Tuning MrBayes - Swapfreq, Nswaps, Nchains, and Temp
Hi,
For an alignment containing 1000s of sequences, what is the best setting for Swapfreq, Nswaps, Nchains, and Temp in MrBayes? I am running an MPI version and tried to raise the number of chains (Nchains) each running on a different CPU core. But did not get significant time decrease compared to 8 chains on 8 cores. Convergence diagnostics and automatic stopping was ON (stoprule=yes stopval=0.01).
Thx,
Victor
0 answers
No answers yet.
Log in to answer this question.
More posts like this
See What do I do when it is difficult to get convergence.
Thanks for the tip, checked it out. It just seems impractical. You have your large dataset, the analysis converges, but takes ages. If it is done already, you don't start to optimise it. Is there any way/hint to set up these values correctly for a large number of chains/cores?