Add aom_subtract_block_avx2
Speed up about 0.4% without RD change
test sequence: BasketballDrill_832x480_50.y4m
test command line:./aomenc --cpu-used=1 --psnr -D \
-q --end-usage=vbr --target-bitrate=1200 --limit=20 \
BasketballDrill_832x480_50.y4m -otest.webm
Change-Id: I0f294fad578d6ea94a41a23765655e6ad4b82f46
diff --git a/aom_dsp/aom_dsp.cmake b/aom_dsp/aom_dsp.cmake
index 9208706..7c0111a 100644
--- a/aom_dsp/aom_dsp.cmake
+++ b/aom_dsp/aom_dsp.cmake
@@ -192,6 +192,7 @@
list(APPEND AOM_DSP_ENCODER_INTRIN_AVX2
"${AOM_ROOT}/aom_dsp/x86/masked_sad_intrin_avx2.c"
+ "${AOM_ROOT}/aom_dsp/x86/subtract_avx2.c"
"${AOM_ROOT}/aom_dsp/x86/highbd_quantize_intrin_avx2.c"
"${AOM_ROOT}/aom_dsp/x86/sad4d_avx2.c"
"${AOM_ROOT}/aom_dsp/x86/sad_avx2.c"