Search Results: "Dirk Eddelbuettel"

26 January 2014

Dirk Eddelbuettel: RcppZiggurat 0.1.2

A new maintenance release of RcppZiggurat is now on CRAN. And just like yesterday's RcppClassic upload, this is motivated chiefly by some minor caretaking for an upcoming Rcpp release -- hopefully soon, and with more details to follow. This release also contains some small extensions and work-in-progress. By using R's uniform generator, we can fall back to R's seeding of the RNG -- but going through the R data structures has a performance hit. Similarly, and somewhat surprisingly, the default generators in both Boost and the C++11 library are not all that speedy. The NEWS file entry shown below lists all changes.
Changes in version 0.1.2 (2013-12-28)
  • Add a new generator deploying R's unif_rand which faciliates use of Ziggurat as the user-supplied generator for N(0,1)
  • Update a 'local' demo comparing normal distribution RNGs from Boost, C++11 and Armadillo none of which are particularly speedy
  • Add declaration to import a symbol from Rcpp to NAMESPACE to ensure proper instantiation with the upcoming Rcpp version
Courtesy of CRANberries, there are diffstat reports for the most recent release relative to the preceding version. More detailed information is on the RcppZiggurat page. Questions, comments etc should go to the rcpp-devel mailing list off the R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

Dirk Eddelbuettel: RcppClassic 0.9.5

A maintenance release of RcppClassic, now at version 0.9.5, went out to CRAN today. Nothing changed in user-facing code. The build was triggered by a change in an upcoming Rcpp release (more about that another time), and a few minor updates were needed for extended checks done by R. Courtesy of CRANberries, there is the set of changes relative to the previous release 0.9.4. Questions, comments etc should go to the rcpp-devel mailing list off the R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

22 January 2014

Dirk Eddelbuettel: RcppArmadillo 0.4.000.2

A bug fix release of Armadillo came out and was turned into a new RcppArmadillo release which is now on CRAN and will go shortly to Debian. The complete list of changes is below.
Changes in RcppArmadillo version 0.4.000.2 (2014-01-21)
  • Upgraded to Armadillo release Version 4.000.2
    • fix for randi() generating out-of-interval values
    • workaround for a bug in the Intel compiler
Courtesy of CRANberries, there is also a diffstat report for the most recent release As always, more detailed information is on the RcppArmadillo page. Questions, comments etc should go to the rcpp-devel mailing list off the R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

20 January 2014

Dirk Eddelbuettel: RProtoBuf 0.4.0: A whole lot of goodies and Windoze support

A new major release 0.4.0 of RProtoBuf, is now on CRAN. RProtoBuf provides GNU R bindings for the Google Protocol Buffers ("Protobuf") data encoding library used and released by Google, and deployed as a language and operating-system agnostic protocol by numerous projects. With this release, we are welcoming Jeroen Ooms to the team. Jeroen had already worked on some RProtoBuf extensions in the context of his OpenCPU project; we have now integrated those Protocol Buffers functions. And Jeroen pushed all the right buttons to finally get RProtoBuf built on everybody's least-favourite operating system by provding a static library for use by win-builder and CRAN. Murray once again did a lot of work on internals. His use of the LLVM tool llvm-format was particular helpful to make our coding style a little more consistent. The complete NEWS file entry for this release follows:
Changes in RProtoBuf version 0.4.0 (2014-01-14)
  • Changes to support CRAN builds for MS Windows.
  • Added functions serialize_pb, unserialize_pb, and can_serialize_pb plus documentation from Jeroen Ooms RProtoBufUtils package.
  • New dir inst/python with some Python examples.
  • Added Jeroen Ooms as author.
  • Vignettes have been converted to the R 3.0.0 or later use of external vignette builders, no longer need a Makefile
  • Added missing methods to dollar completion list for Message, Descriptor, EnumValueDescriptor, and FileDescriptor classes.
  • Add missing export for .DollarNames EnumValueDescriptor to allow completion on that class.
  • Add more than 15 additional pages to the main Intro vignette documenting better all of the S4 classes implemented by RProtoBuf, updating the type mapping tables to take into account 64-bit support, and documenting advanced features such as Extensions.
  • Added better error checking in EnumDescriptors to catch the case when wrong types are provided.
  • Updated the FileDescriptor name() method to accept a boolean for full paths just like the generic name() method.
  • Correct a bug that incorrectly dispatched as.character() when as.list() was called on Descriptor objects.
  • Update FileDescriptor $ dispatch to work properly for the names of fields defined in the FileDescriptor, instead of just returning NULL even for types returned by $ completion.
  • Added a reservation for extension fields in the example tutorial.Person schema.
  • Support setting int32 fields with character representations and raise an R-level stop() error if the provided string can not be parsed as a 32-bit integer, rather than crashing the R instance.
  • Update the project TODO file.
  • Add better documentation and tests for all of the above.
  • Corrected the handling of uint32 and fixed32 types in protocol buffers to ensure that they work with numbers as large as 2^32 - 1, which is larger than an integer can hold in R since R does not have an unsigned integer class. These values are stored as doubles internally now to avoid losing precision.
  • Added unit tests to verify behavior of RProtoBuf with extreme values for uint32 types.
  • Removed old exception handling code and instead rely on the more modern Rcpp::stop method maintained in Rcpp.
  • Add better error messages when setting a repeated field of messages to inform the user which element index was of the wrong type and what the expected type was.
  • Add an optional 'partial' argument to readASCII allowing uninitialized message fragments to be read in.
  • (internal) Added const qualifiers in more places throughout the C++ code for type safety.
  • (internal) Standardize coding conventions of the C++ files and run them through clang-format for consistency. A STYLE file has been submitted to R-Forge with details about the coding standards and how they are enforced with Emacs and clang-format.
  • Applied changes suggested by Kevin Ushey to the S4 class handling to support both the currently released Rcpp as well as the currently pending next version.
CRANberries also provides a diff to the previous release 0.3.2. More information is at the RProtoBuf page which has a draft package vignette, a 'quick' overview vignette and a unit test summary vignette. Questions, comments etc should go to the rprotobuf mailing list off the RProtoBuf page at R-Forge.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

7 January 2014

Dirk Eddelbuettel: RcppArmadillo 0.400

A new major release 4.000 of Armadillo came out out yesterday morning. This is a new stable release by Conrad with a new major number, and it brings a few nice changes as detailed below. I rolled up RcppArmadillo version 0.400 yesterday morning, and following a little bit of discussion this version is now on CRAN and in Debian. The complete list of changes is below.
Changes in RcppArmadillo version 0.4.000 (2014-01-05)
  • Upgraded to Armadillo release Version 4.000 (Feral Steamroller)
    • added eigen decompositions of sparse matrices: eigs_sym() and eigs_gen() [ but this requires linking against ARPACK which RcppArmadillo as a pure-template package does not do, and R is not linked against ARPACK either. ]
    • added eigen decomposition for pair of matrices: eig_pair()
    • added simpler forms of eig_gen()
    • added condition number of matrices: cond()
    • expanded find() to handle cubes
    • expanded subcube views to access elements specified in a vector
    • template argument for running_stat_vec expanded to accept vector types
    • more robust fast inverse of 4x4 matrices
    • faster divide-and-conquer decompositions are now used by default for eig_sym(), pinv(), princomp(), rank(), svd(), svd_econ()
    • the form inv(sympd(X)) no longer assumes that X is positive definite; use inv_sympd() instead
    • added MEX connector for interfacing Octave/Matlab with Armadillo matrices (contributed by George Yammine)
Courtesy of CRANberries, there is also a diffstat report for the most recent release As always, more detailed information is on the RcppArmadillo page. Questions, comments etc should go to the rcpp-devel mailing list off the R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

3 January 2014

Dirk Eddelbuettel: New BH release 1.51.1-4

A new release of the BH package is now on CRAN and its mirrors. BH provides (a sizeable subset of) the Boost library for C++, particularly the (large) parts delivered as pure template headers not requiring linking. See the BH page for more details. This release comes from a rewritten build script for the package. We no longer drive the initial set of Boost libraries by the requirements of the bigmemory (and synchronicity) packages, but rather explicitly enumerate an equivalent set of Boost libraries. This ensures that these libraries are included completely. We also move the project repository to this GitHub repo. The complete list changes follows below.
Changes in version 1.51.0-4 (2014-01-01)
  • Rewritten main package creation script to no longer scan for what bigmemory and synchronicity use, but rather explicitly copy over an (equivalent) list of explicitly-enumerated Boost libraries
  • Repository moved from R-Forge to GitHub, scripts and layout adjusted accordingly
  • Besides the implicitly expanded coverage by including the complete libraries, we also expanded from math/distributions to all of math.
Courtesy of CRANberries, there is also a diffstat report for the most recent release. Comments and suggestions are welcome via the mailing list or issue tracker still available via the package page at R-Forge.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

24 December 2013

Dirk Eddelbuettel: RcppZiggurat 0.1.0 (and 0.1.1): Faster N(0,1) RNGs

Over the last few weeks I have been working on getting the Ziggurat normal random number generator updated and available in R. The Ziggurat generator provides a pretty unique combination of speed and good statistical properties for (standard) normal random numbers (as opposed to uniform draws as is commonn for most RNGs). Generation of N(0,1) draws may not by itself be the dominant slowdown in a simulation, yet when large number of draws are required it may be helpful to have a generator that is faster than the defaults in R (which have excellent properties, but not the fastest speed). A first release 0.0.1 went to CRAN a couple of weeks ago. This was followed up by a more thorough release 0.1.0 this last weekend which, as it happens, needed a minor follow-up 0.1.1 to clean up some dependencies on the right R version, as well as vignette building procedures. I added a web page about RcppZiggurat to group together some basic information, but the single best starting point may be the detailed pdf vignette included in the package. Courtesy of CRANberries, there are diffstat reports for the most recent release as well as for the preceding release two days earlier. More detailed information is on the RcppZiggurat page. Questions, comments etc should go to the rcpp-devel mailing list off the R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

20 December 2013

Dirk Eddelbuettel: New RcppEigen release 0.3.2.0.1 -- and a new maintainer

In a recent email to the Rcpp and lme4 mailing lists, Doug Bates announced that was turning away from R, Rcpp, lme4 and hence also RcppEigen for which he had been both the primary author and maintainer. This is huge loss for the R community. I have known Doug since the 1990s. He has been a fairly central figure around R during all those years in which I got more and more involved with R. I have learned a lot from him, and enjoyed the work together---initially on the Debian R package (which I took over from him), and all the way to joint work on Rcpp and RcppEigen, including our JSS paper. I am certain to miss him around R. Now, in order to keep RcppEigen viable within CRAN and the R ecosystem, I have offered to maintain it. A first new upload is now on CRAN (and I also uploaded it to Debian where I started to maintain it too as a depedency for lme4). I have also started to make a few minor changes such as tightening Suggests: a little, and editing a few descriptive files. Details are in the Github repo. Questions, comments etc about RcppEigen should go to the rcpp-devel mailing list off the R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

16 December 2013

Dirk Eddelbuettel: RProtoBuf 0.3.2

A new version 0.3.2 of RProtoBuf, is now on CRAN. RProtoBuf provides GNU R bindings for the Google Protobuf data encoding library used and released by Google. As for the last few releases, Murray took charge of most changes. The NEWS file entry follows:
Changes in RProtoBuf version 0.3.2 (2013-12-15)
  • Fixed a bug that erroneously prevented users from setting raw byte fields in protocol buffers under certain circumstances.
  • Give a user friendly error message when seting an extension to a message of the wrong type instead of causing a C++ check failure that terminates the Rsession.
  • Change object table lookup slightly to allow users to use the <<- operator in code using RProtoBuf without hitting a stop() error in the lookup routine.
  • Add missing enum_type method and improve show method for EnumValueDescriptors.
  • Improve documentation and tests for all of the above.
  • Rewrote tests/ script calling RUnit tests
CRANberries also provides a diff to the previous release 0.3.1. More information is at the RProtoBuf page which has a draft package vignette, a 'quick' overview vignette and a unit test summary vignette. Questions, comments etc should go to the rprotobuf mailing list off the RProtoBuf page at R-Forge.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

14 December 2013

Dirk Eddelbuettel: RcppDE 0.1.2

A maintenance release (now at version 0.1.2) of my RcppDE package (previously described in these two posts) is now CRAN. More details about the package are available in the vignette also included in the RcppDE R package. Changes were minimal and driven mostly by some CRAN Policy changes which now prefer vignette sources files in (top-level) directory vignettes/ Courtesy of CRANberries, there is also a diffstat report for the most recent release. Current and previous releases are available here as well as on CRAN.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

Dirk Eddelbuettel: random 0.2.2

A maintenance release of my random package for truly (hardware-based) random numbers (pulled from random.org) is now on CRAN. It's been a while since previous releases. The package is described in a detailed vignette as well as in a essay by Mads Haahr. Changes were minimal and driven mostly by some CRAN Policy changes which now prefer vignette sources files in (top-level) directory vignettes/ Courtesy of CRANberries, there is also a diffstat report for the most recent release. Current and previous releases are available here as well as on CRAN.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

Dirk Eddelbuettel: gcbd 0.2.5

A maintenance release (now at version 0.2.5) of my gcbd package (described only in these two posts) is now CRAN. More details about the package are available in the paper which is also included in the gcbd R package. Changes were minimal and driven mostly by some CRAN Policy changes which now prefer vignette sources files in (top-level) directory vignettes/ Courtesy of CRANberries, there is also a diffstat report for the most recent release.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

10 December 2013

Dirk Eddelbuettel: RcppArmadillo 0.3.930.1

A new Armadillo release 3.930 came out a few days ago, with a very nice set of changes (see below). I rolled this into RcppArmadillo 0.3.930.0. However, one of these changes revealed that R shipped only the standard SVD for complex-valued matrices, and not the more performant divide-and-conquer approach. So in R builds using the default built-in Lapack, at least one CRAN package no longer built. After some back and forth, Conrad put some branching in the library to fall back to the standard SVD, and I added a built-time configuration test for an appropriate preprocessor directive used by the fallback code. This is now on which is now on CRAN and in Debian as RcppArmadillorelease 0.3.930.1, and Conrad will probably update the Armadillo page as well (though the fix is only needed with R's builtin Rlapack). Also of note is that R Core already added the missing Fortran routine zgesdd to R 3.1.0 (aka "R-devel") so this issue goes away with the next release. Also of note, I wrote up a short Rcpp Gallery post illustrating the performance gains available from divide-and-conquer SVD. The complete list of changes is below.
Changes in RcppArmadillo version 0.3.930.1 (2013-12-09)
  • Upgraded to Armadillo release Version 3.930.1
    • Armadillo falls back to standard complex svd if the more performant divide-and-conquer variant is unavailable
  • Added detection for Lapack library and distinguish between R's own version (withhout zgesdd) and system Lapack; a preprocessor define is set accordingly
Changes in RcppArmadillo version 0.3.930.0 (2013-12-06)
  • Upgraded to Armadillo release Version 3.930 ("Dragon's Back")
    • added divide-and-conquer variant of svd_econ(), for faster SVD
    • added divide-and-conquer variant of pinv(), for faster pseudo-inverse
    • added element-wise variants of min() and max()
    • added size() based specifications of submatrix view sizes
    • added randi() for generating matrices with random integer values
    • added more intuitive specification of sort direction in sort() and sort_index()
    • added more intuitive specification of method in det(), .i(), inv() and solve()
    • added more precise timer for the wall_clock class when using C++11
  • New unit tests for complex matrices and vectors
Courtesy of CRANberries, there is also a diffstat report for the most recent release As always, more detailed information is on the RcppArmadillo page. Questions, comments etc should go to the rcpp-devel mailing list off the R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

7 December 2013

Dirk Eddelbuettel: R and Big Data at Big Data Summit at UI Research Park

I spent yesterday at the very enjoyable Big Data Summit held at the University of Illinois Research Park at the edge of the University of Illinois at Urbana-Champaign. campus. My (short) presentation was part of a panel session on R and Big Data which Doug Simpson of the UIUC Statistics department had put together very well. We heard from a vendor / technology provider with Christopher Nguyen from Adatao talking about their "Big R", from industry with Andy Stevens talking about a number of some real-life challenges with big data at John Deere, from academia with Jonathon Greenberg talking about R and HPC for geospatial research and I added a few short comments and links about R, HPC and Rcpp. My few slides are now up on my talks / presentations page. Overall, a good day with a number of interesting presentations and of course a number of engaging hallway discussions.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

3 December 2013

Dirk Eddelbuettel: digest 0.6.4

digest version 0.6.4 is now on CRAN and in Debian. This is a pure maintenance release which should help with a build issue affecting users on Solaris. CRANberries provides the usual summary of changes to version 0.6.3. Our package is available via the R-Forge page leading to svn and tarball access, my digest page, the local directory here as well as via Debian and its mirrors.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

2 December 2013

Dirk Eddelbuettel: Recent Rcpp talks at U Chicago / Booth and U Kansas

In early October, I had an opportunity to talk about Rcpp and RcppArmadillo at the Statistical Computing Seminar at the Booth School of Business at the University of Chicago. And then two weeks ago, I had an invitation to talk at the Center for Research Methods and Data Analysis at the University of Kansas where I covered similar material as well as ongoing work on the RcppZiggurat package (for which I should have an updated version soon). Slides from both talks are now at the top of my talks / presentations page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

30 November 2013

Dirk Eddelbuettel: RcppCNPy 0.2.2

Right on the heels of release 0.2.1 of RcppCNPy, a new version 0.2.2 is now on CRAN. RcppCNPy uses the CNPY library by Carl Rogers to provide R with easy read and write access to NumPy files. The reason for the new version that I had experimented with a different way to test endianness (as needed for the NumPy file headers) but accidentally sent an interim tarball to CRAN which still wanted to include endian.h promptly breaking Windows builds. So now we do something even simpler and just rely on the (even more complete) test for endianness when R is built, which prevents all sorts of complications for us and builds everywhere (with thanks to Brian Ripley for the suggestion). While we were at it, we also added a new unit test. Full changes are listed below.
Changes in version 0.2.2 (2013-11-29)
  • Switched to using the result from the compile-time configuration for R to determine big or little endian (as needed for the NPy headers)
  • Added a new test (and test validation result file) for a complete save-reload cycle and comparison
CRANberries also provides a diffstat report for 0.2.2 relative to 0.2.1. As always, feedback is welcome and the rcpp-devel mailing list off the R-Forge page for Rcpp is the best place to start a discussion.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

28 November 2013

Dirk Eddelbuettel: RcppCNPy 0.2.1

Version 0.2.1 of RcppCNPy is now on CRAN earlier. This version is a minor upgrade with a few small changes as detailed in the NEWS file:
Changes in version 0.2.1 (2013-11-28)
  • Synchronized code with the cnpy repository
  • Added new function to test from R whether integers supported
  • Updated tests for integer support, if available
  • Updated vignette and discussion about need for rebuilding only RcppCNPy with the -std=c++11 flag if integer support is desired
  • Updated tests for integer support, if available
  • Updated THANKS file
  • Ensure that inclusion of Rinternal.h does not affect other headers by adding RF_NO_REMAP
CRANberries also provides a diffstat report for 0.2.1 relative to 0.2.0. As always, feedback is welcome and the rcpp-devel mailing list off the R-Forge page for Rcpp is the best place to start a discussion.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

24 November 2013

Dirk Eddelbuettel: Website and blog updated

Earlier this year the blog had its tenth anniversary. I had meant to celebrate the occassion by revamping the site and blog a little. Having set up the updated R/Finance site, the Rcpp Gallery and Rcpp sites as well as the much-needed overhaul of the html side of the CRANberries RSS feed (which also integrates it with the static blog compiler I use), I figured I just needed a little time to get this done. Well, eight months later we're there. It still uses Twitter Bootstrap for layout, and a slightly modified Bootswatch theme. Comments welcome, and please let me know if links are missing or going nowhere in places.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

20 November 2013

Dirk Eddelbuettel: RcppArmadillo 0.3.920.3

Conrad released a bug fix version 3.920.3 for his Armadillo library. So I prepared a new release 0.3.920.3 of RcppArmadillo which is nor on on CRAN and also already in Debian. Note that we skipped interim version .2 which addressed an Atlas linking issue which doesn't affect us in the R context as we get LAPACK and BLAS via the R process we get used from. The complete list of changes is below.
Changes in RcppArmadillo version 0.3.920.3 (2013-11-20)
  • Upgraded to Armadillo release Version 3.920.3
    • fix for handling of tiny matrices by .swap()
Courtesy of CRANberries, there is also a diffstat report for the most recent release As always, more detailed information is on the RcppArmadillo page. Questions, comments etc should go to the rcpp-devel mailing list off the R-Forge page.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

Next.

Previous.