[R-br] byte-compiler no R

[Paulo Justiniano]

A noticia abaixo é relevante
para novas versões e ver tendencias no projeto
P.J.

------------------------------------------------------------------------
With byte compiler, R 2.14 will be even faster
August 2, 2011
By David Smith

This article was first published on Revolutions, and kindly contributed to 
R-bloggers

In a presentation at the JSM 2011 conference in Miami yesterday, R core 
member Luke Tierney revealed that the next major update to R, R 2.14, will 
feature improved speed when processing interpreted R code, thanks to 
standard use of the new byte compiler feature.

The byte compiler was introduced with R 2.13, but while R developers could 
use it for their own functions, the standard base and recommended packages 
were unaffected by the byte compiler. Starting with R 2.14, all of the 
standard functions and packages in R will be pre-compiled into byte-code, 
which in some cases can speed up performance by a factor of 5x or more. 
(In an experimental version of the compiler, which may make it into 2.14, 
even greater speedups are possible.) The benefits accrue mainly to pure R 
functions which deal with scalars and very short vectors -- R functions 
which call out to C code and operations on large vectors won't be affected 
much. With the new compiler, Tierney says there should be fewer occasions 
where R programmers need to turn to C or other external languages to speed 
up R code.

In other news for R 2.14, Tierney says that the new version may also make 
transparent use of parallel processing for some operations on multi-core 
machines. The colSums and dist functions already include hidden features 
for parallel processing, and if tests go well features like this may 
become the default in future versions of R. Features from Tierney's 
experimental pnmath package, which parallelizes some basic math routines 
in R, may also make it into the next release.

R 2.14 is expected to be released later this year.