uint256: optimize Mul, squared (#152)

Using four local `uint64` variables enables register allocation instead of memory allocation, which improves the performance of `Mul`, `squared` and `Exp` about 40%

uint256.go

@@ -372,26 +372,26 @@ func umul(x, y *Int) [8]uint64 {
 // Mul sets z to the product x*y
 func (z *Int) Mul(x, y *Int) *Int {
 	var (
-		res              Int
-		carry            uint64
-		res1, res2, res3 uint64
+		carry                  uint64
+		res0, res1, res2, res3 uint64
 	)
 
-	carry, res[0] = bits.Mul64(x[0], y[0])
+	carry, res0 = bits.Mul64(x[0], y[0])
 	carry, res1 = umulHop(carry, x[1], y[0])
 	carry, res2 = umulHop(carry, x[2], y[0])
 	res3 = x[3]*y[0] + carry
 
-	carry, res[1] = umulHop(res1, x[0], y[1])
+	carry, res1 = umulHop(res1, x[0], y[1])
 	carry, res2 = umulStep(res2, x[1], y[1], carry)
 	res3 = res3 + x[2]*y[1] + carry
 
-	carry, res[2] = umulHop(res2, x[0], y[2])
+	carry, res2 = umulHop(res2, x[0], y[2])
 	res3 = res3 + x[1]*y[2] + carry
 
-	res[3] = res3 + x[0]*y[3]
+	res3 = res3 + x[0]*y[3]
 
-	return z.Set(&res)
+	z[0], z[1], z[2], z[3] = res0, res1, res2, res3
+	return z
 }
 
 // MulOverflow sets z to the product x*y, and returns z and  whether overflow occurred
@@ -403,23 +403,22 @@ func (z *Int) MulOverflow(x, y *Int) (*Int, bool) {
 
 func (z *Int) squared() {
 	var (
-		res                    Int
 		carry0, carry1, carry2 uint64
-		res1, res2             uint64
+		res0, res1, res2, res3 uint64
 	)
 
-	carry0, res[0] = bits.Mul64(z[0], z[0])
+	carry0, res0 = bits.Mul64(z[0], z[0])
 	carry0, res1 = umulHop(carry0, z[0], z[1])
 	carry0, res2 = umulHop(carry0, z[0], z[2])
 
-	carry1, res[1] = umulHop(res1, z[0], z[1])
+	carry1, res1 = umulHop(res1, z[0], z[1])
 	carry1, res2 = umulStep(res2, z[1], z[1], carry1)
 
-	carry2, res[2] = umulHop(res2, z[0], z[2])
+	carry2, res2 = umulHop(res2, z[0], z[2])
 
-	res[3] = 2*(z[0]*z[3]+z[1]*z[2]) + carry0 + carry1 + carry2
+	res3 = 2*(z[0]*z[3]+z[1]*z[2]) + carry0 + carry1 + carry2
 
-	z.Set(&res)
+	z[0], z[1], z[2], z[3] = res0, res1, res2, res3
 }
 
 // isBitSet returns true if bit n-th is set, where n = 0 is LSB.