sw_engine avx: addad ALPHA_BLEND and avxRasterTranslucentRect functions
authorMira Grudzinska <m.grudzinska@samsung.com>
Tue, 24 Aug 2021 23:19:25 +0000 (01:19 +0200)
committerJunsuChoi <jsuya.choi@samsung.com>
Mon, 30 Aug 2021 00:44:59 +0000 (09:44 +0900)
commitad6ce083abd9df56a570f4793e78802966548e19
tree2166d9664dac18cf9ab5fc234e89002b2dcc3c10
parent383772f61d9a918e366e801e4f0c7df8895ebb9a
sw_engine avx: addad ALPHA_BLEND and avxRasterTranslucentRect functions

Both functions implemented using 128-bit registers.
avxRasterTranslucentRect is around 5 times faster than cRasterTranslucentRect (i7-8700 CPU - Coffe Lake)
src/lib/sw_engine/tvgSwRaster.cpp
src/lib/sw_engine/tvgSwRasterAvx.h