The results should not change based on the LS-RUN version you are using. LS-RUN is only used to run LS-DYNA and to send the LS-DYNA execution command lines in the background. The LS-DYNA solver crunches the data. So, which LS-DYNA version are you using? The latest version is R13, but you could be running any version (R12, R11, R10, R9, R8, R7, etc.). You can download any LS-DYNA solver here:
You can also check the release notes to see what has changed between versions:
After you run a model, you can open the d3hsp file in a text editor to see which version you used:
To get consistent results between runs in SMP, make sure you use consistency checking by setting ncpu=-4 (this is negative 4) to a negative number. This way, your results should be consistent when using different number of cores:
If you are running MPP, you will get slightly different results with different number of cores. So, pick a number of cores and stick with it to calibrate your model.
At last, improvements in newer versions of LS-DYNA can lead to different results compared with older LS-DYNA solvers. For this reason, please calibrate your model and stick to a revision of LS-DYNA. For example, if you calibrate your model with R13, then stick with R13. This is why some compagnies (car manufacturers, aerospace companies, etc.) stick to the same version of LS-DYNA for some time, even though many new versions are released.