aboutsummaryrefslogtreecommitdiffstats
path: root/tests/ref/fate/h264-reinit-small_420_9-to-small_420_8
diff options
context:
space:
mode:
authorTimothy Gu <timothygu99@gmail.com>2016-01-30 10:52:44 -0800
committerTimothy Gu <timothygu99@gmail.com>2016-02-08 16:25:44 -0800
commitcb8646af24bd8e9627cc5e1c62b049a00fe0b07b (patch)
tree7094dfe5af383a9c8435b1678b8f8851f6c3466b /tests/ref/fate/h264-reinit-small_420_9-to-small_420_8
parent8e46c7c1e7f5e775a152f6d75a08e93ddb9f7565 (diff)
downloadffmpeg-cb8646af24bd8e9627cc5e1c62b049a00fe0b07b.tar.gz
configure: Enable GCC vectorization on ≥4.9 on x86
4.9 was released precisely nine years after the first GCC version with autovectorizer (4.0) and six years after the first GCC version with `-ftree-vectorize` default to enabled on `-O3` (4.3). We've given GCC enough time to fix those bugs. FATE passes here on a x86-64 machine with both GCC 4.9.2 and 5.3.1. Some optimization hotspots benefit greatly from this change, especially those without handwritten assembly. For instance, the main function in vf_phase is now 1.6x faster (1.2x overall) on my machine.
Diffstat (limited to 'tests/ref/fate/h264-reinit-small_420_9-to-small_420_8')
0 files changed, 0 insertions, 0 deletions