed-size vectors, then Eigen won't attempt to perform any vectorization, whereas for dynamic-size vectors, since Eigen cannot know that the vector will be that small, then the same logic will be applied as for any dynamic vectors.


Bookmarks



Who is online

Registered users: bartoloni, Bing [Bot], Evergrowing, Google [Bot], ourcraft