Slightly different results if I clean up the eigenFloor version and
use unif_rand() instead of runif. The sugar version is still faster,
however.
R version 2.15.1 (2012-06-22) -- "Roasted Marshmallows"
Copyright (C) 2012 The R Foundation for Statistical Computing
ISBN 3-900051-07-0
Platform: x86_
Odd, armaFloor() never won on my end (Intel i7, Ubuntu 64bit). If
eigenFloor() is slow, could is reveal an additional copy of something?
Dirk
--
Dirk Eddelbuettel | e...@debian.org | http://dirk.eddelbuettel.com
___
Rcpp-devel mailing list
Rcpp-dev
Dirk Eddelbuettel recently blogged on "Faster creation of binomial
matrices" (http://dirk.eddelbuettel.com/blog/code/snippets/) comparing
various approaches in R and using C++ callable from R with the Rcpp
package. I used cut-and-paste to create a script from the code in his
posting but omitting t
