diff --git a/src/bytes/bytes.go b/src/bytes/bytes.go
index 305c85d9f41..c35a1c0005f 100644
--- a/src/bytes/bytes.go
+++ b/src/bytes/bytes.go
@@ -93,37 +93,6 @@ func ContainsRune(b []byte, r rune) bool {
 	return IndexRune(b, r) >= 0
 }
 
-// Index returns the index of the first instance of sep in s, or -1 if sep is not present in s.
-func Index(s, sep []byte) int {
-	n := len(sep)
-	if n == 0 {
-		return 0
-	}
-	if n > len(s) {
-		return -1
-	}
-	c := sep[0]
-	if n == 1 {
-		return IndexByte(s, c)
-	}
-	i := 0
-	t := s[:len(s)-n+1]
-	for i < len(t) {
-		if t[i] != c {
-			o := IndexByte(t[i:], c)
-			if o < 0 {
-				break
-			}
-			i += o
-		}
-		if Equal(s[i:i+n], sep) {
-			return i
-		}
-		i++
-	}
-	return -1
-}
-
 func indexBytePortable(s []byte, c byte) int {
 	for i, b := range s {
 		if b == c {
diff --git a/src/bytes/bytes_amd64.go b/src/bytes/bytes_amd64.go
new file mode 100644
index 00000000000..e8be28b51d1
--- /dev/null
+++ b/src/bytes/bytes_amd64.go
@@ -0,0 +1,69 @@
+// Copyright 2016 The Go Authors.  All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package bytes
+
+// indexShortStr returns the index of the first instance of c in s, or -1 if c is not present in s.
+// indexShortStr requires 2 <= len(c) <= shortStringLen
+func indexShortStr(s, c []byte) int // ../runtime/asm_$GOARCH.s
+const shortStringLen = 31
+
+// Index returns the index of the first instance of sep in s, or -1 if sep is not present in s.
+func Index(s, sep []byte) int {
+	n := len(sep)
+	switch {
+	case n == 0:
+		return 0
+	case n == 1:
+		return IndexByte(s, sep[0])
+	case n <= shortStringLen:
+		return indexShortStr(s, sep)
+	case n == len(s):
+		if Equal(sep, s) {
+			return 0
+		}
+		return -1
+	case n > len(s):
+		return -1
+	}
+	// Rabin-Karp search
+	hashsep, pow := hashStr(sep)
+	var h uint32
+	for i := 0; i < n; i++ {
+		h = h*primeRK + uint32(s[i])
+	}
+	if h == hashsep && Equal(s[:n], sep) {
+		return 0
+	}
+	for i := n; i < len(s); {
+		h *= primeRK
+		h += uint32(s[i])
+		h -= pow * uint32(s[i-n])
+		i++
+		if h == hashsep && Equal(s[i-n:i], sep) {
+			return i - n
+		}
+	}
+	return -1
+}
+
+// primeRK is the prime base used in Rabin-Karp algorithm.
+const primeRK = 16777619
+
+// hashStr returns the hash and the appropriate multiplicative
+// factor for use in Rabin-Karp algorithm.
+func hashStr(sep []byte) (uint32, uint32) {
+	hash := uint32(0)
+	for i := 0; i < len(sep); i++ {
+		hash = hash*primeRK + uint32(sep[i])
+	}
+	var pow, sq uint32 = 1, primeRK
+	for i := len(sep); i > 0; i >>= 1 {
+		if i&1 != 0 {
+			pow *= sq
+		}
+		sq *= sq
+	}
+	return hash, pow
+}
diff --git a/src/bytes/bytes_generic.go b/src/bytes/bytes_generic.go
new file mode 100644
index 00000000000..88e232eccf7
--- /dev/null
+++ b/src/bytes/bytes_generic.go
@@ -0,0 +1,41 @@
+// Copyright 2015 The Go Authors.  All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build !amd64
+
+package bytes
+
+// TODO: implements short string optimization on non amd64 platforms
+// and get rid of bytes_amd64.go
+
+// Index returns the index of the first instance of sep in s, or -1 if sep is not present in s.
+func Index(s, sep []byte) int {
+	n := len(sep)
+	if n == 0 {
+		return 0
+	}
+	if n > len(s) {
+		return -1
+	}
+	c := sep[0]
+	if n == 1 {
+		return IndexByte(s, c)
+	}
+	i := 0
+	t := s[:len(s)-n+1]
+	for i < len(t) {
+		if t[i] != c {
+			o := IndexByte(t[i:], c)
+			if o < 0 {
+				break
+			}
+			i += o
+		}
+		if Equal(s[i:i+n], sep) {
+			return i
+		}
+		i++
+	}
+	return -1
+}
diff --git a/src/runtime/asm_amd64.s b/src/runtime/asm_amd64.s
index f44fc1166a4..c9d6b90d800 100644
--- a/src/runtime/asm_amd64.s
+++ b/src/runtime/asm_amd64.s
@@ -1695,13 +1695,31 @@ big_loop_avx2_exit:
 	JMP loop
 
 
-// TODO: Also use this in bytes.Index
 TEXT strings·indexShortStr(SB),NOSPLIT,$0-40
 	MOVQ s+0(FP), DI
 	// We want len in DX and AX, because PCMPESTRI implicitly consumes them
 	MOVQ s_len+8(FP), DX
 	MOVQ c+16(FP), BP
 	MOVQ c_len+24(FP), AX
+	MOVQ DI, R10
+	LEAQ ret+32(FP), R11
+	JMP  runtime·indexShortStr(SB)
+
+TEXT bytes·indexShortStr(SB),NOSPLIT,$0-56
+	MOVQ s+0(FP), DI
+	MOVQ s_len+8(FP), DX
+	MOVQ c+24(FP), BP
+	MOVQ c_len+32(FP), AX
+	MOVQ DI, R10
+	LEAQ ret+48(FP), R11
+	JMP  runtime·indexShortStr(SB)
+
+// AX: length of string, that we are searching for
+// DX: length of string, in which we are searching
+// DI: pointer to string, in which we are searching
+// BP: pointer to string, that we are searching for
+// R11: address, where to put return value
+TEXT runtime·indexShortStr(SB),NOSPLIT,$0
 	CMPQ AX, DX
 	JA fail
 	CMPQ DX, $16
@@ -1853,7 +1871,7 @@ partial_success17to31:
 	CMPQ DI,DX
 	JB loop17to31
 fail:
-	MOVQ $-1, ret+32(FP)
+	MOVQ $-1, (R11)
 	RET
 sse42:
 	MOVL runtime·cpuid_ecx(SB), CX
@@ -1893,8 +1911,8 @@ loop_sse42:
 sse42_success:
 	ADDQ CX, DI
 success:
-	SUBQ s+0(FP), DI
-	MOVQ DI, ret+32(FP)
+	SUBQ R10, DI
+	MOVQ DI, (R11)
 	RET