shithub: libvpx

Download patch

ref: 1b833d63d9c82270e4ea588541d14e9111c64c79
parent: fe96dbda156bf5cbaeec9535180e2a02c2b93814
author: Johann <[email protected]>
date: Wed Jun 29 14:15:49 EDT 2016 

vpx_dsp: remove x86inc.asm distinction

BUG=b:29583530

Change-Id: I397d77536b0d3cee0a92cdfe8b76bc4e434d0720


--- a/vpx_dsp/vpx_dsp.mk
+++ b/vpx_dsp/vpx_dsp.mk
@@ -40,18 +40,14 @@
 # intra predictions
 DSP_SRCS-yes += intrapred.c
 
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSE) += x86/intrapred_sse2.asm
 DSP_SRCS-$(HAVE_SSE2) += x86/intrapred_sse2.asm
 DSP_SRCS-$(HAVE_SSSE3) += x86/intrapred_ssse3.asm
 DSP_SRCS-$(HAVE_SSSE3) += x86/vpx_subpixel_8t_ssse3.asm
-endif  # CONFIG_USE_X86INC
 
 ifeq ($(CONFIG_VP9_HIGHBITDEPTH),yes)
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSE)  += x86/highbd_intrapred_sse2.asm
 DSP_SRCS-$(HAVE_SSE2) += x86/highbd_intrapred_sse2.asm
-endif  # CONFIG_USE_X86INC
 endif  # CONFIG_VP9_HIGHBITDEPTH
 
 ifneq ($(filter yes,$(CONFIG_POSTPROC) $(CONFIG_VP9_POSTPROC)),)
@@ -87,9 +83,8 @@
 DSP_SRCS-$(HAVE_SSE2)  += x86/vpx_high_subpixel_8t_sse2.asm
 DSP_SRCS-$(HAVE_SSE2)  += x86/vpx_high_subpixel_bilinear_sse2.asm
 endif
-ifeq ($(CONFIG_USE_X86INC),yes)
+
 DSP_SRCS-$(HAVE_SSE2)  += x86/vpx_convolve_copy_sse2.asm
-endif
 
 ifeq ($(HAVE_NEON_ASM),yes)
 DSP_SRCS-yes += arm/vpx_convolve_copy_neon_asm$(ASM)
@@ -179,10 +174,8 @@
 DSP_SRCS-$(HAVE_SSE2)   += x86/fwd_txfm_impl_sse2.h
 DSP_SRCS-$(HAVE_SSE2)   += x86/fwd_dct32x32_impl_sse2.h
 ifeq ($(ARCH_X86_64),yes)
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSSE3)  += x86/fwd_txfm_ssse3_x86_64.asm
 endif
-endif
 DSP_SRCS-$(HAVE_AVX2)   += x86/fwd_txfm_avx2.c
 DSP_SRCS-$(HAVE_AVX2)   += x86/fwd_dct32x32_impl_avx2.h
 DSP_SRCS-$(HAVE_NEON)   += arm/fwd_txfm_neon.c
@@ -197,12 +190,10 @@
 DSP_SRCS-yes            += inv_txfm.c
 DSP_SRCS-$(HAVE_SSE2)   += x86/inv_txfm_sse2.h
 DSP_SRCS-$(HAVE_SSE2)   += x86/inv_txfm_sse2.c
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSE2)   += x86/inv_wht_sse2.asm
 ifeq ($(ARCH_X86_64),yes)
 DSP_SRCS-$(HAVE_SSSE3)  += x86/inv_txfm_ssse3_x86_64.asm
 endif  # ARCH_X86_64
-endif  # CONFIG_USE_X86INC
 
 ifeq ($(HAVE_NEON_ASM),yes)
 DSP_SRCS-yes  += arm/save_reg_neon$(ASM)
@@ -254,11 +245,9 @@
 DSP_SRCS-$(HAVE_SSE2)   += x86/highbd_quantize_intrin_sse2.c
 endif
 ifeq ($(ARCH_X86_64),yes)
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSSE3)  += x86/quantize_ssse3_x86_64.asm
 DSP_SRCS-$(HAVE_AVX)    += x86/quantize_avx_x86_64.asm
 endif
-endif
 
 # avg
 DSP_SRCS-yes           += avg.c
@@ -267,10 +256,8 @@
 DSP_SRCS-$(HAVE_MSA)   += mips/avg_msa.c
 DSP_SRCS-$(HAVE_NEON)  += arm/hadamard_neon.c
 ifeq ($(ARCH_X86_64),yes)
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSSE3) += x86/avg_ssse3_x86_64.asm
 endif
-endif
 
 endif  # CONFIG_VP9_ENCODER
 
@@ -292,7 +279,6 @@
 DSP_SRCS-$(HAVE_AVX2)   += x86/sad4d_avx2.c
 DSP_SRCS-$(HAVE_AVX2)   += x86/sad_avx2.c
 
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSE)    += x86/sad4d_sse2.asm
 DSP_SRCS-$(HAVE_SSE)    += x86/sad_sse2.asm
 DSP_SRCS-$(HAVE_SSE2)   += x86/sad4d_sse2.asm
@@ -303,7 +289,6 @@
 DSP_SRCS-$(HAVE_SSE2) += x86/highbd_sad4d_sse2.asm
 DSP_SRCS-$(HAVE_SSE2) += x86/highbd_sad_sse2.asm
 endif  # CONFIG_VP9_HIGHBITDEPTH
-endif  # CONFIG_USE_X86INC
 
 endif  # CONFIG_ENCODERS
 
@@ -334,17 +319,13 @@
 DSP_SRCS-$(HAVE_SSE2)   += x86/ssim_opt_x86_64.asm
 endif  # ARCH_X86_64
 
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSE)    += x86/subpel_variance_sse2.asm
 DSP_SRCS-$(HAVE_SSE2)   += x86/subpel_variance_sse2.asm  # Contains SSE2 and SSSE3
-endif  # CONFIG_USE_X86INC
 
 ifeq ($(CONFIG_VP9_HIGHBITDEPTH),yes)
 DSP_SRCS-$(HAVE_SSE2)   += x86/highbd_variance_sse2.c
 DSP_SRCS-$(HAVE_SSE2)   += x86/highbd_variance_impl_sse2.asm
-ifeq ($(CONFIG_USE_X86INC),yes)
 DSP_SRCS-$(HAVE_SSE2)   += x86/highbd_subpel_variance_impl_sse2.asm
-endif  # CONFIG_USE_X86INC
 endif  # CONFIG_VP9_HIGHBITDEPTH
 endif  # CONFIG_ENCODERS || CONFIG_POSTPROC || CONFIG_VP9_POSTPROC
 
--- a/vpx_dsp/vpx_dsp_rtcd_defs.pl
+++ b/vpx_dsp/vpx_dsp_rtcd_defs.pl
@@ -11,29 +11,6 @@
 }
 forward_decls qw/vpx_dsp_forward_decls/;
 
-# x86inc.asm had specific constraints. break it out so it's easy to disable.
-# zero all the variables to avoid tricky else conditions.
-$mmx_x86inc = $sse_x86inc = $sse2_x86inc = $ssse3_x86inc = $avx_x86inc =
-  $avx2_x86inc = '';
-$mmx_x86_64_x86inc = $sse_x86_64_x86inc = $sse2_x86_64_x86inc =
-  $ssse3_x86_64_x86inc = $avx_x86_64_x86inc = $avx2_x86_64_x86inc = '';
-if (vpx_config("CONFIG_USE_X86INC") eq "yes") {
-  $mmx_x86inc = 'mmx';
-  $sse_x86inc = 'sse';
-  $sse2_x86inc = 'sse2';
-  $ssse3_x86inc = 'ssse3';
-  $avx_x86inc = 'avx';
-  $avx2_x86inc = 'avx2';
-  if ($opts{arch} eq "x86_64") {
-    $mmx_x86_64_x86inc = 'mmx';
-    $sse_x86_64_x86inc = 'sse';
-    $sse2_x86_64_x86inc = 'sse2';
-    $ssse3_x86_64_x86inc = 'ssse3';
-    $avx_x86_64_x86inc = 'avx';
-    $avx2_x86_64_x86inc = 'avx2';
-  }
-}
-
 # optimizations which depend on multiple features
 $avx2_ssse3 = '';
 if ((vpx_config("HAVE_AVX2") eq "yes") && (vpx_config("HAVE_SSSE3") eq "yes")) {
@@ -55,19 +32,19 @@
 #
 
 add_proto qw/void vpx_d207_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d207_predictor_4x4/, "$sse2_x86inc";
+specialize qw/vpx_d207_predictor_4x4 sse2/;
 
 add_proto qw/void vpx_d207e_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d207e_predictor_4x4/;
 
 add_proto qw/void vpx_d45_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d45_predictor_4x4 neon/, "$sse2_x86inc";
+specialize qw/vpx_d45_predictor_4x4 neon sse2/;
 
 add_proto qw/void vpx_d45e_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d45e_predictor_4x4/;
 
 add_proto qw/void vpx_d63_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d63_predictor_4x4/, "$ssse3_x86inc";
+specialize qw/vpx_d63_predictor_4x4 ssse3/;
 
 add_proto qw/void vpx_d63e_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d63e_predictor_4x4/;
@@ -76,7 +53,7 @@
 specialize qw/vpx_d63f_predictor_4x4/;
 
 add_proto qw/void vpx_h_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_h_predictor_4x4 neon dspr2 msa/, "$sse2_x86inc";
+specialize qw/vpx_h_predictor_4x4 neon dspr2 msa sse2/;
 
 add_proto qw/void vpx_he_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_he_predictor_4x4/;
@@ -88,49 +65,49 @@
 specialize qw/vpx_d135_predictor_4x4 neon/;
 
 add_proto qw/void vpx_d153_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d153_predictor_4x4/, "$ssse3_x86inc";
+specialize qw/vpx_d153_predictor_4x4 ssse3/;
 
 add_proto qw/void vpx_v_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_v_predictor_4x4 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_v_predictor_4x4 neon msa sse2/;
 
 add_proto qw/void vpx_ve_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_ve_predictor_4x4/;
 
 add_proto qw/void vpx_tm_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_tm_predictor_4x4 neon dspr2 msa/, "$sse2_x86inc";
+specialize qw/vpx_tm_predictor_4x4 neon dspr2 msa sse2/;
 
 add_proto qw/void vpx_dc_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_predictor_4x4 dspr2 msa neon/, "$sse2_x86inc";
+specialize qw/vpx_dc_predictor_4x4 dspr2 msa neon sse2/;
 
 add_proto qw/void vpx_dc_top_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_top_predictor_4x4 msa neon/, "$sse2_x86inc";
+specialize qw/vpx_dc_top_predictor_4x4 msa neon sse2/;
 
 add_proto qw/void vpx_dc_left_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_left_predictor_4x4 msa neon/, "$sse2_x86inc";
+specialize qw/vpx_dc_left_predictor_4x4 msa neon sse2/;
 
 add_proto qw/void vpx_dc_128_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_128_predictor_4x4 msa neon/, "$sse2_x86inc";
+specialize qw/vpx_dc_128_predictor_4x4 msa neon sse2/;
 
 add_proto qw/void vpx_d207_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d207_predictor_8x8/, "$ssse3_x86inc";
+specialize qw/vpx_d207_predictor_8x8 ssse3/;
 
 add_proto qw/void vpx_d207e_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d207e_predictor_8x8/;
 
 add_proto qw/void vpx_d45_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d45_predictor_8x8 neon/, "$sse2_x86inc";
+specialize qw/vpx_d45_predictor_8x8 neon sse2/;
 
 add_proto qw/void vpx_d45e_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d45e_predictor_8x8/;
 
 add_proto qw/void vpx_d63_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d63_predictor_8x8/, "$ssse3_x86inc";
+specialize qw/vpx_d63_predictor_8x8 ssse3/;
 
 add_proto qw/void vpx_d63e_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d63e_predictor_8x8/;
 
 add_proto qw/void vpx_h_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_h_predictor_8x8 neon dspr2 msa/, "$sse2_x86inc";
+specialize qw/vpx_h_predictor_8x8 neon dspr2 msa sse2/;
 
 add_proto qw/void vpx_d117_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d117_predictor_8x8/;
@@ -139,46 +116,46 @@
 specialize qw/vpx_d135_predictor_8x8/;
 
 add_proto qw/void vpx_d153_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d153_predictor_8x8/, "$ssse3_x86inc";
+specialize qw/vpx_d153_predictor_8x8 ssse3/;
 
 add_proto qw/void vpx_v_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_v_predictor_8x8 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_v_predictor_8x8 neon msa sse2/;
 
 add_proto qw/void vpx_tm_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_tm_predictor_8x8 neon dspr2 msa/, "$sse2_x86inc";
+specialize qw/vpx_tm_predictor_8x8 neon dspr2 msa sse2/;
 
 add_proto qw/void vpx_dc_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_predictor_8x8 dspr2 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_dc_predictor_8x8 dspr2 neon msa sse2/;
 
 add_proto qw/void vpx_dc_top_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_top_predictor_8x8 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_dc_top_predictor_8x8 neon msa sse2/;
 
 add_proto qw/void vpx_dc_left_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_left_predictor_8x8 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_dc_left_predictor_8x8 neon msa sse2/;
 
 add_proto qw/void vpx_dc_128_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_128_predictor_8x8 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_dc_128_predictor_8x8 neon msa sse2/;
 
 add_proto qw/void vpx_d207_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d207_predictor_16x16/, "$ssse3_x86inc";
+specialize qw/vpx_d207_predictor_16x16 ssse3/;
 
 add_proto qw/void vpx_d207e_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d207e_predictor_16x16/;
 
 add_proto qw/void vpx_d45_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d45_predictor_16x16 neon/, "$ssse3_x86inc";
+specialize qw/vpx_d45_predictor_16x16 neon ssse3/;
 
 add_proto qw/void vpx_d45e_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d45e_predictor_16x16/;
 
 add_proto qw/void vpx_d63_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d63_predictor_16x16/, "$ssse3_x86inc";
+specialize qw/vpx_d63_predictor_16x16 ssse3/;
 
 add_proto qw/void vpx_d63e_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d63e_predictor_16x16/;
 
 add_proto qw/void vpx_h_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_h_predictor_16x16 neon dspr2 msa/, "$sse2_x86inc";
+specialize qw/vpx_h_predictor_16x16 neon dspr2 msa sse2/;
 
 add_proto qw/void vpx_d117_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d117_predictor_16x16/;
@@ -187,46 +164,46 @@
 specialize qw/vpx_d135_predictor_16x16/;
 
 add_proto qw/void vpx_d153_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d153_predictor_16x16/, "$ssse3_x86inc";
+specialize qw/vpx_d153_predictor_16x16 ssse3/;
 
 add_proto qw/void vpx_v_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_v_predictor_16x16 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_v_predictor_16x16 neon msa sse2/;
 
 add_proto qw/void vpx_tm_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_tm_predictor_16x16 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_tm_predictor_16x16 neon msa sse2/;
 
 add_proto qw/void vpx_dc_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_predictor_16x16 dspr2 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_dc_predictor_16x16 dspr2 neon msa sse2/;
 
 add_proto qw/void vpx_dc_top_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_top_predictor_16x16 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_dc_top_predictor_16x16 neon msa sse2/;
 
 add_proto qw/void vpx_dc_left_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_left_predictor_16x16 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_dc_left_predictor_16x16 neon msa sse2/;
 
 add_proto qw/void vpx_dc_128_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_128_predictor_16x16 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_dc_128_predictor_16x16 neon msa sse2/;
 
 add_proto qw/void vpx_d207_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d207_predictor_32x32/, "$ssse3_x86inc";
+specialize qw/vpx_d207_predictor_32x32 ssse3/;
 
 add_proto qw/void vpx_d207e_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d207e_predictor_32x32/;
 
 add_proto qw/void vpx_d45_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d45_predictor_32x32/, "$ssse3_x86inc";
+specialize qw/vpx_d45_predictor_32x32 ssse3/;
 
 add_proto qw/void vpx_d45e_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d45e_predictor_32x32/;
 
 add_proto qw/void vpx_d63_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d63_predictor_32x32/, "$ssse3_x86inc";
+specialize qw/vpx_d63_predictor_32x32 ssse3/;
 
 add_proto qw/void vpx_d63e_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d63e_predictor_32x32/;
 
 add_proto qw/void vpx_h_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_h_predictor_32x32 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_h_predictor_32x32 neon msa sse2/;
 
 add_proto qw/void vpx_d117_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vpx_d117_predictor_32x32/;
@@ -235,25 +212,25 @@
 specialize qw/vpx_d135_predictor_32x32/;
 
 add_proto qw/void vpx_d153_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_d153_predictor_32x32/, "$ssse3_x86inc";
+specialize qw/vpx_d153_predictor_32x32 ssse3/;
 
 add_proto qw/void vpx_v_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_v_predictor_32x32 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_v_predictor_32x32 neon msa sse2/;
 
 add_proto qw/void vpx_tm_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_tm_predictor_32x32 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_tm_predictor_32x32 neon msa sse2/;
 
 add_proto qw/void vpx_dc_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_predictor_32x32 msa neon/, "$sse2_x86inc";
+specialize qw/vpx_dc_predictor_32x32 msa neon sse2/;
 
 add_proto qw/void vpx_dc_top_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_top_predictor_32x32 msa neon/, "$sse2_x86inc";
+specialize qw/vpx_dc_top_predictor_32x32 msa neon sse2/;
 
 add_proto qw/void vpx_dc_left_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_left_predictor_32x32 msa neon/, "$sse2_x86inc";
+specialize qw/vpx_dc_left_predictor_32x32 msa neon sse2/;
 
 add_proto qw/void vpx_dc_128_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vpx_dc_128_predictor_32x32 msa neon/, "$sse2_x86inc";
+specialize qw/vpx_dc_128_predictor_32x32 msa neon sse2/;
 
 # High bitdepth functions
 if (vpx_config("CONFIG_VP9_HIGHBITDEPTH") eq "yes") {
@@ -288,13 +265,13 @@
   specialize qw/vpx_highbd_d153_predictor_4x4/;
 
   add_proto qw/void vpx_highbd_v_predictor_4x4/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_v_predictor_4x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_v_predictor_4x4 sse2/;
 
   add_proto qw/void vpx_highbd_tm_predictor_4x4/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_tm_predictor_4x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_tm_predictor_4x4 sse2/;
 
   add_proto qw/void vpx_highbd_dc_predictor_4x4/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_dc_predictor_4x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_dc_predictor_4x4 sse2/;
 
   add_proto qw/void vpx_highbd_dc_top_predictor_4x4/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
   specialize qw/vpx_highbd_dc_top_predictor_4x4/;
@@ -336,13 +313,13 @@
   specialize qw/vpx_highbd_d153_predictor_8x8/;
 
   add_proto qw/void vpx_highbd_v_predictor_8x8/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_v_predictor_8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_v_predictor_8x8 sse2/;
 
   add_proto qw/void vpx_highbd_tm_predictor_8x8/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_tm_predictor_8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_tm_predictor_8x8 sse2/;
 
   add_proto qw/void vpx_highbd_dc_predictor_8x8/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_dc_predictor_8x8/, "$sse2_x86inc";;
+  specialize qw/vpx_highbd_dc_predictor_8x8 sse2/;;
 
   add_proto qw/void vpx_highbd_dc_top_predictor_8x8/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
   specialize qw/vpx_highbd_dc_top_predictor_8x8/;
@@ -384,13 +361,13 @@
   specialize qw/vpx_highbd_d153_predictor_16x16/;
 
   add_proto qw/void vpx_highbd_v_predictor_16x16/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_v_predictor_16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_v_predictor_16x16 sse2/;
 
   add_proto qw/void vpx_highbd_tm_predictor_16x16/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_tm_predictor_16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_tm_predictor_16x16 sse2/;
 
   add_proto qw/void vpx_highbd_dc_predictor_16x16/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_dc_predictor_16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_dc_predictor_16x16 sse2/;
 
   add_proto qw/void vpx_highbd_dc_top_predictor_16x16/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
   specialize qw/vpx_highbd_dc_top_predictor_16x16/;
@@ -432,13 +409,13 @@
   specialize qw/vpx_highbd_d153_predictor_32x32/;
 
   add_proto qw/void vpx_highbd_v_predictor_32x32/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_v_predictor_32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_v_predictor_32x32 sse2/;
 
   add_proto qw/void vpx_highbd_tm_predictor_32x32/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_tm_predictor_32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_tm_predictor_32x32 sse2/;
 
   add_proto qw/void vpx_highbd_dc_predictor_32x32/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
-  specialize qw/vpx_highbd_dc_predictor_32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_dc_predictor_32x32 sse2/;
 
   add_proto qw/void vpx_highbd_dc_top_predictor_32x32/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
   specialize qw/vpx_highbd_dc_top_predictor_32x32/;
@@ -454,10 +431,10 @@
 # Sub Pixel Filters
 #
 add_proto qw/void vpx_convolve_copy/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vpx_convolve_copy neon dspr2 msa/, "$sse2_x86inc";
+specialize qw/vpx_convolve_copy neon dspr2 msa sse2/;
 
 add_proto qw/void vpx_convolve_avg/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vpx_convolve_avg neon dspr2 msa/, "$sse2_x86inc";
+specialize qw/vpx_convolve_avg neon dspr2 msa sse2/;
 
 add_proto qw/void vpx_convolve8/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
 specialize qw/vpx_convolve8 sse2 ssse3 neon dspr2 msa/, "$avx2_ssse3";
@@ -500,10 +477,10 @@
   # Sub Pixel Filters
   #
   add_proto qw/void vpx_highbd_convolve_copy/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
-  specialize qw/vpx_highbd_convolve_copy/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_convolve_copy sse2/;
 
   add_proto qw/void vpx_highbd_convolve_avg/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
-  specialize qw/vpx_highbd_convolve_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_convolve_avg sse2/;
 
   add_proto qw/void vpx_highbd_convolve8/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
   specialize qw/vpx_highbd_convolve8/, "$sse2_x86_64";
@@ -674,7 +651,7 @@
   specialize qw/vpx_fdct4x4_1 sse2/;
 
   add_proto qw/void vpx_fdct8x8/, "const int16_t *input, tran_low_t *output, int stride";
-  specialize qw/vpx_fdct8x8 sse2 neon msa/, "$ssse3_x86_64_x86inc";
+  specialize qw/vpx_fdct8x8 sse2 neon msa/, "$ssse3_x86_64";
 
   add_proto qw/void vpx_fdct8x8_1/, "const int16_t *input, tran_low_t *output, int stride";
   specialize qw/vpx_fdct8x8_1 sse2 neon msa/;
@@ -706,7 +683,7 @@
   specialize qw/vpx_iwht4x4_1_add/;
 
   add_proto qw/void vpx_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-  specialize qw/vpx_iwht4x4_16_add/, "$sse2_x86inc";
+  specialize qw/vpx_iwht4x4_16_add sse2/;
 
   add_proto qw/void vpx_highbd_idct4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
   specialize qw/vpx_highbd_idct4x4_1_add/;
@@ -792,10 +769,10 @@
     specialize qw/vpx_idct4x4_1_add sse2/;
 
     add_proto qw/void vpx_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct8x8_64_add sse2/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct8x8_64_add sse2/, "$ssse3_x86_64";
 
     add_proto qw/void vpx_idct8x8_12_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct8x8_12_add sse2/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct8x8_12_add sse2/, "$ssse3_x86_64";
 
     add_proto qw/void vpx_idct8x8_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vpx_idct8x8_1_add sse2/;
@@ -810,15 +787,15 @@
     specialize qw/vpx_idct16x16_1_add sse2/;
 
     add_proto qw/void vpx_idct32x32_1024_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct32x32_1024_add sse2/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct32x32_1024_add sse2/, "$ssse3_x86_64";
 
     add_proto qw/void vpx_idct32x32_135_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct32x32_135_add sse2/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct32x32_135_add sse2/, "$ssse3_x86_64";
     # Need to add 135 eob idct32x32 implementations.
     $vpx_idct32x32_135_add_sse2=vpx_idct32x32_1024_add_sse2;
 
     add_proto qw/void vpx_idct32x32_34_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct32x32_34_add sse2/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct32x32_34_add sse2/, "$ssse3_x86_64";
 
     add_proto qw/void vpx_idct32x32_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vpx_idct32x32_1_add sse2/;
@@ -893,10 +870,10 @@
     specialize qw/vpx_idct8x8_1_add sse2 neon dspr2 msa/;
 
     add_proto qw/void vpx_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct8x8_64_add sse2 neon dspr2 msa/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct8x8_64_add sse2 neon dspr2 msa/, "$ssse3_x86_64";
 
     add_proto qw/void vpx_idct8x8_12_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct8x8_12_add sse2 neon dspr2 msa/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct8x8_12_add sse2 neon dspr2 msa/, "$ssse3_x86_64";
 
     add_proto qw/void vpx_idct16x16_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vpx_idct16x16_1_add sse2 neon dspr2 msa/;
@@ -908,10 +885,10 @@
     specialize qw/vpx_idct16x16_10_add sse2 neon dspr2 msa/;
 
     add_proto qw/void vpx_idct32x32_1024_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct32x32_1024_add sse2 neon dspr2 msa/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct32x32_1024_add sse2 neon dspr2 msa/, "$ssse3_x86_64";
 
     add_proto qw/void vpx_idct32x32_135_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct32x32_135_add sse2 neon dspr2 msa/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct32x32_135_add sse2 neon dspr2 msa/, "$ssse3_x86_64";
     # Need to add 135 eob idct32x32 implementations.
     $vpx_idct32x32_135_add_sse2=vpx_idct32x32_1024_add_sse2;
     $vpx_idct32x32_135_add_neon=vpx_idct32x32_1024_add_neon;
@@ -919,7 +896,7 @@
     $vpx_idct32x32_135_add_msa=vpx_idct32x32_1024_add_msa;
 
     add_proto qw/void vpx_idct32x32_34_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_idct32x32_34_add sse2 neon dspr2 msa/, "$ssse3_x86_64_x86inc";
+    specialize qw/vpx_idct32x32_34_add sse2 neon dspr2 msa/, "$ssse3_x86_64";
     # Need to add 34 eob idct32x32 neon implementation.
     $vpx_idct32x32_34_add_neon=vpx_idct32x32_1024_add_neon;
 
@@ -930,7 +907,7 @@
     specialize qw/vpx_iwht4x4_1_add msa/;
 
     add_proto qw/void vpx_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vpx_iwht4x4_16_add msa/, "$sse2_x86inc";
+    specialize qw/vpx_iwht4x4_16_add msa sse2/;
   }  # CONFIG_EMULATE_HARDWARE
 }  # CONFIG_VP9_HIGHBITDEPTH
 }  # CONFIG_VP9
@@ -940,10 +917,10 @@
 #
 if (vpx_config("CONFIG_VP9_ENCODER") eq "yes") {
   add_proto qw/void vpx_quantize_b/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
-  specialize qw/vpx_quantize_b sse2/, "$ssse3_x86_64_x86inc", "$avx_x86_64_x86inc";
+  specialize qw/vpx_quantize_b sse2/, "$ssse3_x86_64", "$avx_x86_64";
 
   add_proto qw/void vpx_quantize_b_32x32/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
-  specialize qw/vpx_quantize_b_32x32/, "$ssse3_x86_64_x86inc", "$avx_x86_64_x86inc";
+  specialize qw/vpx_quantize_b_32x32/, "$ssse3_x86_64", "$avx_x86_64";
 
   if (vpx_config("CONFIG_VP9_HIGHBITDEPTH") eq "yes") {
     add_proto qw/void vpx_highbd_quantize_b/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
@@ -959,49 +936,49 @@
 # Block subtraction
 #
 add_proto qw/void vpx_subtract_block/, "int rows, int cols, int16_t *diff_ptr, ptrdiff_t diff_stride, const uint8_t *src_ptr, ptrdiff_t src_stride, const uint8_t *pred_ptr, ptrdiff_t pred_stride";
-specialize qw/vpx_subtract_block neon msa/, "$sse2_x86inc";
+specialize qw/vpx_subtract_block neon msa sse2/;
 
 #
 # Single block SAD
 #
 add_proto qw/unsigned int vpx_sad64x64/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad64x64 avx2 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad64x64 avx2 neon msa sse2/;
 
 add_proto qw/unsigned int vpx_sad64x32/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad64x32 avx2 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad64x32 avx2 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad32x64/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad32x64 avx2 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x64 avx2 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad32x32/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad32x32 avx2 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x32 avx2 neon msa sse2/;
 
 add_proto qw/unsigned int vpx_sad32x16/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad32x16 avx2 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x16 avx2 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad16x32/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad16x32 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x32 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad16x16/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad16x16 media neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x16 media neon msa sse2/;
 
 add_proto qw/unsigned int vpx_sad16x8/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad16x8 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x8 neon msa sse2/;
 
 add_proto qw/unsigned int vpx_sad8x16/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad8x16 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x16 neon msa sse2/;
 
 add_proto qw/unsigned int vpx_sad8x8/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad8x8 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x8 neon msa sse2/;
 
 add_proto qw/unsigned int vpx_sad8x4/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad8x4 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x4 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad4x8/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad4x8 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad4x8 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad4x4/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-specialize qw/vpx_sad4x4 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad4x4 neon msa sse2/;
 
 #
 # Avg
@@ -1017,7 +994,7 @@
   specialize qw/vpx_minmax_8x8 sse2 neon/;
 
   add_proto qw/void vpx_hadamard_8x8/, "const int16_t *src_diff, int src_stride, int16_t *coeff";
-  specialize qw/vpx_hadamard_8x8 sse2 neon/, "$ssse3_x86_64_x86inc";
+  specialize qw/vpx_hadamard_8x8 sse2 neon/, "$ssse3_x86_64";
 
   add_proto qw/void vpx_hadamard_16x16/, "const int16_t *src_diff, int src_stride, int16_t *coeff";
   specialize qw/vpx_hadamard_16x16 sse2 neon/;
@@ -1036,43 +1013,43 @@
 }  # CONFIG_VP9_ENCODER
 
 add_proto qw/unsigned int vpx_sad64x64_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad64x64_avg avx2 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad64x64_avg avx2 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad64x32_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad64x32_avg avx2 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad64x32_avg avx2 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad32x64_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad32x64_avg avx2 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x64_avg avx2 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad32x32_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad32x32_avg avx2 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x32_avg avx2 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad32x16_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad32x16_avg avx2 msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x16_avg avx2 msa sse2/;
 
 add_proto qw/unsigned int vpx_sad16x32_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad16x32_avg msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x32_avg msa sse2/;
 
 add_proto qw/unsigned int vpx_sad16x16_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad16x16_avg msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x16_avg msa sse2/;
 
 add_proto qw/unsigned int vpx_sad16x8_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad16x8_avg msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x8_avg msa sse2/;
 
 add_proto qw/unsigned int vpx_sad8x16_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad8x16_avg msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x16_avg msa sse2/;
 
 add_proto qw/unsigned int vpx_sad8x8_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad8x8_avg msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x8_avg msa sse2/;
 
 add_proto qw/unsigned int vpx_sad8x4_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad8x4_avg msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x4_avg msa sse2/;
 
 add_proto qw/unsigned int vpx_sad4x8_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad4x8_avg msa/, "$sse2_x86inc";
+specialize qw/vpx_sad4x8_avg msa sse2/;
 
 add_proto qw/unsigned int vpx_sad4x4_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-specialize qw/vpx_sad4x4_avg msa/, "$sse2_x86inc";
+specialize qw/vpx_sad4x4_avg msa sse2/;
 
 #
 # Multi-block SAD, comparing a reference to N blocks 1 pixel apart horizontally
@@ -1131,43 +1108,43 @@
 # Multi-block SAD, comparing a reference to N independent blocks
 #
 add_proto qw/void vpx_sad64x64x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad64x64x4d avx2 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad64x64x4d avx2 neon msa sse2/;
 
 add_proto qw/void vpx_sad64x32x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad64x32x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad64x32x4d msa sse2/;
 
 add_proto qw/void vpx_sad32x64x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad32x64x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x64x4d msa sse2/;
 
 add_proto qw/void vpx_sad32x32x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad32x32x4d avx2 neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x32x4d avx2 neon msa sse2/;
 
 add_proto qw/void vpx_sad32x16x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad32x16x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad32x16x4d msa sse2/;
 
 add_proto qw/void vpx_sad16x32x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad16x32x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x32x4d msa sse2/;
 
 add_proto qw/void vpx_sad16x16x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad16x16x4d neon msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x16x4d neon msa sse2/;
 
 add_proto qw/void vpx_sad16x8x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad16x8x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad16x8x4d msa sse2/;
 
 add_proto qw/void vpx_sad8x16x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad8x16x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x16x4d msa sse2/;
 
 add_proto qw/void vpx_sad8x8x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad8x8x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x8x4d msa sse2/;
 
 add_proto qw/void vpx_sad8x4x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad8x4x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad8x4x4d msa sse2/;
 
 add_proto qw/void vpx_sad4x8x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad4x8x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad4x8x4d msa sse2/;
 
 add_proto qw/void vpx_sad4x4x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[], int ref_stride, uint32_t *sad_array";
-specialize qw/vpx_sad4x4x4d msa/, "$sse2_x86inc";
+specialize qw/vpx_sad4x4x4d msa sse2/;
 
 #
 # Structured Similarity (SSIM)
@@ -1191,37 +1168,37 @@
   # Single block SAD
   #
   add_proto qw/unsigned int vpx_highbd_sad64x64/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad64x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad64x64 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad64x32/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad64x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad64x32 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad32x64/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad32x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x64 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad32x32/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x32 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad32x16/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad32x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x16 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad16x32/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad16x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x32 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad16x16/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x16 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad16x8/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad16x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x8 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad8x16/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad8x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x16 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad8x8/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x8 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad8x4/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-  specialize qw/vpx_highbd_sad8x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x4 sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad4x8/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
   specialize qw/vpx_highbd_sad4x8/;
@@ -1240,37 +1217,37 @@
   specialize qw/vpx_highbd_minmax_8x8/;
 
   add_proto qw/unsigned int vpx_highbd_sad64x64_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad64x64_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad64x64_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad64x32_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad64x32_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad64x32_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad32x64_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad32x64_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x64_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad32x32_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad32x32_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x32_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad32x16_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad32x16_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x16_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad16x32_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad16x32_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x32_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad16x16_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad16x16_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x16_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad16x8_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad16x8_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x8_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad8x16_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad8x16_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x16_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad8x8_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad8x8_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x8_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad8x4_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_sad8x4_avg/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x4_avg sse2/;
 
   add_proto qw/unsigned int vpx_highbd_sad4x8_avg/, "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
   specialize qw/vpx_highbd_sad4x8_avg/;
@@ -1335,43 +1312,43 @@
   # Multi-block SAD, comparing a reference to N independent blocks
   #
   add_proto qw/void vpx_highbd_sad64x64x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad64x64x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad64x64x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad64x32x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad64x32x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad64x32x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad32x64x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad32x64x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x64x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad32x32x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad32x32x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x32x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad32x16x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad32x16x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad32x16x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad16x32x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad16x32x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x32x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad16x16x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad16x16x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x16x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad16x8x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad16x8x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad16x8x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad8x16x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad8x16x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x16x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad8x8x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad8x8x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x8x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad8x4x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad8x4x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad8x4x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad4x8x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad4x8x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad4x8x4d sse2/;
 
   add_proto qw/void vpx_highbd_sad4x4x4d/, "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, uint32_t *sad_array";
-  specialize qw/vpx_highbd_sad4x4x4d/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_sad4x4x4d sse2/;
 
   #
   # Structured Similarity (SSIM)
@@ -1460,82 +1437,82 @@
 # Subpixel Variance
 #
 add_proto qw/uint32_t vpx_sub_pixel_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance64x64 avx2 neon msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance64x64 avx2 neon msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance64x32 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance64x32 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance32x64 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance32x64 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance32x32 avx2 neon msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance32x32 avx2 neon msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance32x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance32x16 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance16x32 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance16x32 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance16x16 media neon msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance16x16 media neon msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance16x8 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance16x8 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance8x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance8x16 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance8x8 media neon msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance8x8 media neon msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance8x4 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance8x4 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance4x8 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance4x8 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_sub_pixel_variance4x4 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_variance4x4 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance64x64 avx2 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance64x64 avx2 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance64x32 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance64x32 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance32x64 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance32x64 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance32x32 avx2 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance32x32 avx2 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance32x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance32x16 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance16x32 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance16x32 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance16x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance16x16 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance16x8 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance16x8 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance8x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance8x16 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance8x8 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance8x8 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance8x4 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance8x4 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance4x8 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance4x8 msa sse2 ssse3/;
 
 add_proto qw/uint32_t vpx_sub_pixel_avg_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_sub_pixel_avg_variance4x4 msa/, "$sse2_x86inc", "$ssse3_x86inc";
+  specialize qw/vpx_sub_pixel_avg_variance4x4 msa sse2 ssse3/;
 
 #
 # Specialty Subpixel
@@ -1691,217 +1668,217 @@
   # Subpixel Variance
   #
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance64x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance64x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance64x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance64x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance32x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance32x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance32x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance32x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance32x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance16x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance16x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance16x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance16x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance16x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance8x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance8x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance8x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_12_sub_pixel_variance8x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_variance8x4 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance64x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance64x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance64x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance64x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance32x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance32x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance32x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance32x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance32x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance16x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance16x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance16x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance16x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance16x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance8x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance8x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance8x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_10_sub_pixel_variance8x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_variance8x4 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance64x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance64x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance64x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance64x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance32x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance32x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance32x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance32x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance32x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance16x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance16x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance16x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance16x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance16x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance8x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance8x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance8x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
-  specialize qw/vpx_highbd_8_sub_pixel_variance8x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_variance8x4 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse";
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance64x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance64x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance64x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance64x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance32x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance32x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance32x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance32x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance32x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance16x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance16x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance16x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance16x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance16x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance8x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance8x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance8x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_12_sub_pixel_avg_variance8x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_12_sub_pixel_avg_variance8x4 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
   add_proto qw/uint32_t vpx_highbd_12_sub_pixel_avg_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance64x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance64x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance64x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance64x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance32x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance32x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance32x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance32x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance32x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance16x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance16x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance16x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance16x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance16x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance8x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance8x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance8x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_10_sub_pixel_avg_variance8x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_10_sub_pixel_avg_variance8x4 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
   add_proto qw/uint32_t vpx_highbd_10_sub_pixel_avg_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance64x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance64x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance64x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance64x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance32x64/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance32x64 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance32x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance32x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance32x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance32x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance16x32/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance16x32 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance16x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance16x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance16x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance16x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance8x16/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance8x16 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance8x8/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance8x8 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
-  specialize qw/vpx_highbd_8_sub_pixel_avg_variance8x4/, "$sse2_x86inc";
+  specialize qw/vpx_highbd_8_sub_pixel_avg_variance8x4 sse2/;
 
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
   add_proto qw/uint32_t vpx_highbd_8_sub_pixel_avg_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
--- a/vpx_dsp/x86/highbd_variance_sse2.c
+++ b/vpx_dsp/x86/highbd_variance_sse2.c
@@ -246,7 +246,6 @@
   return *sse;
 }
 
-#if CONFIG_USE_X86INC
 // The 2 unused parameters are place holders for PIC enabled build.
 // These definitions are for functions defined in
 // highbd_subpel_variance_impl_sse2.asm
@@ -593,4 +592,3 @@
 
 #undef FNS
 #undef FN
-#endif  // CONFIG_USE_X86INC
--- a/vpx_dsp/x86/variance_sse2.c
+++ b/vpx_dsp/x86/variance_sse2.c
@@ -308,7 +308,6 @@
   return *sse;
 }
 
-#if CONFIG_USE_X86INC
 // The 2 unused parameters are place holders for PIC enabled build.
 // These definitions are for functions defined in subpel_variance.asm
 #define DECL(w, opt) \
@@ -474,4 +473,3 @@
 
 #undef FNS
 #undef FN
-#endif  // CONFIG_USE_X86INC