]> granicus.if.org Git - libvpx/commit
vp8 bilinear: rewrite in intrinsics
authorJohann <johann.koenig@duck.com>
Wed, 24 Oct 2018 19:22:35 +0000 (12:22 -0700)
committerJohann <johann.koenig@duck.com>
Wed, 24 Oct 2018 23:01:39 +0000 (16:01 -0700)
commit6f35ac956b8315b466c7f66af1172fbc473f1464
tree73de6089d910ab9ad7cae8a28b94be11b3d5ac4b
parent979f0c0e5a901e9d6d21bcc5d566024acaaa22f9
vp8 bilinear: rewrite in intrinsics

8x8 is 15% faster than the assembly. 8x4 is 200% faster than MMX.

Remove MMX version.

Change-Id: I55642ebd276db265911f2c79616177a3a9a7e04f
test/predict_test.cc
vp8/common/rtcd_defs.pl
vp8/common/x86/bilinear_filter_sse2.c [new file with mode: 0644]
vp8/common/x86/subpixel_mmx.asm
vp8/common/x86/subpixel_sse2.asm
vp8/vp8_common.mk