diff --git a/tz/avx2_inline_amd64.s b/tz/avx2_amd64.s
similarity index 62%
rename from tz/avx2_inline_amd64.s
rename to tz/avx2_amd64.s
index fb7f83a..200163c 100644
--- a/tz/avx2_inline_amd64.s
+++ b/tz/avx2_amd64.s
@@ -44,3 +44,29 @@ TEXT ·mulByteRightx2(SB),NOSPLIT,$0
     VMOVDQA Y0, (AX)
 
     RET
+
+// func mulBitRightx2(c00c10, c01c11 *[4]uint64, e *[2]uint64)
+TEXT ·mulBitRightx2(SB),NOSPLIT,$0
+    MOVQ c00c10+0(FP), AX
+    VMOVDQA (AX), Y0
+    MOVQ c01c11+8(FP), BX
+    VMOVDQA (BX), Y8
+
+    VPSLLQ $1, Y0, Y1
+    VPALIGNR $8, Y1, Y0, Y2
+    VPSRLQ $63, Y2, Y2
+    VPXOR Y1, Y2, Y2
+    VPSRLQ $63, Y1, Y3
+    VPSLLQ $63, Y3, Y3
+    VPUNPCKHQDQ Y3, Y3, Y3
+    VPXOR Y2, Y3, Y3
+
+    MOVQ e+16(FP), CX
+    VBROADCASTI128 (CX), Y2
+
+    VPXOR Y3, Y8, Y3
+    VPAND Y3, Y2, Y4
+    VPXOR Y4, Y0, Y8
+    VMOVDQA Y8, (BX)
+    VMOVDQA Y3, (AX)
+    RET
diff --git a/tz/avx_amd64.s b/tz/avx_amd64.s
index a380b3a..d8b5e53 100644
--- a/tz/avx_amd64.s
+++ b/tz/avx_amd64.s
@@ -96,29 +96,3 @@ TEXT ·mulByteRight(SB),NOSPLIT,$0
     MOVUPD X3, (DX)
 
     RET
-
-// func mulBitRightx2(c00c10, c01c11 *[4]uint64, e *[2]uint64)
-TEXT ·mulBitRightx2(SB),NOSPLIT,$0
-    MOVQ c00c10+0(FP), AX
-    VMOVDQA (AX), Y0
-    MOVQ c01c11+8(FP), BX
-    VMOVDQA (BX), Y8
-
-    VPSLLQ $1, Y0, Y1
-    VPALIGNR $8, Y1, Y0, Y2
-    VPSRLQ $63, Y2, Y2
-    VPXOR Y1, Y2, Y2
-    VPSRLQ $63, Y1, Y3
-    VPSLLQ $63, Y3, Y3
-    VPUNPCKHQDQ Y3, Y3, Y3
-    VPXOR Y2, Y3, Y3
-
-    MOVQ e+16(FP), CX
-    VBROADCASTI128 (CX), Y2
-
-    VPXOR Y3, Y8, Y3
-    VPAND Y3, Y2, Y4
-    VPXOR Y4, Y0, Y8
-    VMOVDQA Y8, (BX)
-    VMOVDQA Y3, (AX)
-    RET