Data Science. TileDB. Open Source. Quant Research. R. C++. Debian. Linux. Adjunct Clinical Professor, University of Illinois. Lots of coffee. And some running.
Rcpp release candidate 1.0.3.8 at Rcpp drat
If you can, please help test prior to the 1.0.4 release.
Details in a post on the rcpp-devel list
lists.r-forge.r-project.org/…#rcpp#rstats
Finally able to catch up to this, and it is fantastic.
Kudos to @stanleynelson1 for an excellent retrospective on @MilesDavis who changed so much in music, and so often.
Streams at pbs.org/video/miles-davis-bi… if you're in the right region (i.e. could not see it in DK last week).
Yes! Related tips:
- use wrappers that encode `Ncpus`, all `littler` ones do
- blog posts about this as even faster than parallel installs are installs that
- use `ccache`
- use binaries instead
- build once, eg in Docker, and reuse
are below dirk.eddelbuettel.com/blog/c…
Quick #rstats Tip: Use `Ncpus` when installing packages
🔹install.packages() has an Ncpus argument determining how many parallel processes to use
🔹This can e.g. speed up installation of {{tidyverse}} from ~11min to ~4.6min
🔹Also works with remotes & devtools install_ functions
RcppAnnoy 0.0.15 on CRAN: Small fixes, more coming!
(Very Fast) Approximate Nearest Neighbours for R
dirk.eddelbuettel.com/blog/2…#rcpp#rstats
Delayed tweet / blog post as I was out at the lovely @celebRtion2020
/cc @fulhack and twitterless Aaron and James
And the @Debian (source) package has been prepared and uploaded -- for once while being in the same room as @pdalgd on the very happy occassion of @celebRtion !
@Ubuntu packages to follow once @marutterstat gets to it.
And the @Debian (source) package has been prepared and uploaded -- for once while being in the same room as @pdalgd on the very happy occassion of @celebRtion !
@Ubuntu packages to follow once @marutterstat gets to it.
One option: use Docker image r-base (same as rocker/r-base)
It uses @Debian testing which has gcc 10 now. So
apt-get update
apt-get install gcc-10 g++-10 gfortran-10
then adjust your package. "We all" need to work on replicatin CRAN's setup better.
#rstats#rockerproject
Nice overview of the 'new data economy' in @TheEconomist by Ludwig Siegele (aka @EconoScribe): new tech, new data sources, new goverment response, new challenges, possibly new markets, likely new snake oil salesmen -- a useful survey
economist.com/special-report…#data
digest 0.6.25 on CRAN: spookyhash fix
Hash digests for arbitrary R objects
dirk.eddelbuettel.com/blog/2…#rstats
spookyhash fix thanks to (twitterless) Aaron Lun and Kendon Bell
CRAN takes its Policy very seriously. 108 packages removed in one day.
Follow CRANberries via html, rss or here via @CRANberriesFeed
Follow CRAN Policy changes on @GitHub or via @CRANPolicyWatch#rstats
Watched the first two: the @tvaneerd talk about 'Postmodern C++' is just fantastic.
About programming, what is said and unsaid in code, lots (LOTS) of Dr Seuss (!!), some Derrida (!! -- whom this @EHESS_fr grad never read) and more puns than you can shake a stick at.
Superb.
R-devel now officially has 'raw' character constant strings like C++ as it just made the NEWS.Rd file (and its nightly RSS feed). Thanks once again to @LukeTierney4 for what may become quite widely used.
#rstats#rdevel
Or keep an older version of package:roxygen2 around that manages to build Rd files just fine without requiring package re-compilation---as my `roxy.r` helper does.
github.com/eddelbuettel/litt…#rstats#rcpp
Small #Rcpp + @RStudio beginner tip.
If you configure your project build tools to roxygenise at each "Install and restart", your Rcpp code will be compiled with debug flags and no optimisation.
To make your code fast again, choose "More" > "Clean and Rebuild".
#rstats