rm unused aom_sad_skip functions

This change is the same as the one that removed unused
aom_highbd_sad_skip versions.

These functions are only assigned to the sdf and sdx4df function
pointers in mcomp.c when the block height is >= 16.

Bug: aomedia:3416
Change-Id: I1ae313163ba074236f86ce92d3e03ad00ca87a36
diff --git a/aom_dsp/aom_dsp_rtcd_defs.pl b/aom_dsp/aom_dsp_rtcd_defs.pl
index 4373ea0..bc0f596 100755
--- a/aom_dsp/aom_dsp_rtcd_defs.pl
+++ b/aom_dsp/aom_dsp_rtcd_defs.pl
@@ -838,7 +838,9 @@
   foreach (@encoder_block_sizes) {
     ($w, $h) = @$_;
     add_proto qw/unsigned int/, "aom_sad${w}x${h}", "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
-    add_proto qw/unsigned int/, "aom_sad_skip_${w}x${h}", "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
+    if ($h >= 16) {
+      add_proto qw/unsigned int/, "aom_sad_skip_${w}x${h}", "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride";
+    }
     add_proto qw/unsigned int/, "aom_sad${w}x${h}_avg", "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred";
     add_proto qw/unsigned int/, "aom_dist_wtd_sad${w}x${h}_avg", "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred, const DIST_WTD_COMP_PARAMS *jcp_param";
   }
@@ -881,15 +883,9 @@
   specialize qw/aom_sad_skip_16x16           sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_16x8            sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_8x16            sse2 neon/;
-  specialize qw/aom_sad_skip_8x8             sse2 neon/;
-  specialize qw/aom_sad_skip_8x4                  neon/;
-  specialize qw/aom_sad_skip_4x8             sse2 neon/;
-  specialize qw/aom_sad_skip_4x4                  neon/;
 
   specialize qw/aom_sad_skip_4x16            sse2 neon/;
-  specialize qw/aom_sad_skip_16x4                 neon neon_dotprod/;
   specialize qw/aom_sad_skip_8x32            sse2 neon/;
-  specialize qw/aom_sad_skip_32x8            sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_16x64           sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_64x16           sse2 neon neon_dotprod/;
 
@@ -1069,7 +1065,9 @@
     ($w, $h) = @$_;
     add_proto qw/void/, "aom_sad${w}x${h}x4d", "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[4], int ref_stride, uint32_t sad_array[4]";
     add_proto qw/void/, "aom_sad${w}x${h}x3d", "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[4], int ref_stride, uint32_t sad_array[4]";
-    add_proto qw/void/, "aom_sad_skip_${w}x${h}x4d", "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[4], int ref_stride, uint32_t sad_array[4]";
+    if ($h >= 16) {
+      add_proto qw/void/, "aom_sad_skip_${w}x${h}x4d", "const uint8_t *src_ptr, int src_stride, const uint8_t * const ref_ptr[4], int ref_stride, uint32_t sad_array[4]";
+    }
   }
 
   specialize qw/aom_sad128x128x4d avx2 sse2 neon neon_dotprod/;
@@ -1106,20 +1104,14 @@
   specialize qw/aom_sad_skip_32x64x4d   avx2 sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_32x32x4d   avx2 sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_32x16x4d   avx2 sse2 neon neon_dotprod/;
-  specialize qw/aom_sad_skip_32x8x4d    avx2 sse2 neon neon_dotprod/;
 
   specialize qw/aom_sad_skip_16x64x4d   avx2 sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_16x32x4d   avx2 sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_16x16x4d   avx2 sse2 neon neon_dotprod/;
   specialize qw/aom_sad_skip_16x8x4d    avx2 sse2 neon neon_dotprod/;
-  specialize qw/aom_sad_skip_16x4x4d    avx2      neon neon_dotprod/;
   specialize qw/aom_sad_skip_8x32x4d         sse2 neon/;
   specialize qw/aom_sad_skip_8x16x4d         sse2 neon/;
-  specialize qw/aom_sad_skip_8x8x4d          sse2 neon/;
-  specialize qw/aom_sad_skip_8x4x4d               neon/;
   specialize qw/aom_sad_skip_4x16x4d         sse2 neon/;
-  specialize qw/aom_sad_skip_4x8x4d          sse2 neon/;
-  specialize qw/aom_sad_skip_4x4x4d               neon/;
 
   specialize qw/aom_sad128x128x3d avx2 neon neon_dotprod/;
   specialize qw/aom_sad128x64x3d  avx2 neon neon_dotprod/;
diff --git a/aom_dsp/arm/sad_neon.c b/aom_dsp/arm/sad_neon.c
index 639b9ba..121af83 100644
--- a/aom_dsp/arm/sad_neon.c
+++ b/aom_dsp/arm/sad_neon.c
@@ -263,14 +263,8 @@
            sad##w##xh_neon(src, 2 * src_stride, ref, 2 * ref_stride, (h) / 2); \
   }
 
-SAD_SKIP_WXH_NEON(4, 4)
-SAD_SKIP_WXH_NEON(4, 8)
-
-SAD_SKIP_WXH_NEON(8, 4)
-SAD_SKIP_WXH_NEON(8, 8)
 SAD_SKIP_WXH_NEON(8, 16)
 
-SAD_SKIP_WXH_NEON(16, 8)
 SAD_SKIP_WXH_NEON(16, 16)
 SAD_SKIP_WXH_NEON(16, 32)
 
@@ -288,9 +282,7 @@
 #if !CONFIG_REALTIME_ONLY
 SAD_SKIP_WXH_NEON(4, 16)
 SAD_SKIP_WXH_NEON(8, 32)
-SAD_SKIP_WXH_NEON(16, 4)
 SAD_SKIP_WXH_NEON(16, 64)
-SAD_SKIP_WXH_NEON(32, 8)
 SAD_SKIP_WXH_NEON(64, 16)
 #endif  // !CONFIG_REALTIME_ONLY
 
diff --git a/aom_dsp/arm/sad_neon_dotprod.c b/aom_dsp/arm/sad_neon_dotprod.c
index c24db42..5403c0a 100644
--- a/aom_dsp/arm/sad_neon_dotprod.c
+++ b/aom_dsp/arm/sad_neon_dotprod.c
@@ -143,7 +143,6 @@
                                        2 * ref_stride, (h) / 2); \
   }
 
-SAD_SKIP_WXH_NEON_DOTPROD(16, 8)
 SAD_SKIP_WXH_NEON_DOTPROD(16, 16)
 SAD_SKIP_WXH_NEON_DOTPROD(16, 32)
 
@@ -159,9 +158,7 @@
 SAD_SKIP_WXH_NEON_DOTPROD(128, 128)
 
 #if !CONFIG_REALTIME_ONLY
-SAD_SKIP_WXH_NEON_DOTPROD(16, 4)
 SAD_SKIP_WXH_NEON_DOTPROD(16, 64)
-SAD_SKIP_WXH_NEON_DOTPROD(32, 8)
 SAD_SKIP_WXH_NEON_DOTPROD(64, 16)
 #endif  // !CONFIG_REALTIME_ONLY
 
diff --git a/aom_dsp/arm/sadxd_neon.c b/aom_dsp/arm/sadxd_neon.c
index b528281..3b88069 100644
--- a/aom_dsp/arm/sadxd_neon.c
+++ b/aom_dsp/arm/sadxd_neon.c
@@ -480,14 +480,8 @@
     res[3] <<= 1;                                                           \
   }
 
-SAD_SKIP_WXH_4D_NEON(4, 4)
-SAD_SKIP_WXH_4D_NEON(4, 8)
-
-SAD_SKIP_WXH_4D_NEON(8, 4)
-SAD_SKIP_WXH_4D_NEON(8, 8)
 SAD_SKIP_WXH_4D_NEON(8, 16)
 
-SAD_SKIP_WXH_4D_NEON(16, 8)
 SAD_SKIP_WXH_4D_NEON(16, 16)
 SAD_SKIP_WXH_4D_NEON(16, 32)
 
@@ -505,9 +499,7 @@
 #if !CONFIG_REALTIME_ONLY
 SAD_SKIP_WXH_4D_NEON(4, 16)
 SAD_SKIP_WXH_4D_NEON(8, 32)
-SAD_SKIP_WXH_4D_NEON(16, 4)
 SAD_SKIP_WXH_4D_NEON(16, 64)
-SAD_SKIP_WXH_4D_NEON(32, 8)
 SAD_SKIP_WXH_4D_NEON(64, 16)
 #endif  // !CONFIG_REALTIME_ONLY
 
diff --git a/aom_dsp/arm/sadxd_neon_dotprod.c b/aom_dsp/arm/sadxd_neon_dotprod.c
index 317e7ce..339e428 100644
--- a/aom_dsp/arm/sadxd_neon_dotprod.c
+++ b/aom_dsp/arm/sadxd_neon_dotprod.c
@@ -264,7 +264,6 @@
     res[3] <<= 1;                                                             \
   }
 
-SAD_SKIP_WXH_4D_NEON_DOTPROD(16, 8)
 SAD_SKIP_WXH_4D_NEON_DOTPROD(16, 16)
 SAD_SKIP_WXH_4D_NEON_DOTPROD(16, 32)
 
@@ -280,9 +279,7 @@
 SAD_SKIP_WXH_4D_NEON_DOTPROD(128, 128)
 
 #if !CONFIG_REALTIME_ONLY
-SAD_SKIP_WXH_4D_NEON_DOTPROD(16, 4)
 SAD_SKIP_WXH_4D_NEON_DOTPROD(16, 64)
-SAD_SKIP_WXH_4D_NEON_DOTPROD(32, 8)
 SAD_SKIP_WXH_4D_NEON_DOTPROD(64, 16)
 #endif  // !CONFIG_REALTIME_ONLY
 
diff --git a/aom_dsp/sad.c b/aom_dsp/sad.c
index 7552019..9609672 100644
--- a/aom_dsp/sad.c
+++ b/aom_dsp/sad.c
@@ -35,7 +35,7 @@
   return sad;
 }
 
-#define SADMXN(m, n)                                                          \
+#define SADMXN_NO_SKIP(m, n)                                                  \
   unsigned int aom_sad##m##x##n##_c(const uint8_t *src, int src_stride,       \
                                     const uint8_t *ref, int ref_stride) {     \
     return sad(src, src_stride, ref, ref_stride, m, n);                       \
@@ -54,7 +54,10 @@
     aom_dist_wtd_comp_avg_pred_c(comp_pred, second_pred, m, n, ref,           \
                                  ref_stride, jcp_param);                      \
     return sad(src, src_stride, comp_pred, m, m, n);                          \
-  }                                                                           \
+  }
+
+#define SADMXN(m, n)                                                          \
+  SADMXN_NO_SKIP(m, n)                                                        \
   unsigned int aom_sad_skip_##m##x##n##_c(const uint8_t *src, int src_stride, \
                                           const uint8_t *ref,                 \
                                           int ref_stride) {                   \
@@ -62,16 +65,19 @@
   }
 
 // Calculate sad against 4 reference locations and store each in sad_array
+#define SAD_MXNX4D_NO_SKIP(m, n)                                           \
+  void aom_sad##m##x##n##x4d_c(const uint8_t *src, int src_stride,         \
+                               const uint8_t *const ref_array[4],          \
+                               int ref_stride, uint32_t sad_array[4]) {    \
+    int i;                                                                 \
+    for (i = 0; i < 4; ++i) {                                              \
+      sad_array[i] =                                                       \
+          aom_sad##m##x##n##_c(src, src_stride, ref_array[i], ref_stride); \
+    }                                                                      \
+  }
+
 #define SAD_MXNX4D(m, n)                                                      \
-  void aom_sad##m##x##n##x4d_c(const uint8_t *src, int src_stride,            \
-                               const uint8_t *const ref_array[4],             \
-                               int ref_stride, uint32_t sad_array[4]) {       \
-    int i;                                                                    \
-    for (i = 0; i < 4; ++i) {                                                 \
-      sad_array[i] =                                                          \
-          aom_sad##m##x##n##_c(src, src_stride, ref_array[i], ref_stride);    \
-    }                                                                         \
-  }                                                                           \
+  SAD_MXNX4D_NO_SKIP(m, n)                                                    \
   void aom_sad_skip_##m##x##n##x4d_c(const uint8_t *src, int src_stride,      \
                                      const uint8_t *const ref_array[4],       \
                                      int ref_stride, uint32_t sad_array[4]) { \
@@ -140,8 +146,8 @@
 SAD_MXNX3D(16, 16)
 
 // 16x8
-SADMXN(16, 8)
-SAD_MXNX4D(16, 8)
+SADMXN_NO_SKIP(16, 8)
+SAD_MXNX4D_NO_SKIP(16, 8)
 SAD_MXNX3D(16, 8)
 
 // 8x16
@@ -150,34 +156,34 @@
 SAD_MXNX3D(8, 16)
 
 // 8x8
-SADMXN(8, 8)
-SAD_MXNX4D(8, 8)
+SADMXN_NO_SKIP(8, 8)
+SAD_MXNX4D_NO_SKIP(8, 8)
 SAD_MXNX3D(8, 8)
 
 // 8x4
-SADMXN(8, 4)
-SAD_MXNX4D(8, 4)
+SADMXN_NO_SKIP(8, 4)
+SAD_MXNX4D_NO_SKIP(8, 4)
 SAD_MXNX3D(8, 4)
 
 // 4x8
-SADMXN(4, 8)
-SAD_MXNX4D(4, 8)
+SADMXN_NO_SKIP(4, 8)
+SAD_MXNX4D_NO_SKIP(4, 8)
 SAD_MXNX3D(4, 8)
 
 // 4x4
-SADMXN(4, 4)
-SAD_MXNX4D(4, 4)
+SADMXN_NO_SKIP(4, 4)
+SAD_MXNX4D_NO_SKIP(4, 4)
 SAD_MXNX3D(4, 4)
 
 #if !CONFIG_REALTIME_ONLY
 SADMXN(4, 16)
 SAD_MXNX4D(4, 16)
-SADMXN(16, 4)
-SAD_MXNX4D(16, 4)
+SADMXN_NO_SKIP(16, 4)
+SAD_MXNX4D_NO_SKIP(16, 4)
 SADMXN(8, 32)
 SAD_MXNX4D(8, 32)
-SADMXN(32, 8)
-SAD_MXNX4D(32, 8)
+SADMXN_NO_SKIP(32, 8)
+SAD_MXNX4D_NO_SKIP(32, 8)
 SADMXN(16, 64)
 SAD_MXNX4D(16, 64)
 SADMXN(64, 16)
diff --git a/aom_dsp/x86/sad4d_avx2.c b/aom_dsp/x86/sad4d_avx2.c
index 324b142..0c1a03c 100644
--- a/aom_dsp/x86/sad4d_avx2.c
+++ b/aom_dsp/x86/sad4d_avx2.c
@@ -182,7 +182,6 @@
 SAD_SKIP_MXN_AVX2(128, 128)
 
 #if !CONFIG_REALTIME_ONLY
-SAD_SKIP_MXN_AVX2(32, 8)
 SAD_SKIP_MXN_AVX2(64, 16)
 #endif  // !CONFIG_REALTIME_ONLY
 
@@ -324,9 +323,7 @@
 
 SAD_SKIP_16XN_AVX2(32)
 SAD_SKIP_16XN_AVX2(16)
-SAD_SKIP_16XN_AVX2(8)
 
 #if !CONFIG_REALTIME_ONLY
 SAD_SKIP_16XN_AVX2(64)
-SAD_SKIP_16XN_AVX2(4)
 #endif  // !CONFIG_REALTIME_ONLY
diff --git a/aom_dsp/x86/sad4d_sse2.asm b/aom_dsp/x86/sad4d_sse2.asm
index 3384215..11eba08 100644
--- a/aom_dsp/x86/sad4d_sse2.asm
+++ b/aom_dsp/x86/sad4d_sse2.asm
@@ -419,19 +419,10 @@
 SADNXN4D  32,  16, 1
 SADNXN4D  16,  32, 1
 SADNXN4D  16,  16, 1
-SADNXN4D  16,   8, 1
 SADNXN4D   8,  16, 1
-SADNXN4D   8,   8, 1
-SADNXN4D   4,   8, 1
 %if CONFIG_REALTIME_ONLY==0
 SADNXN4D   4,  16, 1
 SADNXN4D   8,  32, 1
-SADNXN4D  32,   8, 1
 SADNXN4D  16,  64, 1
 SADNXN4D  64,  16, 1
 %endif
-
-; Different assembly is needed when the height gets subsampled to 2
-; SADNXN4D 16,  4, 1
-; SADNXN4D  8,  4, 1
-; SADNXN4D  4,  4, 1
diff --git a/aom_dsp/x86/sad_sse2.asm b/aom_dsp/x86/sad_sse2.asm
index 424a954..5cec603 100644
--- a/aom_dsp/x86/sad_sse2.asm
+++ b/aom_dsp/x86/sad_sse2.asm
@@ -259,7 +259,6 @@
 %if CONFIG_REALTIME_ONLY==0
 SAD32XN  8    ; sad_32x8_sse2
 SAD32XN  8, 1 ; sad_32x8_avg_sse2
-SAD32XN  8, 2 ; sad_32x8_skip_sse2
 %endif
 
 ; unsigned int aom_sad16x{8,16}_sse2(uint8_t *src, int src_stride,
@@ -316,7 +315,6 @@
 SAD16XN  8, 1 ; sad16x8_avg_sse2
 SAD16XN 32, 2 ; sad16x32_skip_sse2
 SAD16XN 16, 2 ; sad16x16_skip_sse2
-SAD16XN  8, 2 ; sad16x8_skip_sse2
 %if CONFIG_REALTIME_ONLY==0
 SAD16XN 64    ; sad_16x64_sse2
 SAD16XN  4    ; sad_16x4_sse2
@@ -434,7 +432,6 @@
 SAD4XN  4 ; sad4x4_sse2
 SAD4XN  8, 1 ; sad4x8_avg_sse2
 SAD4XN  4, 1 ; sad4x4_avg_sse2
-SAD4XN  8, 2 ; sad4x8_skip_sse2
 %if CONFIG_REALTIME_ONLY==0
 SAD4XN 16 ; sad_4x16_sse2
 SAD4XN 16, 1 ; sad_4x16_avg_sse2
diff --git a/av1/encoder/encoder.c b/av1/encoder/encoder.c
index ce142ec..6df5e2e 100644
--- a/av1/encoder/encoder.c
+++ b/av1/encoder/encoder.c
@@ -1272,16 +1272,11 @@
 
   SDSFP(BLOCK_16X32, aom_sad_skip_16x32, aom_sad_skip_16x32x4d)
   SDSFP(BLOCK_16X16, aom_sad_skip_16x16, aom_sad_skip_16x16x4d)
-  SDSFP(BLOCK_16X8, aom_sad_skip_16x8, aom_sad_skip_16x8x4d)
   SDSFP(BLOCK_8X16, aom_sad_skip_8x16, aom_sad_skip_8x16x4d)
-  SDSFP(BLOCK_8X8, aom_sad_skip_8x8, aom_sad_skip_8x8x4d)
-
-  SDSFP(BLOCK_4X8, aom_sad_skip_4x8, aom_sad_skip_4x8x4d)
 
 #if !CONFIG_REALTIME_ONLY
   SDSFP(BLOCK_64X16, aom_sad_skip_64x16, aom_sad_skip_64x16x4d)
   SDSFP(BLOCK_16X64, aom_sad_skip_16x64, aom_sad_skip_16x64x4d)
-  SDSFP(BLOCK_32X8, aom_sad_skip_32x8, aom_sad_skip_32x8x4d)
   SDSFP(BLOCK_8X32, aom_sad_skip_8x32, aom_sad_skip_8x32x4d)
   SDSFP(BLOCK_4X16, aom_sad_skip_4x16, aom_sad_skip_4x16x4d)
 #endif
diff --git a/test/sad_test.cc b/test/sad_test.cc
index 96585ef..fb8c4b4 100644
--- a/test/sad_test.cc
+++ b/test/sad_test.cc
@@ -1153,18 +1153,11 @@
   make_tuple(32, 16, &aom_sad_skip_32x16_c, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32_c, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16_c, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8_c, -1),
   make_tuple(8, 16, &aom_sad_skip_8x16_c, -1),
-  make_tuple(8, 8, &aom_sad_skip_8x8_c, -1),
-  make_tuple(8, 4, &aom_sad_skip_8x4_c, -1),
-  make_tuple(4, 8, &aom_sad_skip_4x8_c, -1),
-  make_tuple(4, 4, &aom_sad_skip_4x4_c, -1),
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16_c, -1),
   make_tuple(16, 64, &aom_sad_skip_16x64_c, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8_c, -1),
   make_tuple(8, 32, &aom_sad_skip_8x32_c, -1),
-  make_tuple(16, 4, &aom_sad_skip_16x4_c, -1),
   make_tuple(4, 16, &aom_sad_skip_4x16_c, -1),
 #endif
 #if CONFIG_AV1_HIGHBITDEPTH
@@ -1571,14 +1564,10 @@
   make_tuple(32, 16, &aom_sad_skip_32x16x4d_c, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32x4d_c, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16x4d_c, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8x4d_c, -1),
   make_tuple(8, 16, &aom_sad_skip_8x16x4d_c, -1),
-  make_tuple(8, 8, &aom_sad_skip_8x8x4d_c, -1),
-  make_tuple(4, 8, &aom_sad_skip_4x8x4d_c, -1),
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16x4d_c, -1),
   make_tuple(16, 64, &aom_sad_skip_16x64x4d_c, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8x4d_c, -1),
   make_tuple(8, 32, &aom_sad_skip_8x32x4d_c, -1),
   make_tuple(4, 16, &aom_sad_skip_4x16x4d_c, -1),
 #endif
@@ -1850,12 +1839,7 @@
   make_tuple(32, 16, &aom_sad_skip_32x16_neon, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32_neon, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16_neon, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8_neon, -1),
   make_tuple(8, 16, &aom_sad_skip_8x16_neon, -1),
-  make_tuple(8, 8, &aom_sad_skip_8x8_neon, -1),
-  make_tuple(8, 4, &aom_sad_skip_8x4_neon, -1),
-  make_tuple(4, 8, &aom_sad_skip_4x8_neon, -1),
-  make_tuple(4, 4, &aom_sad_skip_4x4_neon, -1),
 #if CONFIG_AV1_HIGHBITDEPTH
   make_tuple(128, 128, &aom_highbd_sad_skip_128x128_neon, 8),
   make_tuple(128, 64, &aom_highbd_sad_skip_128x64_neon, 8),
@@ -1893,9 +1877,7 @@
 #endif  // CONFIG_AV1_HIGHBITDEPTH
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16_neon, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8_neon, -1),
   make_tuple(16, 64, &aom_sad_skip_16x64_neon, -1),
-  make_tuple(16, 4, &aom_sad_skip_16x4_neon, -1),
   make_tuple(8, 32, &aom_sad_skip_8x32_neon, -1),
   make_tuple(4, 16, &aom_sad_skip_4x16_neon, -1),
 #if CONFIG_AV1_HIGHBITDEPTH
@@ -1928,12 +1910,7 @@
   make_tuple(32, 16, &aom_sad_skip_32x16x4d_neon, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32x4d_neon, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16x4d_neon, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8x4d_neon, -1),
   make_tuple(8, 16, &aom_sad_skip_8x16x4d_neon, -1),
-  make_tuple(8, 8, &aom_sad_skip_8x8x4d_neon, -1),
-  make_tuple(8, 4, &aom_sad_skip_8x4x4d_neon, -1),
-  make_tuple(4, 8, &aom_sad_skip_4x8x4d_neon, -1),
-  make_tuple(4, 4, &aom_sad_skip_4x4x4d_neon, -1),
 #if CONFIG_AV1_HIGHBITDEPTH
   make_tuple(128, 128, &aom_highbd_sad_skip_128x128x4d_neon, 8),
   make_tuple(128, 64, &aom_highbd_sad_skip_128x64x4d_neon, 8),
@@ -1971,9 +1948,7 @@
 #endif  // CONFIG_AV1_HIGHBITDEPTH
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16x4d_neon, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8x4d_neon, -1),
   make_tuple(16, 64, &aom_sad_skip_16x64x4d_neon, -1),
-  make_tuple(16, 4, &aom_sad_skip_16x4x4d_neon, -1),
   make_tuple(8, 32, &aom_sad_skip_8x32x4d_neon, -1),
   make_tuple(4, 16, &aom_sad_skip_4x16x4d_neon, -1),
 #if CONFIG_AV1_HIGHBITDEPTH
@@ -2258,12 +2233,9 @@
   make_tuple(32, 16, &aom_sad_skip_32x16_neon_dotprod, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32_neon_dotprod, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16_neon_dotprod, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8_neon_dotprod, -1),
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16_neon_dotprod, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8_neon_dotprod, -1),
   make_tuple(16, 64, &aom_sad_skip_16x64_neon_dotprod, -1),
-  make_tuple(16, 4, &aom_sad_skip_16x4_neon_dotprod, -1),
 #endif  // !CONFIG_REALTIME_ONLY
 };
 INSTANTIATE_TEST_SUITE_P(NEON_DOTPROD, SADSkipTest,
@@ -2369,12 +2341,9 @@
   make_tuple(32, 16, &aom_sad_skip_32x16x4d_neon_dotprod, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32x4d_neon_dotprod, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16x4d_neon_dotprod, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8x4d_neon_dotprod, -1),
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16x4d_neon_dotprod, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8x4d_neon_dotprod, -1),
   make_tuple(16, 64, &aom_sad_skip_16x64x4d_neon_dotprod, -1),
-  make_tuple(16, 4, &aom_sad_skip_16x4x4d_neon_dotprod, -1),
 #endif  // !CONFIG_REALTIME_ONLY
 };
 INSTANTIATE_TEST_SUITE_P(NEON_DOTPROD, SADSkipx4Test,
@@ -2488,14 +2457,10 @@
   make_tuple(32, 16, &aom_sad_skip_32x16_sse2, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32_sse2, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16_sse2, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8_sse2, -1),
   make_tuple(8, 16, &aom_sad_skip_8x16_sse2, -1),
-  make_tuple(8, 8, &aom_sad_skip_8x8_sse2, -1),
-  make_tuple(4, 8, &aom_sad_skip_4x8_sse2, -1),
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16_sse2, -1),
   make_tuple(16, 64, &aom_sad_skip_16x64_sse2, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8_sse2, -1),
   make_tuple(8, 32, &aom_sad_skip_8x32_sse2, -1),
   make_tuple(4, 16, &aom_sad_skip_4x16_sse2, -1),
 #endif
@@ -2746,14 +2711,10 @@
   make_tuple(32, 16, &aom_sad_skip_32x16x4d_sse2, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32x4d_sse2, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16x4d_sse2, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8x4d_sse2, -1),
   make_tuple(8, 16, &aom_sad_skip_8x16x4d_sse2, -1),
-  make_tuple(8, 8, &aom_sad_skip_8x8x4d_sse2, -1),
-  make_tuple(4, 8, &aom_sad_skip_4x8x4d_sse2, -1),
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16x4d_sse2, -1),
   make_tuple(16, 64, &aom_sad_skip_16x64x4d_sse2, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8x4d_sse2, -1),
   make_tuple(8, 32, &aom_sad_skip_8x32x4d_sse2, -1),
   make_tuple(4, 16, &aom_sad_skip_4x16x4d_sse2, -1),
 #endif
@@ -3032,7 +2993,6 @@
   make_tuple(32, 16, &aom_sad_skip_32x16x4d_avx2, -1),
   make_tuple(16, 32, &aom_sad_skip_16x32x4d_avx2, -1),
   make_tuple(16, 16, &aom_sad_skip_16x16x4d_avx2, -1),
-  make_tuple(16, 8, &aom_sad_skip_16x8x4d_avx2, -1),
 
 #if CONFIG_AV1_HIGHBITDEPTH
   make_tuple(128, 128, &aom_highbd_sad_skip_128x128x4d_avx2, 8),
@@ -3082,10 +3042,8 @@
 
 #if !CONFIG_REALTIME_ONLY
   make_tuple(64, 16, &aom_sad_skip_64x16x4d_avx2, -1),
-  make_tuple(32, 8, &aom_sad_skip_32x8x4d_avx2, -1),
 
   make_tuple(16, 64, &aom_sad_skip_16x64x4d_avx2, -1),
-  make_tuple(16, 4, &aom_sad_skip_16x4x4d_avx2, -1),
 #endif
 };