For my handful of followers, you may have noticed something strange. Over the past two days, I released updates to all four of my packages on CRAN. I decided to practice what I preach and actively reduce dependency hell by moving from testhat to tinytest. I have been programming for ...

This post outlines the steps needed for building R 4.2 for Windows with OpenBLAS using the the new Rtools42 build system. With the soon-to-be released new UCRT toolchain for building R on Windows, the build process changes once again. First and foremost, we Windows users owe a deep debt of gratitude ...

This post outlines the steps needed to build R 4+ for Windows with OpenBLAS. The release of R 4.0 includes significant changes to the Windows build system from prior versions—for the better! Before anything, we all owe Jeroen Ooms significant gratitude for the many hours he spent working on the build ...

Searching for Speed In my previous post, I described the method I use for compiling Fortran (or C) into an R package using the .Call interface. This post will compare the speed of various implementations of the layer loss cost function. The Function Often, insurance or reinsurance is bought in ...

Searching for Speed Everyone who has ever used R has, at one time or another, wished for an increase in R’s speed. If you haven’t, you’re not using R hard enough! Recently, as part of some research on credibility, I was calculating layer loss costs for millions ...

Recently, the R-consortium accepted a new project called histoRicalg. The main goal of the project is to document and transfer knowledge of some older algorithms used by R and by other computational systems. There is a lot of R written in Fortran—much of which is in the old F77 ...

With the recent releases of R 3.2.4 and OpenBLAS 2.17, I decided it was time to re-benchmark R speed. I’ve settled on a particular set of tests, based on my experience as well as some of Simon Urbanek’s work which I separated into two groups: those focusing on BLAS-heavy operations ...

A number of years ago, someone asked me "why does my company need actuaries to fit curves, once I have the mean and standard deviation of my losses, isn't that enough?" I explained to him that not every distribution is completely determined by its mean...