diff options
| author | Tobias Grosser <grosser@google.com> | 2013-07-11 17:58:47 -0700 |
|---|---|---|
| committer | Tobias Grosser <grosser@google.com> | 2013-07-15 14:07:20 -0700 |
| commit | da7ddd8477dc802c8736c7ab860fc09f33689ce9 (patch) | |
| tree | aa2adccfb8659aeef55ae83ef2f1164699c4cb30 /cpp/Script.cpp | |
| parent | 574854bcb2eb25a85b9b52faf2fb3e743fa7aa14 (diff) | |
Simplify code of convolve3x3
Instead of first doing all multiplications and then adding the results in
a tree manner, we just repetitively perform a load/multiply/add patter.
With and without tuning for A15, this yields a 5% performance increase for N10.
This commit also exposes more instructions to be transformed into fused
multiply adds.
Change-Id: I1215d75da236e6b2d6b6aa48b3ab35606cdba7b8
Diffstat (limited to 'cpp/Script.cpp')
0 files changed, 0 insertions, 0 deletions