Feature request: speed up calculation

[richelbilderbeek]

From @dksam:

Dear Dr Richel Bilderbeek, will you at least make iit possible to check a
total of 9 or 16 models comparing 2 by 2 or 3 by 3 or 4 by 4, then it will
help make a grand statement, out of 16 models tested, this had the lowest
log ratios and this the highest, I have a good mac which has 64 GB RAM, and
I can run for 2 weeks if needed, but this statement will help (not that all
will understand it,) thanks regards,Samuel

[richelbilderbeek]

Hi @dksam, as far as I understand, you would like the calculation to be faster. Am I correct?

If yes, I could have suggested to contact the BEAST2 developers, especially the ones that created the BEAST2 NS package, as that is were the big calculations take place (mcbette only allows one to do so from R).

I do not suggest to contact these developers, because Bayesian inference simply is a hard calculation: also they cannot help that.

Instead, I suggest to accept the 'slow' calculations as a given, and simply use less models in the comparison.

Cheers, Richel Bilderbeek