diff --git a/qsort.h b/qsort.h
index 46edbd5..dc32535 100644
--- a/qsort.h
+++ b/qsort.h
@@ -102,14 +102,15 @@ static inline pret3 partition3(uint32_t array[], int low, int high, uint32_t piv
 
 	/* index until smaller or eq elements lay */
 	int i = (low - 1);
-	/* index until smaller or eq elements lay */
-	int i2 = (high + 1);
+	uint32_t pc = 0;
 
 	/* traverse each element of the array */
 	/* compare them with the pivot */
-	int j2 = high; // j2 > i; --j2) {
 	#pragma GCC unroll 4
 	for (int j = low; j <= high; ++j) {
+		/* Branchless pivot-count */
+		pc += (array[j] == pivot);
+
 		if(array[j] < pivot) {
 			/* if element smaller than pivot is found */
 			/* swap it with the greater element pointed by i */
@@ -118,16 +119,28 @@ static inline pret3 partition3(uint32_t array[], int low, int high, uint32_t piv
 			/* swap element at i with element at j */
 			swapit(&array[i], &array[j]);
 		}
-		if(j2 > i) {
-			if (array[j2] > pivot) {
-				/* if element smaller than pivot is found */
-				/* swap it with the greater element pointed by i */
-				--i2;
+	}
 
-				/* swap element at i with element at j */
-				swapit(&array[i2], &array[j2]);
-			}
-			--j2;
+	/* Can spare out the second loop in these cases */
+	if(pc < 2) {
+		pret3 ret;
+		ret.leftend = i;
+		ret.rightend = i + 1;
+		return ret;
+	}
+
+	/* index until smaller or eq elements lay */
+	int i2 = (high + 1);
+
+	#pragma GCC unroll 4
+	for (int j = high; j > i; --j) {
+		if (array[j] > pivot) {
+			/* if element smaller than pivot is found */
+			/* swap it with the greater element pointed by i */
+			--i2;
+
+			/* swap element at i with element at j */
+			swapit(&array[i2], &array[j]);
 		}
 	}