By Joseph Rickert In a recent blog post, David Smith reported on a talk that Steve Yun and I gave at STRATA in NYC about building and benchmarking Poisson GLM models on various platforms. The results presented showed that the rxGlm function from Revolution Analytics’ RevoScaleR package running on a five node cluster outperformed a Map Reduce/ Hadoop implementation as well as an implementation of legacy software running on a large server. An alert R user posted the following comment on the blog: As a poisson regression was used, it would be nice to also see as a benchmark the computational speed when using the biglm package in open source R? Just import your csv in sqlite and run biglm to obtain your poisson regression. Biglm also loads in data in R in chunks in order to update the model so that looks more similar to the RevoScaleR setup then just running plain glm in R. This seemed like a reasonable, simple enough experiment. So we tried it. The benchmark results presented at STRATA were done on a 145 million record file, but as a first step, I thought that I would try it on a 14 million record subset that I already had loaded on my PC, a quad core Dell, with i7 processors and 8GB of RAM.  It took almost an hour to build the SQLite data base: # make a SQLite database out of the csv file library(sqldf) sqldf("attach AdataT2SQL as new") file <- file.path(getwd(),"adatat2.csv") read.csv.sql(file, sql = "create table main.AT2_10Pct as select * from file">