diff --git a/README.md b/README.md
index 2222df4..d929f53 100644
--- a/README.md
+++ b/README.md
@@ -13,7 +13,7 @@
 ## Packages
 * **SM2** - This is a SM2 sm2p256v1 implementation whose performance is similar like golang native NIST P256 under **amd64** and **arm64**, for implementation detail, please refer [SM2实现细节](https://github.com/emmansun/gmsm/wiki/SM2%E6%80%A7%E8%83%BD%E4%BC%98%E5%8C%96). It supports ShangMi sm2 digital signature, public key encryption algorithm and also key exchange.
 
-* **SM3** - This is also a SM3 implementation whose performance is similar like golang native SHA 256 with SIMD under **amd64**, for implementation detail, please refer [SM3性能优化](https://github.com/emmansun/gmsm/wiki/SM3%E6%80%A7%E8%83%BD%E4%BC%98%E5%8C%96). It also provides A64 cryptographic instructions SM3 tested with QEMU.
+* **SM3** - This is also a SM3 implementation whose performance is similar like golang native SHA 256 with SIMD under **amd64** and **arm64**, for implementation detail, please refer [SM3性能优化](https://github.com/emmansun/gmsm/wiki/SM3%E6%80%A7%E8%83%BD%E4%BC%98%E5%8C%96). It also provides A64 cryptographic instructions SM3 tested with QEMU.
 
 * **SM4** - For SM4 implementation, SIMD & AES-NI are used under **amd64** and **arm64**, for detail please refer [SM4性能优化](https://github.com/emmansun/gmsm/wiki/SM4%E6%80%A7%E8%83%BD%E4%BC%98%E5%8C%96), it supports ECB/CBC/CFB/OFB/CTR/GCM/CCM/XTS modes. It also provides A64 cryptographic instructions SM4 tested with QEMU.
 
diff --git a/sm3/sm3_test.go b/sm3/sm3_test.go
index 29136ae..782e6b3 100644
--- a/sm3/sm3_test.go
+++ b/sm3/sm3_test.go
@@ -432,8 +432,8 @@ func round3(a, b, c, d, e, f, g, h string, i int) {
 	fmt.Printf("w[%d] = p1(w[%d]^w[%d]^bits.RotateLeft32(w[%d], 15)) ^ bits.RotateLeft32(w[%d], 7) ^ w[%d]\n", i+4, i-12, i-5, i+1, i-9, i-2)
 	fmt.Printf("tt2 = bits.RotateLeft32(%s, 12)\n", a)
 	fmt.Printf("ss1 = bits.RotateLeft32(tt2+%s+_K[%d], 7)\n", e, i)
-	fmt.Printf("%s = %s&(%s|%s) | (%s & %s) + %s + (ss1 ^ tt2) + (w[%d] ^ w[%d])\n", d, c, a, b, a, b, d, i, i+4)
-	fmt.Printf("tt2 = (%s & %s) | (^%s & %s) + %s + ss1 + w[%d]\n", e, f, e, g, h, i)
+	fmt.Printf("%s = %s&(%s|%s) | (%s & %s) + %s + (ss1 ^ tt2) + (w[%d] ^ w[%d])\n", d, a, b, c, b, c, d, i, i+4)
+	fmt.Printf("tt2 = (%s^%s)&%s ^ %s + %s + ss1 + w[%d]\n", f, g, e, g, h, i)
 	fmt.Printf("%s = bits.RotateLeft32(%s, 9)\n", b, b)
 	fmt.Printf("%s = bits.RotateLeft32(%s, 19)\n", f, f)
 	fmt.Printf("%s = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)\n\n", h)
diff --git a/sm3/sm3block.go b/sm3/sm3block.go
index 2a8eb52..c0ea485 100644
--- a/sm3/sm3block.go
+++ b/sm3/sm3block.go
@@ -201,7 +201,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[16], 7)
 		d = c&(a|b) | (a & b) + d + (ss1 ^ tt2) + (w[16] ^ w[20])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[16]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[16]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -211,7 +211,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[17], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[17] ^ w[21])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[17]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[17]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -221,7 +221,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[18], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[18] ^ w[22])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[18]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[18]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -231,7 +231,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[19], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[19] ^ w[23])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[19]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[19]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -241,7 +241,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[20], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[20] ^ w[24])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[20]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[20]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -251,7 +251,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[21], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[21] ^ w[25])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[21]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[21]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -261,7 +261,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[22], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[22] ^ w[26])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[22]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[22]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -271,7 +271,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[23], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[23] ^ w[27])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[23]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[23]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -281,7 +281,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[24], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[24] ^ w[28])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[24]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[24]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -291,7 +291,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[25], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[25] ^ w[29])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[25]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[25]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -301,7 +301,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[26], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[26] ^ w[30])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[26]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[26]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -311,7 +311,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[27], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[27] ^ w[31])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[27]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[27]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -321,7 +321,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[28], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[28] ^ w[32])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[28]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[28]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -331,7 +331,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[29], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[29] ^ w[33])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[29]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[29]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -341,7 +341,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[30], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[30] ^ w[34])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[30]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[30]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -351,7 +351,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[31], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[31] ^ w[35])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[31]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[31]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -361,7 +361,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[32], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[32] ^ w[36])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[32]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[32]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -371,7 +371,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[33], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[33] ^ w[37])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[33]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[33]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -381,7 +381,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[34], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[34] ^ w[38])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[34]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[34]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -391,7 +391,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[35], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[35] ^ w[39])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[35]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[35]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -401,7 +401,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[36], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[36] ^ w[40])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[36]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[36]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -411,7 +411,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[37], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[37] ^ w[41])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[37]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[37]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -421,7 +421,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[38], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[38] ^ w[42])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[38]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[38]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -431,7 +431,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[39], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[39] ^ w[43])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[39]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[39]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -441,16 +441,17 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[40], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[40] ^ w[44])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[40]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[40]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
+
 		// Round 42
 		w[45] = p1(w[29]^w[36]^bits.RotateLeft32(w[42], 15)) ^ bits.RotateLeft32(w[32], 7) ^ w[39]
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[41], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[41] ^ w[45])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[41]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[41]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -460,7 +461,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[42], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[42] ^ w[46])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[42]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[42]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -470,7 +471,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[43], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[43] ^ w[47])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[43]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[43]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -480,7 +481,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[44], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[44] ^ w[48])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[44]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[44]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -490,7 +491,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[45], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[45] ^ w[49])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[45]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[45]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -500,7 +501,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[46], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[46] ^ w[50])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[46]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[46]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -510,7 +511,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[47], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[47] ^ w[51])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[47]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[47]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -520,7 +521,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[48], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[48] ^ w[52])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[48]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[48]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -530,7 +531,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[49], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[49] ^ w[53])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[49]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[49]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -540,7 +541,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[50], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[50] ^ w[54])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[50]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[50]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -550,7 +551,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[51], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[51] ^ w[55])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[51]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[51]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -560,7 +561,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[52], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[52] ^ w[56])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[52]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[52]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -570,7 +571,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[53], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[53] ^ w[57])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[53]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[53]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -580,7 +581,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[54], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[54] ^ w[58])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[54]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[54]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -590,7 +591,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[55], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[55] ^ w[59])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[55]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[55]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -600,7 +601,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[56], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[56] ^ w[60])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[56]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[56]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -610,7 +611,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[57], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[57] ^ w[61])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[57]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[57]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -620,7 +621,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[58], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[58] ^ w[62])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[58]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[58]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -630,7 +631,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[59], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[59] ^ w[63])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[59]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[59]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -640,7 +641,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(a, 12)
 		ss1 = bits.RotateLeft32(tt2+e+_K[60], 7)
 		d = a&(b|c) | (b & c) + d + (ss1 ^ tt2) + (w[60] ^ w[64])
-		tt2 = (e & f) | (^e & g) + h + ss1 + w[60]
+		tt2 = (f^g)&e ^ g + h + ss1 + w[60]
 		b = bits.RotateLeft32(b, 9)
 		f = bits.RotateLeft32(f, 19)
 		h = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -650,7 +651,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(d, 12)
 		ss1 = bits.RotateLeft32(tt2+h+_K[61], 7)
 		c = d&(a|b) | (a & b) + c + (ss1 ^ tt2) + (w[61] ^ w[65])
-		tt2 = (h & e) | (^h & f) + g + ss1 + w[61]
+		tt2 = (e^f)&h ^ f + g + ss1 + w[61]
 		a = bits.RotateLeft32(a, 9)
 		e = bits.RotateLeft32(e, 19)
 		g = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -660,7 +661,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(c, 12)
 		ss1 = bits.RotateLeft32(tt2+g+_K[62], 7)
 		b = c&(d|a) | (d & a) + b + (ss1 ^ tt2) + (w[62] ^ w[66])
-		tt2 = (g & h) | (^g & e) + f + ss1 + w[62]
+		tt2 = (h^e)&g ^ e + f + ss1 + w[62]
 		d = bits.RotateLeft32(d, 9)
 		h = bits.RotateLeft32(h, 19)
 		f = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
@@ -670,7 +671,7 @@ func blockGeneric(dig *digest, p []byte) {
 		tt2 = bits.RotateLeft32(b, 12)
 		ss1 = bits.RotateLeft32(tt2+f+_K[63], 7)
 		a = b&(c|d) | (c & d) + a + (ss1 ^ tt2) + (w[63] ^ w[67])
-		tt2 = (f & g) | (^f & h) + e + ss1 + w[63]
+		tt2 = (g^h)&f ^ h + e + ss1 + w[63]
 		c = bits.RotateLeft32(c, 9)
 		g = bits.RotateLeft32(g, 19)
 		e = tt2 ^ bits.RotateLeft32(tt2, 9) ^ bits.RotateLeft32(tt2, 17)
diff --git a/sm3/sm3block_amd64.s b/sm3/sm3block_amd64.s
index 72ffa9b..b21bfb8 100644
--- a/sm3/sm3block_amd64.s
+++ b/sm3/sm3block_amd64.s
@@ -115,10 +115,10 @@
 	MOVL ((index)*4)(BP), DX; \
 	ADDL h, DX; \   // Wt + h
 	ADDL BX, DX; \  // h + ss1 + Wt
-	MOVL f, AX; \   
-	ANDL e, AX; \  // e AND f
-	ANDNL g, e, BX; \ // NOT(e) AND g
-	ORL  AX, BX; \
+	MOVL f, BX; \   
+	XORL g, BX; \
+	ANDL e, BX; \
+	XORL g, BX; \ // GG2(e, f, g)
 	ADDL DX, BX
 
 #define COPYRESULT(b, d, f, h) \
diff --git a/sm3/sm3block_arm64.s b/sm3/sm3block_arm64.s
index 2f394c2..9ee2707 100644
--- a/sm3/sm3block_arm64.s
+++ b/sm3/sm3block_arm64.s
@@ -30,11 +30,10 @@
 #define y0 R8
 #define y1 R9
 #define y2 R10
-#define y3 R11
 
-#define NUM_BYTES R12
-#define INP	R13
-#define CTX R14 // Beginning of digest in memory (a, b, c, ... , h)
+#define NUM_BYTES R11
+#define INP	R12
+#define CTX R13 // Beginning of digest in memory (a, b, c, ... , h)
 
 #define a1 R15
 #define b1 R16
@@ -210,10 +209,10 @@
 	VEOR XTMP1.B16, XTMP0.B16, XTMP0.B16;         \ // XTMP0 = W[-6] ^ (W[-13] rol 7)
 	ADDW  y0, h;                                  \ // h = FF(a, b, c) + d + SS2 + W' = tt1
 	; \
-	ANDW  e, f, y1;                               \
-	BICW  e, g, y3;                               \
+	EORW  f, g, y1;                               \
+	ANDW  e, y1;                                  \	
 	VEXT $12, XWORD2.B16, XWORD1.B16, XTMP1.B16;  \ // XTMP1 = W[-9] = {w10,w9,w8,w7}, Vm = XWORD2, Vn = XWORD1
-	ORRW  y3, y1;                                 \ // y1 = (e AND f) OR (NOT(e) AND g)
+	EORW  g, y1;                                  \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)	
 	ADDW  y1, y2;                                 \ // y2 = GG(e, f, g) + h + SS1 + W = tt2 
 	; \
 	RORW  $23, b;                                 \
@@ -250,9 +249,9 @@
 	VSRI $17, XTMP2.S4, XTMP4.S4;                 \ // XTMP4 =  = XTMP2 rol 15 {xxBA}
 	ADDW  y0, h;                                  \ // h = FF(a, b, c) + d + SS2 + W' = tt1
 	; \
-	ANDW  e, f, y1;                               \
-	BICW  e, g, y3;                               \
-	ORRW  y3, y1;                                 \ // y1 = (e AND f) OR (NOT(e) AND g)
+	EORW  f, g, y1;                               \
+	ANDW  e, y1;                                  \	
+	EORW  g, y1;                                  \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)	
 	VSHL $8, XTMP4.S4, XTMP3.S4;                  \
 	ADDW  y1, y2;                                 \ // y2 = GG(e, f, g) + h + SS1 + W = tt2 
 	; \
@@ -287,10 +286,10 @@
 	VSHL $15, XTMP3.S4, XTMP4.S4;                 \
 	ORRW  y1, h;                                  \ // h =  (a AND b) OR (a AND c) OR (b AND c)	
 	ADDW  y0, h;                                  \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	ANDW  e, f, y1;                               \
-	BICW  e, g, y3;                               \
+	EORW  f, g, y1;                               \
+	ANDW  e, y1;                                  \	
 	VSRI $17, XTMP3.S4, XTMP4.S4;                 \ // XTMP4 = W[-3] rol 15 {DCBA}
-	ORRW  y3, y1;                                 \ // y1 = (e AND f) OR (NOT(e) AND g)
+	EORW  g, y1;                                  \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)	
 	ADDW  y1, y2;                                 \ // y2 = GG(e, f, g) + h + SS1 + W = tt2 
 	RORW  $23, b;                                 \
 	RORW  $13, f;                                 \
@@ -321,10 +320,10 @@
 	VSRI $24, XTMP3.S4, XTMP1.S4;                 \ // XTMP1 = XTMP4 rol 23 {DCBA}
 	ORRW  y1, h;                                  \ // h =  (a AND b) OR (a AND c) OR (b AND c)
 	ADDW  y0, h;                                  \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	ANDW  e, f, y1;                               \
-	BICW  e, g, y3;                               \
+	EORW  f, g, y1;                               \
+	ANDW  e, y1;                                  \
 	VEOR XTMP3.B16, XTMP4.B16, XTMP3.B16;         \ // XTMP3 = XTMP4 XOR (XTMP4 rol 15 {DCBA})
-	ORRW  y3, y1;                                 \ // y1 = (e AND f) OR (NOT(e) AND g)
+	EORW  g, y1;                                  \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
 	ADDW  y1, y2;                                 \ // y2 = GG(e, f, g) + h + SS1 + W = tt2 
 	RORW  $23, b;                                 \
 	RORW  $13, f;                                 \
@@ -355,10 +354,10 @@
 	ORRW  y1, h;                               \ // h =  (a AND b) OR (a AND c) OR (b AND c)
 	ADDW  y0, h;                               \ // h = FF(a, b, c) + d + SS2 + W' = tt1
 	; \
-	ANDW  e, f, y1;                            \
-	BICW  e, g, y3;                            \
-	ORRW  y3, y1;                              \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDW  y1, y2;                              \ // y2 = GG(e, f, g) + h + SS1 + W = tt2 
+	EORW  f, g, y1;                            \
+	ANDW  e, y1;                               \
+	EORW  g, y1;                               \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDW  y1, y2;                              \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2 
 	; \
 	RORW  $23, b;                              \
 	RORW  $13, f;                              \
diff --git a/sm3/sm3block_avx2_amd64.s b/sm3/sm3block_avx2_amd64.s
index 3ace21a..d892d1d 100644
--- a/sm3/sm3block_avx2_amd64.s
+++ b/sm3/sm3block_avx2_amd64.s
@@ -52,17 +52,14 @@
 #define y0 R12
 #define y1 R13
 #define y2 R14
-#define y3 DI
 
 // Offsets
 #define XFER_SIZE 4*64*4
 #define INP_END_SIZE 8
-#define INP_SIZE 8
 
 #define _XFER 0
 #define _INP_END _XFER + XFER_SIZE
-#define _INP _INP_END + INP_END_SIZE
-#define STACK_SIZE _INP + INP_SIZE
+#define STACK_SIZE _INP_END + INP_END_SIZE
 
 #define P0(tt2, tmp, out) \
 	RORXL    $23, tt2, tmp;                        \
@@ -221,12 +218,12 @@
 	ORL      y1, h;                            \ // h =  (a AND b) OR (a AND c) OR (b AND c)  
 	VPALIGNR $8, XDWORD2, XDWORD3, XTMP0;      \ // XTMP0 = W[-6] = {w13,w12,w11,w10}
 	ADDL     y0, h;                            \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	MOVL     f, y3;                            \
-	ANDL     e, y3;                            \ // y3 = e AND f
-	ANDNL    g, e, y1;                         \ // y1 = NOT(e) AND g
+	MOVL     f, y1;                            \
+	XORL     g, y1;                            \
+	ANDL     e, y1;                            \
 	VPXOR   XTMP1, XTMP0, XTMP0;               \ // XTMP0 = W[-6] ^ (W[-13] rol 7) 
-	ORL      y3, y1;                           \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                           \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	XORL     g, y1;                            \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDL     y1, y2;                           \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2  	 
 	ROLL     $9, b;                            \
 	ROLL     $19, f;                           \
 	VPALIGNR $12, XDWORD1, XDWORD2, XTMP1;     \ // XTMP1 = W[-9] = {w10,w9,w8,w7}
@@ -256,12 +253,12 @@
 	ORL      y1, h;                            \ // h =  (a AND b) OR (a AND c) OR (b AND c)     
 	VPSHUFD $0x00, XTMP2, XTMP2;               \ // XTMP2 = {AAAA}
 	ADDL     y0, h;                            \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	MOVL     f, y3;                            \
-	ANDL     e, y3;                            \ // y3 = e AND f
-	ANDNL    g, e, y1;                         \ // y1 = NOT(e) AND g
+	MOVL     f, y1;                            \
+	XORL     g, y1;                            \
+	ANDL     e, y1;                            \
 	VPSRLQ  $17, XTMP2, XTMP3;                 \ // XTMP3 = XTMP2 rol 15 {xxxA}
-	ORL      y3, y1;                           \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                           \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	XORL     g, y1;                            \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDL     y1, y2;                           \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2  	
 	ROLL     $9, b;                            \
 	ROLL     $19, f;                           \
 	VPSRLQ  $9, XTMP2, XTMP4;                  \ // XTMP4 = XTMP2 rol 23 {xxxA}
@@ -291,12 +288,12 @@
 	ORL      y1, h;                            \ // h =  (a AND b) OR (a AND c) OR (b AND c)     
 	VPSLLD   $15, XTMP3, XTMP4;                \
 	ADDL     y0, h;                            \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	MOVL     f, y3;                            \
-	ANDL     e, y3;                            \ // y3 = e AND f
-	ANDNL    g, e, y1;                         \ // y1 = NOT(e) AND g
+	MOVL     f, y1;                            \
+	XORL     g, y1;                            \
+	ANDL     e, y1;                            \
 	VPSRLD   $(32-15), XTMP3, XTMP3;           \
-	ORL      y3, y1;                           \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                           \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	XORL     g, y1;                            \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDL     y1, y2;                           \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2  
 	ROLL     $9, b;                            \
 	ROLL     $19, f;                           \
 	VPOR    XTMP3, XTMP4, XTMP4;               \ // XTMP4 = (W[-3] rol 15) {DCBA}
@@ -326,12 +323,12 @@
 	ORL      y1, h;                            \ // h =  (a AND b) OR (a AND c) OR (b AND c)
 	VPSHUFB  R08_SHUFFLE_MASK, XTMP3, XTMP1;   \ // XTMP1 = XTMP4 rol 23 {DCBA}
 	ADDL     y0, h;                            \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	MOVL     f, y3;                            \
-	ANDL     e, y3;                            \ // y3 = e AND f
-	ANDNL    g, e, y1;                         \ // y1 = NOT(e) AND g
+	MOVL     f, y1;                            \
+	XORL     g, y1;                            \
+	ANDL     e, y1;                            \
 	VPXOR    XTMP3, XTMP4, XTMP3;              \ // XTMP3 = XTMP4 ^ (XTMP4 rol 15 {DCBA})
-	ORL      y3, y1;                           \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                           \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	XORL     g, y1;                            \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDL     y1, y2;                           \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2  
 	ROLL     $9, b;                            \
 	ROLL     $19, f;                           \
 	VPXOR    XTMP3, XTMP1, XTMP1;              \ // XTMP1 = XTMP4 ^ (XTMP4 rol 15 {DCBA}) ^ (XTMP4 rol 23 {DCBA})
@@ -387,18 +384,18 @@
 	ORL      y1, h;                              \ // h =  (a AND b) OR (a AND c) OR (b AND c)     
 	ADDL     y0, h;                              \ // h = FF(a, b, c) + d + SS2 + W' = tt1
 	;                                            \
-	MOVL     f, y3;                              \
-	ANDL     e, y3;                              \ // y3 = e AND f
-	ANDNL    g, e, y1;                           \ // y1 = NOT(e) AND g
-	ORL      y3, y1;                             \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                             \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	MOVL     f, y1;                              \
+	XORL     g, y1;                              \
+	ANDL     e, y1;                              \
+	XORL     g, y1;                              \ // y1 = GG2(e, f, g)
+	ADDL     y1, y2;                             \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2  
 	;                                            \
 	ROLL     $9, b;                              \
 	ROLL     $19, f;                             \
 	;                                            \
 	P0(y2, y0, d)
 
-TEXT ·blockAVX2(SB), 0, $1048-32
+TEXT ·blockAVX2(SB), 0, $1040-32
 	MOVQ dig+0(FP), CTX          // d.h[8]
 	MOVQ p_base+8(FP), INP
 	MOVQ p_len+16(FP), NUM_BYTES
@@ -443,7 +440,6 @@ avx2_loop: // at each iteration works with one block (512 bit)
 
 avx2_last_block_enter:
 	ADDQ $64, INP
-	MOVQ INP, _INP(SP)
 
 avx2_schedule_compress: // for w0 - w47
 	// Do 4 rounds and scheduling
@@ -592,8 +588,6 @@ avx2_schedule_compress: // for w0 - w47
 	DO_ROUND_N_1(_XFER + 30*32, 2, T62, c, d, e, f, g, h, a, b)
 	DO_ROUND_N_1(_XFER + 30*32, 3, T63, b, c, d, e, f, g, h, a)
 
-	MOVQ _INP(SP), INP
-
 	xorm(  0(CTX), a)
 	xorm(  4(CTX), b)
 	xorm(  8(CTX), c)
@@ -687,7 +681,6 @@ avx2_compress: // Do second block using previously scheduled results
 	DO_ROUND_N_1(_XFER + 30*32 + 16, 2, T62, c, d, e, f, g, h, a, b)
 	DO_ROUND_N_1(_XFER + 30*32 + 16, 3, T63, b, c, d, e, f, g, h, a)
 
-	MOVQ _INP(SP), INP
 	ADDQ $64, INP
 
 	xorm(  0(CTX), a)
diff --git a/sm3/sm3block_simd_amd64.s b/sm3/sm3block_simd_amd64.s
index a4f557f..b028d90 100644
--- a/sm3/sm3block_simd_amd64.s
+++ b/sm3/sm3block_simd_amd64.s
@@ -44,17 +44,14 @@
 #define y0 R12
 #define y1 R13
 #define y2 R14
-#define y3 DI
 
 // Offsets
 #define XFER_SIZE 2*16
 #define INP_END_SIZE 8
-#define INP_SIZE 8
 
 #define _XFER 0
 #define _INP_END _XFER + XFER_SIZE
-#define _INP _INP_END + INP_END_SIZE
-#define STACK_SIZE _INP + INP_SIZE
+#define STACK_SIZE _INP_END + INP_END_SIZE
 
 #define SS12(a, e, const, ss1, ss2) \
 	MOVL     a, ss2;                            \
@@ -249,12 +246,12 @@
 	ORL      y1, h;                            \ // h =  (a AND b) OR (a AND c) OR (b AND c)  
 	VPALIGNR $8, XWORD2, XWORD3, XTMP0;        \ // XTMP0 = W[-6] = {w13,w12,w11,w10}
 	ADDL     y0, h;                            \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	MOVL     f, y3;                            \
-	ANDL     e, y3;                            \ // y3 = e AND f
-	ANDNL    g, e, y1;                         \ // y1 = NOT(e) AND g
+	MOVL     f, y1;                            \
+	XORL     g, y1;                            \
+	ANDL     e, y1;                            \
 	VPXOR   XTMP1, XTMP0, XTMP0;               \ // XTMP0 = W[-6] ^ (W[-13] rol 7) 
-	ORL      y3, y1;                           \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                           \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	XORL     g, y1;                            \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDL     y1, y2;                           \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2  	
 	ROLL     $9, b;                            \
 	ROLL     $19, f;                           \
 	VPALIGNR $12, XWORD1, XWORD2, XTMP1;       \ // XTMP1 = W[-9] = {w10,w9,w8,w7}
@@ -285,12 +282,12 @@
 	ORL      y1, h;                            \ // h =  (a AND b) OR (a AND c) OR (b AND c)     
 	VPSHUFD $0x00, XTMP2, XTMP2;               \ // XTMP2 = {AAAA}
 	ADDL     y0, h;                            \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	MOVL     f, y3;                            \
-	ANDL     e, y3;                            \ // y3 = e AND f
-	ANDNL    g, e, y1;                         \ // y1 = NOT(e) AND g
+	MOVL     f, y1;                            \
+	XORL     g, y1;                            \
+	ANDL     e, y1;                            \
 	VPSRLQ  $17, XTMP2, XTMP3;                 \ // XTMP3 = XTMP2 rol 15 {xxxA}
-	ORL      y3, y1;                           \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                           \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	XORL     g, y1;                            \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDL     y1, y2;                           \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2  	
 	ROLL     $9, b;                            \
 	ROLL     $19, f;                           \
 	VPSRLQ  $9, XTMP2, XTMP4;                  \ // XTMP4 = XTMP2 rol 23 {xxxA}
@@ -321,12 +318,12 @@
 	ORL      y1, h;                            \ // h =  (a AND b) OR (a AND c) OR (b AND c)  
 	VPSLLD   $15, XTMP3, XTMP4;                \
 	ADDL     y0, h;                            \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	MOVL     f, y3;                            \
-	ANDL     e, y3;                            \ // y3 = e AND f
-	ANDNL    g, e, y1;                         \ // y1 = NOT(e) AND g
+	MOVL     f, y1;                            \
+	XORL     g, y1;                            \
+	ANDL     e, y1;                            \
 	VPSRLD   $(32-15), XTMP3, XTMP3;           \
-	ORL      y3, y1;                           \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                           \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	XORL     g, y1;                            \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDL     y1, y2;                           \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2 
 	ROLL     $9, b;                            \
 	ROLL     $19, f;                           \
 	VPOR    XTMP3, XTMP4, XTMP4;               \ // XTMP4 = (W[-3] rol 15) {DCBA}
@@ -357,12 +354,12 @@
 	ORL      y1, h;                            \ // h =  (a AND b) OR (a AND c) OR (b AND c)  
 	VPSHUFB  R08_SHUFFLE_MASK, XTMP3, XTMP1;   \ // XTMP1 = XTMP4 rol 23 {DCBA}
 	ADDL     y0, h;                            \ // h = FF(a, b, c) + d + SS2 + W' = tt1
-	MOVL     f, y3;                            \
-	ANDL     e, y3;                            \ // y3 = e AND f
-	ANDNL    g, e, y1;                         \ // y1 = NOT(e) AND g
+	MOVL     f, y1;                            \
+	XORL     g, y1;                            \
+	ANDL     e, y1;                            \
 	VPXOR    XTMP3, XTMP4, XTMP3;              \ // XTMP3 = XTMP4 ^ (XTMP4 rol 15 {DCBA})
-	ORL      y3, y1;                           \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                           \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	XORL     g, y1;                            \ // y1 = GG2(e, f, g) = (e AND f) OR (NOT(e) AND g)
+	ADDL     y1, y2;                           \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2 
 	ROLL     $9, b;                            \
 	ROLL     $19, f;                           \
 	VPXOR    XTMP3, XTMP1, XTMP1;              \ // XTMP1 = XTMP4 ^ (XTMP4 rol 15 {DCBA}) ^ (XTMP4 rol 23 {DCBA})
@@ -410,11 +407,11 @@
 	ORL      y1, h;                              \ // h =  (a AND b) OR (a AND c) OR (b AND c)  
 	ADDL     y0, h;                              \ // h = FF(a, b, c) + d + SS2 + W' = tt1
 	;                                            \
-	MOVL     f, y3;                              \
-	ANDL     e, y3;                              \ // y3 = e AND f
-	ANDNL    g, e, y1;                           \ // y1 = NOT(e) AND g
-	ORL      y3, y1;                             \ // y1 = (e AND f) OR (NOT(e) AND g)
-	ADDL     y1, y2;                             \ // y2 = GG(e, f, g) + h + SS1 + W = tt2  
+	MOVL     f, y1;                              \
+	XORL     g, y1;                              \
+	ANDL     e, y1;                              \
+	XORL     g, y1;                              \ // y1 = GG2(e, f, g)	
+	ADDL     y1, y2;                             \ // y2 = GG2(e, f, g) + h + SS1 + W = tt2  
 	;                                            \
 	ROLL     $9, b;                              \
 	ROLL     $19, f;                             \
@@ -471,7 +468,7 @@
 	MOVOU XTMP0, XWORD0;                     \
 	PXOR XTMP1, XWORD0
 
-TEXT ·blockSIMD(SB), 0, $56-32
+TEXT ·blockSIMD(SB), 0, $48-32
 	MOVQ dig+0(FP), CTX          // d.h[8]
 	MOVQ p_base+8(FP), INP
 	MOVQ p_len+16(FP), NUM_BYTES
@@ -507,7 +504,6 @@ sse_loop: // at each iteration works with one block (512 bit)
 	PSHUFB X_BYTE_FLIP_MASK, XWORD3 // w15, w14, w13, w12
 
 	ADDQ $64, INP
-	MOVQ INP, _INP(SP)
 
 sse_schedule_compress: // for w0 - w47
 	// Do 4 rounds and scheduling
@@ -684,8 +680,6 @@ sse_schedule_compress: // for w0 - w47
 	DO_ROUND_N_1(_XFER, 2, T62, c, d, e, f, g, h, a, b)
 	DO_ROUND_N_1(_XFER, 3, T63, b, c, d, e, f, g, h, a)
 
-	MOVQ _INP(SP), INP
-
 	xorm(  0(CTX), a)
 	xorm(  4(CTX), b)
 	xorm(  8(CTX), c)
@@ -718,7 +712,6 @@ avx_loop: // at each iteration works with one block (512 bit)
 	VPSHUFB X_BYTE_FLIP_MASK, XWORD3, XWORD3 // w15, w14, w13, w12
 
 	ADDQ $64, INP
-	MOVQ INP, _INP(SP)
 
 avx_schedule_compress: // for w0 - w47
 	// Do 4 rounds and scheduling
@@ -868,8 +861,6 @@ avx_schedule_compress: // for w0 - w47
 	DO_ROUND_N_1(_XFER, 2, T62, c, d, e, f, g, h, a, b)
 	DO_ROUND_N_1(_XFER, 3, T63, b, c, d, e, f, g, h, a)
 
-	MOVQ _INP(SP), INP
-
 	xorm(  0(CTX), a)
 	xorm(  4(CTX), b)
 	xorm(  8(CTX), c)