In WinVector/RcppDynProg: 'Rcpp' Dynamic Programming

knitr::opts_chunk$set(fig.width=12, fig.height=8) 
library("RcppDynProg")
library("WVPlots")
library("microbenchmark")
library("rqdatatable")


set.seed(2018)
n <- 500
x <- matrix(runif(n*n), nrow=n, ncol=n)

solve_interval_partition(x, n)

solve_interval_partition_R(x, n)

timings <- microbenchmark(
  solve_interval_partition(x, n),
  solve_interval_partition_R(x, n),
  times = 5L)

print(timings)

p <- data.frame(timings)
p$seconds <- p$time/1e+9
p$method <- as.factor(p$expr)
p$method <- reorder(p$method, p$seconds)

summary <- p %.>%
  project(., 
          mean_seconds = mean(seconds),
          groupby = "method")
print(summary)
ratio <- max(summary$mean_seconds)/min(summary$mean_seconds)
print(ratio)

WVPlots::ScatterBoxPlotH(p, 
                         "seconds", "method", 
                         "performance of same dynamic programming code in R and Rcpp (C++)")

---

Timings on a 2018 Dell XPS-13 laptop, 16 Gib RAM, LPDDR3, 2133 MT/s, Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz (8 cores reported), idle, charged, and plugged into power supply. Ubuntu 18.04.1 LTS.

R.version.string

R.version

sessionInfo()

WinVector/RcppDynProg documentation built on Aug. 26, 2023, 4:47 a.m.