Repeat vx_wmma issue & hardcode dst address

tests/kernel/reductions/main.cpp

@@ -138,6 +138,7 @@ void test_maxu_reduce() {
 	y = reduced;
 }
 
+// assumes NUM_THREADS == 4
 unsigned bit_vectors[4] = {0b11010110000111001100010100100110, 0b10010100011010001010000000001110, 0b10001001010111110001110000000010, 0b00010011010100101101110111001111};
 
 void test_and_reduce() {
@@ -213,4 +214,4 @@ int main()
 	
 
 	return 0;
-}
+}