Enable better vectorization for generic convolution

Browse Source

Break the single dependence chain into two parallel sub-chains.
Provides 2-4% performance uplift as measured on modern ARM systems
when using the generic codepath.

...

This commit is contained in:

Mahesh Madhav 2024-04-24 20:32:08 +00:00

parent 49ab34dfef

commit 1fd5cd8c7e

1 changed files with 692 additions and 641 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1333

src/libFLAC/lpc.c

View File

File diff suppressed because it is too large Load Diff