Although I still think the results we saw for the system.time() on the very large lme run are peculiar, we probably will be able to circumvent the problem. Jose' put in several calls to proc.time() in the S-PLUS version and isolated the problem. We have this very elegant code that calculates a complicated of dimensions. That one piece of auxillary code that piece of code takes up about 90% of the execution time. It does a lot of cbind, rbind, and apply operations where apply is working across rows. It's a nice way to think of things but it takes forever on large matrices. Once we get that piece of code cleaned up we will go back to the question of what system.time() does under Linux. -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._