Speed improvement [R for BE/BA]

posted by ElMaestro  – Denmark, 2020-08-08 12:25 (527 d 15:26 ago) – Posting: # 21839
Views: 10,189

I am seeing about a 30% speed improvement on my machine when using arrays of matrices. Am using lists for that purpose, that's the only way I can think of.
Further improvement may be possible, because I may only need to store principally three different matrices in these arrays (per array) an keep track of which subjects use which matrices (that would be defined by their sequences), but this, I think, for now is not a priority for me to do.

A further option is to only use one kind of matrix and then reorder all periods in the global data list, so that all subjects have e.g. R-T-R, then we'd need to work with only a single matrix type in the profiles; this may improve speed further. Will think about it...

30%...It ain't much but it's honest work. :-)

Pass or fail!
ElMaestro

Complete thread:

Activity
 Admin contact
21,829 posts in 4,566 threads, 1,555 registered users;
online 3 (0 registered, 3 guests [including 3 identified bots]).
Forum time: Tuesday 02:51 CET (Europe/Vienna)

The generation of random numbers is too important
to be left to chance.    Robert R. Coveyou

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5