ACSF and split frequency plots take a REALLY long time

[roblanf]

There must be something we can do about this. I'm trying out some test data and it's pretty intolerable...

[danlwarren]

We can make them faster by moving to bin2d plots instead of points. Not as pretty but much faster draw time.

On Fri, Dec 6, 2019 at 7:34 AM roblanf ***@***.***> wrote: There must be something we can do about this. I'm trying out some test data and it's pretty intolerable... — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#147?email_source=notifications&email_token=AA4YUC6OBKEITP7DKZYLOULQXF6WBA5CNFSM4JWDLY5KYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4H6OVGNA>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA4YUC3QQBFR5TU5BRSUHLDQXF6WBANCNFSM4JWDLY5A> .

-- Dan Warren Email: dan.l.warren@gmail.com Phone (US): 530-848-3809 Skype: dan.l.warren Google Scholar <https://scholar.google.com/citations?user=NTzu9c8AAAAJ&hl=en> Orcid <http://orcid.org/0000-0002-8747-2451> ResearcherID <http://www.researcherid.com/rid/B-3821-2010> Scopus <http://www.scopus.com/authid/detail.url?authorId=7202133982>

[roblanf]

First let's check if it's the drawing time or the calculation time. If it's the calculation time we might need to see if we can draft in some help to speed it up. I'm fairly sure that whatever machinery is used for 'rf.dist()' is the same as would be used for these. So I think in principle the calculations don't need to take longer than a call to rf.dist() for the same trees. It's probably just that our code is really inefficient...

[danlwarren]

My impression from screwing around with it the other day is that it’s draw time. Those xy point plots get slow as hell past a certain number of points. It’s the same reason ggpairs sucks for large data sets.

[roblanf]

We might be getting our wires crossed here. I'm talking about the sliding window and cumulative plots.

[danlwarren]

Oh! I thought you meant the multi chain comparison plots. Sorry.