On Jun 26, 2011, at 5:45 PM, Zheng Yin wrote:

> Hi, I used GP by ECJ to do experiments and had two findings below.  
> Hope you
> can help to explain.
> 1. Same PC, same ECJ parameter setting, same seed number, experiments
> were run at different time. One was run at end of 2009; the other  
> was run
> yesterday. The results were different.
>
> 2. Different PC, same ECJ parameter setting, same seed number,  
> experiments
> were run at different time(around 2 months time). The results from 2  
> pcs
> were same.

ECJ does not promise identical results between versions.

Sean