diff --git a/arch/x86_64.S b/arch/x86_64.S
index d8f805e..927069c 100644
--- a/arch/x86_64.S
+++ b/arch/x86_64.S
@@ -3,6 +3,7 @@
 .globl hw_bswap64
 .globl _cu_memcpy
 .globl _cu_memmove
+.globl _cu_memcmp
 .globl _cu_memset
 .globl cu_memzero
 .globl cu_memtest
@@ -60,6 +61,30 @@ _cu_memmove:
 .retmove:
     ret
 
+_cu_memcmp:
+    mov %rdx, %rcx
+    shr $3, %rcx
+    jz .test1
+.cmp8:
+    cmpsq
+    jnz .not_equal
+    loop .cmp8
+.test1:
+    mov %rdx, %rcx
+    and $7, %rcx
+.cmp1:
+    cmpsb
+    jnz .not_equal
+    loop .cmp1
+    jmp .equal
+.not_equal:
+    xor %rax, %rax
+    inc %rax
+    ret
+.equal:
+    xor %rax, %rax
+    ret
+
 _cu_memset:
     // RDI - ptr, RSI - val, RDX - size
     mov %rdx, %rax
diff --git a/cutil.c b/cutil.c
index c79c133..49696ad 100644
--- a/cutil.c
+++ b/cutil.c
@@ -8,6 +8,7 @@ void (*cu_free)(void*) = NULL;
 void* (*cu_memset)(void*,int,size_t) = _cu_memset;
 void* (*cu_memcpy)(void*,const void*,size_t) = _cu_memcpy;
 void* (*cu_memmove)(void*,const void*,size_t) = _cu_memmove;
+void* (*cu_memcmp)(const void*,const void*, size_t) = _cu_memcmp;
 #else
 #   include <stdlib.h>
 #   include <string.h>
@@ -18,6 +19,7 @@ void (*cu_free)(void*) = free;
 void* (*cu_memset)(void*,int,size_t) = memset;
 void* (*cu_memcpy)(void*,const void*,size_t) = memcpy;
 void* (*cu_memmove)(void*,const void*,size_t) = memmove;
+void* (*cu_memcmp)(const void*,const void*, size_t) = memcmp;
 #endif
 
 void cutil_init()
diff --git a/cutil.h b/cutil.h
index 2a9ec8e..0620343 100644
--- a/cutil.h
+++ b/cutil.h
@@ -7,6 +7,8 @@
 extern void* _cu_memset(void* dst, int val, size_t size);
 extern void* _cu_memcpy(void* dst, const void* src, size_t size);
 extern void* _cu_memmove(void* dst, const void* src, size_t size);
+extern void* _cu_memcmp(const void* dst, const void* src, size_t size);
+
 void __cu_memmove(void* dst, void* src, size_t size);
 
 extern void* (*cu_malloc)(size_t);
@@ -15,6 +17,7 @@ extern void (*cu_free)(void*);
 extern void* (*cu_memset)(void*,int,size_t);
 extern void* (*cu_memcpy)(void*,const void*,size_t);
 extern void* (*cu_memmove)(void*,const void*,size_t);
+extern void* (*cu_memcmp)(const void*,const void*, size_t);
 
 void cutil_init();
 void cutil_exit();