diff --git a/sm4/gcm_arm64.s b/sm4/gcm_arm64.s
index 7ef9d40..80fa535 100644
--- a/sm4/gcm_arm64.s
+++ b/sm4/gcm_arm64.s
@@ -68,16 +68,6 @@ GLOBL r24_mask<>(SB), (NOPTR+RODATA), $16
 #define K1 V20
 #define K2 V21
 #define K3 V22
-#define K4 V23
-#define K5 V24
-#define K6 V25
-#define K7 V26
-#define K8 V27
-#define K9 V28
-#define K10 V29
-#define K11 V30
-#define KLAST V31
-
 #define NIBBLE_MASK V23
 #define INVERSE_SHIFT_ROWS V24
 #define M1L V25
@@ -577,21 +567,13 @@ encOctetsLoop:
 		// Prepare 8 counters
 		VMOV	CTR.B16, B0.B16
 		VADD	B0.S4, INC.S4, B1.S4
-		VREV32	B0.B16, B0.B16
 		VADD	B1.S4, INC.S4, B2.S4
-		VREV32	B1.B16, B1.B16
 		VADD	B2.S4, INC.S4, B3.S4
-		VREV32	B2.B16, B2.B16
 		VADD	B3.S4, INC.S4, B4.S4
-		VREV32	B3.B16, B3.B16
 		VADD	B4.S4, INC.S4, B5.S4
-		VREV32	B4.B16, B4.B16
 		VADD	B5.S4, INC.S4, B6.S4
-		VREV32	B5.B16, B5.B16
 		VADD	B6.S4, INC.S4, B7.S4
-		VREV32	B6.B16, B6.B16
 		VADD	B7.S4, INC.S4, CTR.S4
-		VREV32	B7.B16, B7.B16
 
 		// encryption first 4 blocks
 		PRE_TRANSPOSE_MATRIX(B0, B1, B2, B3)
@@ -684,13 +666,9 @@ encNibblesLoop:
 	// Prepare 4 counters
 	VMOV	CTR.B16, B0.B16
 	VADD	B0.S4, INC.S4, B1.S4
-	VREV32	B0.B16, B0.B16
 	VADD	B1.S4, INC.S4, B2.S4
-	VREV32	B1.B16, B1.B16
 	VADD	B2.S4, INC.S4, B3.S4
-	VREV32	B2.B16, B2.B16
 	VADD	B3.S4, INC.S4, CTR.S4
-	VREV32	B3.B16, B3.B16
 
 	// encryption first 4 blocks
 	PRE_TRANSPOSE_MATRIX(B0, B1, B2, B3)
@@ -733,13 +711,9 @@ encStartSingles:
 	// Prepare 4 counters
 	VMOV	CTR.B16, B0.B16
 	VADD	B0.S4, INC.S4, B1.S4
-	//VREV32	B0.B16, B0.B16
 	VADD	B1.S4, INC.S4, B2.S4
-	//VREV32	B1.B16, B1.B16
 	VADD	B2.S4, INC.S4, B3.S4
-	//VREV32	B2.B16, B2.B16
 	VADD	B3.S4, INC.S4, CTR.S4
-	//VREV32	B3.B16, B3.B16
 
 	// encryption first 4 blocks
 	PRE_TRANSPOSE_MATRIX(B0, B1, B2, B3)
@@ -885,21 +859,13 @@ decOctetsLoop:
 
 		VMOV	CTR.B16, B0.B16
 		VADD	B0.S4, INC.S4, B1.S4
-		VREV32	B0.B16, B0.B16
 		VADD	B1.S4, INC.S4, B2.S4
-		VREV32	B1.B16, B1.B16
 		VADD	B2.S4, INC.S4, B3.S4
-		VREV32	B2.B16, B2.B16
 		VADD	B3.S4, INC.S4, B4.S4
-		VREV32	B3.B16, B3.B16
 		VADD	B4.S4, INC.S4, B5.S4
-		VREV32	B4.B16, B4.B16
 		VADD	B5.S4, INC.S4, B6.S4
-		VREV32	B5.B16, B5.B16
 		VADD	B6.S4, INC.S4, B7.S4
-		VREV32	B6.B16, B6.B16
 		VADD	B7.S4, INC.S4, CTR.S4
-		VREV32	B7.B16, B7.B16
 
 		// encryption first 4 blocks
 		PRE_TRANSPOSE_MATRIX(B0, B1, B2, B3)
@@ -994,13 +960,9 @@ decNibblesLoop:
 	// Prepare 4 counters
 	VMOV	CTR.B16, B0.B16
 	VADD	B0.S4, INC.S4, B1.S4
-	VREV32	B0.B16, B0.B16
 	VADD	B1.S4, INC.S4, B2.S4
-	VREV32	B1.B16, B1.B16
 	VADD	B2.S4, INC.S4, B3.S4
-	VREV32	B2.B16, B2.B16
 	VADD	B3.S4, INC.S4, CTR.S4
-	VREV32	B3.B16, B3.B16
 
 	// encryption first 4 blocks
 	PRE_TRANSPOSE_MATRIX(B0, B1, B2, B3)
@@ -1046,13 +1008,9 @@ decStartSingles:
 	// Prepare 4 counters
 	VMOV	CTR.B16, B0.B16
 	VADD	B0.S4, INC.S4, B1.S4
-	VREV32	B0.B16, B0.B16
 	VADD	B1.S4, INC.S4, B2.S4
-	VREV32	B1.B16, B1.B16
 	VADD	B2.S4, INC.S4, B3.S4
-	VREV32	B2.B16, B2.B16
 	VADD	B3.S4, INC.S4, CTR.S4
-	VREV32	B3.B16, B3.B16
 
 	// encryption first 4 blocks
 	PRE_TRANSPOSE_MATRIX(B0, B1, B2, B3)