memory overflow

[ppxwwtaking]

You mentioned you did the tree analysis on 1M sequences. How do you handle the memory issue? I have ~470k sequence running veryfasttree on local workstation with memory ~750G. But if failed due to memory overflow when running disttbfast for distance matrix calculation. Any suggestion?
Thanks

[cesarpomar]

In our case, the 1M-sequence tree was computed on nodes with 2 TB of RAM, and memory needs depend not only on the number of sequences but also on their length.

If your workstation doesn’t have enough RAM, try enabling the -disk-computing option; and if you have a high-performance storage device, specify it with -disk-computing-path [path] so the temporary files use that location. Also, avoid using all available threads, since that can create an I/O bottleneck. Start with a small number of threads and increase gradually until your disk is fully utilized.

Thanks for using VeryFastTree!