diff --git a/source/compare_neon64.cc b/source/compare_neon64.cc
index fdd9e0f87..2c0b68b2d 100644
--- a/source/compare_neon64.cc
+++ b/source/compare_neon64.cc
@@ -24,24 +24,22 @@ extern "C" {
 uint32 HammingDistance_NEON(const uint8* src_a, const uint8* src_b, int count) {
   uint32 diff;
   asm volatile (
-    "movi       d4, #0                         \n"
+    "movi       v4.4s, #0                      \n"
 
   "1:                                          \n"
-    MEMACCESS(0)
     "ld1        {v0.16b, v1.16b}, [%0], #32    \n"
-    MEMACCESS(1)
     "ld1        {v2.16b, v3.16b}, [%1], #32    \n"
-    "subs       %w2, %w2, #32                  \n"
     "eor        v0.16b, v0.16b, v2.16b         \n"
-    "eor        v1.16b, v1.16b, v3.16b         \n"
+    "eor        v1.16b, v1.16b, v3.16b         \n" 
     "cnt        v0.16b, v0.16b                 \n"
     "cnt        v1.16b, v1.16b                 \n"
-    "uaddlv     h0, v0.16b                     \n"
-    "uaddlv     h1, v1.16b                     \n"
-    "add        d4, d4, d0                     \n"
-    "add        d4, d4, d1                     \n"
+    "subs       %w2, %w2, #32                  \n"
+    "add        v0.16b, v0.16b, v1.16b         \n"
+    "uaddlp     v0.8h, v0.16b                  \n"
+    "uadalp     v4.4s, v0.8h                   \n"
     "b.gt       1b                             \n"
 
+    "addv       s4, v4.4s                      \n"
     "fmov       %w3, s4                        \n"
     : "+r"(src_a),
       "+r"(src_b),