Improved Matrix Multiplication by Changing Loop Order

Table 4

Results extracted by changing row and columns using serial code.

N	Loop order (i is the outermost loop)	Time taken (seconds)
		C++ (serial code)	Java (serial code)	Python (serial code)
500	i, j, k	0.312	0.324	2.345
1000	i, j, k	2.334	3.932	5.456
1500	i, j, k	10.881	16.245	23.453
2000	i, j, k	37.382	43.88	122.304
2500	i, j, k	97.293	124.593	566.452