tree d56858be755ee38d41a278793afd5be869094695
parent 98dc22b814950167a9c7b6f21b32db9983c8e1da
author Rupert Swarbrick <rupert.swarbrick@argondesign.com> 1507106566 +0100
committer Debargha Mukherjee <debargha@google.com> 1507594053 +0000

Add an SSE4.1 implementation of av1_highbd_convolve_2d_scale

For large blocks this is about 8x the speed of the C version. The code
needs SSE 4.1 for the PMULLD instruction that we use to do SIMD 32-bit
multiplies.

The patch uses av1_convolve_scale_test (written already to test the
low bit depth path) to make sure the optimised code matches the C
version.

Change-Id: I9304d6bb3d2cb31390de93ed08ff1a852e3ace86