|
O-Matrix Matrix Function Performance
| Benchmark |
O-Matrix 6.4 |
Matlab 7.01 |
| Matrix Power, A^4 for 750x750 random matrix |
0.287 |
0.322 |
| Frobenius norm of 2048x2048 random matrix |
0.006 |
0.009 |
| LU Factorization of 512x512 random matrix |
0.023 |
0.041 |
| QR Factorization of 512x512 random matrix |
0.174 |
0.125 |
| Cholesky Factoring of 512x512 random matrix |
0.012 |
0.014 |
| SVD of 512x512 random matrix |
0.284 |
0.255 |
| Outer Product of 1000 element vectors |
0.004 |
0.016 |
| Matrix Product of 512x512 random matrices |
0.030 |
0.039 |
| Kronecker Tensor Product of 50x50 identity and random matrix |
0.033 |
0.236 |
All timings are in seconds. - Run on a 3 GHz Pentium 4
On my HP Pavilion computer, built in 2002 with a 2.53 GHz Pentium 4 chip,
I do a speed test, multiplying two 2000 x 2000 element matrices in double precision.
Counting both multiplies and adds as floating-point operations,
O-Matrix version 6.3 runs the test at 3.91 gigaflops.
- Jim Worthey
|
