]> granicus.if.org Git - libvpx/commit
projects / libvpx / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 8101e82) | patch
WORKAROUND FIX FOR GCC4.9.1
authorlevytamar82 <tamar.levy@intel.com>
Thu, 16 Oct 2014 23:56:37 +0000 (16:56 -0700)
committerlevytamar82 <tamar.levy@intel.com>
Sat, 1 Nov 2014 18:27:28 +0000 (11:27 -0700)
commit86175a57881ba5b891fadd2e42a02404facf44a1
treebdd28ea46c68122b704412e3954f9d8c74e0adc8tree | snapshot
parent8101e82a1bbb5dfeaf9014c90f310337592e402acommit | diff
WORKAROUND FIX FOR GCC4.9.1

In the function mb_lpf_horizontal_edge_w_avx2_16 the usage of the intrinsic
_mm256_cvtepu8_epi16 cause a compiler bug in gcc 4.9.1.
until it will be fixed I created a workaround that create the up convert by
using broadcast128+shuffle.
The bug was reported here:
https://code.google.com/p/webm/issues/detail?id=867

Change-Id: I73452e6806f42e0fadcde96b804ea3afa7eeb351
test/lpf_8_test.cc diff | blob | history
vp9/common/x86/vp9_loopfilter_intrin_avx2.c diff | blob | history
Unnamed repository; edit this file 'description' to name the repository.