[ojAlgo-user] Performance tests

[Anders P. <ap...@op...>]

These are the latest timings...

1000x1000 matrices and performing each operation 10 times.

compute() ojAlgo/Jama

Chol: 7,31/8,22
LU: 11.12/11.85
QR: 97.26/100.00

solve() ojAlgo/Jama

Chol: 32.74/293.06
LU: 32.97/52.47
QR: 170.97/324.64

The main news is that the ojAlgo "native" LU.compute() is now faster 
than (or at least roughly as fast as) the Jama version. It used to be 
clearly slower. There is some variation in the timings between different 
runs, but the trend is clear.

http://ojalgo.org/generated/comparison.xml

/Anders