Slight difference of the results for same job

[HUI WANG]

Hello all,

I ran the same case with same number of processors(1024) on HPC two days ago and today respectively, but the results are slightly different. The attached graphs are the evolution of dt and grids. Although most of the features of the results are almost the same, I prefer to get rid of this uncertainty. 

Previously, I noticed that the results may show tiny difference too if I run the same job with different numbers of processors or different clusters, but I am not sure if they are the problem. 

Best!

Hui