fidzu : debian

Debconf17 has come and gone by too fast, so we all could use a moment looing back at all the fun and serious happenings of the main event in the Debian social calendar. You can find my full photo gallery on Google, Flickr and Debconf Share.

Make sure to check out the Debconf17 group photo and as an extra special treat for you - enjoy the "living" Debconf17 group photo!

21 Aug 2017 6:49pm GMT

On the heels of the very recent bi-monthly RcppArmadillo release comes a quick bug-fix release 0.7.960.1.1 which just got onto CRAN (and I will ship a build to Debian in a moment).

There were three distinct issues I addressed in three quick pull requests:

• The excellent Google Summer of Code work by Binxiang Ni had only encountered direct use of sparse matrices as produced by the Matrix. However, while we waited for 0.7.960.1.0 to make it onto CRAN, the quanteda package switched to derived classes---which we now account for via the is() method of our S4 class. Thanks to Kevin Ushey for reminding me we had is().
• We somehow missed to account for the R 3.4.* and Rcpp 0.12.{11,12} changes for package registration (with .registration=TRUE), so ensured we only have one fastLm symbol.
• The build did not take not too well to systems without OpenMP, so we now explicitly unset supported via an Armadillo configuration variable. In general, client packages probably want to enable C++11 support when using OpenMP (explicitly) but we prefer to not upset too many (old) users. However, our configure check now also wants g++ 4.7.2 or later just like Armadillo.

Armadillo is a powerful and expressive C++ template library for linear algebra aiming towards a good balance between speed and ease of use with a syntax deliberately close to a Matlab. RcppArmadillo integrates this library with the R environment and language--and is widely used by (currently) 382 other packages on CRAN---an increase of 52 since the CRAN release in June!

Changes in this release relative to the previous CRAN release are as follows:

Changes in RcppArmadillo version 0.7.960.1.1 (2017-08-20)

• Added improved check for inherited S4 matrix classes (#162 fixing #161)

• Changed fastLm C++ function to fastLm_impl to not clash with R method (#164 fixing #163)

• Added OpenMP check for configure (#166 fixing #165)

Courtesy of CRANberries, there is a diffstat report. 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 Aug 2017 7:20pm GMT

Post-Solskogen, there hasn't been all that many commits in the main Nageru repository, but that doesn't mean the project is standing still. In particular, I've been working with NVIDIA to shake out a crash bug in their drivers (which in itself uncovered some interesting debugging techniques, although in the end, the bug turned out just to be uncovered by the boring standard technique of analyzing crash dumps and writing a minimal program to reproduce). But I've also been looking at intraframe codecs; my sort-of plan was to go to VideoLAN Dev Days to present my findings, but unfortunately, there seems to be a schedule conflict, so instead, you can have some scattered random notes:

• JPEG is really impressive technology. It's made in 1992, and it's still frighteningly close to state of the art! Beating it without getting a lot slower isn't trivial at all; witness that even with WebP etc., we still don't have a JPEG-killer because it's just so incredibly good. (Granted, in 1992, decompressing a JPEG would easily take half a minute or more.)
• rANS is pretty neat; it gives you almost all the advantages of arithmetic coding, but does without divisions in the decoder (and if you have static probabilities and fast high multiplies, also in the encoder). I still don't feel like I understand it fully, though. And it has the odd property of having to encode and decode in different directions, which may or may not be a problem to you. But ryg's interleaving tricks to get SIMD are pretty nice; I won't claim to fully understand those either. Maybe I will need to eventually.
• GPU performance models continue to bend my head. GPUs may have managed to make a parallel programming model that people actually manage to use, but getting really high performance out of compute shaders is still hard to get use to.
• It really annoys me that Quick Sync doesn't support luma-only encoding (ie. 4:0:0); the best you can do is seemingly 4:2:0 and then have empty chroma planes, which wastes resources. It would have been a nice way to hack around the limitation that you can't have 4:2:2 or alpha.
• Time-frequency switching is the solution to a problem that sounds trivial if you're a DSP beginner and nearly impossible if you've actually done some DSP. I'm not intuitively convinced it's worth it if you have fast float muladds, but it's pretty neat. Allowing multiple block sizes to an intraframe codec has a cost in that you'll need a size decision heuristic, though, which pretty much instantly complicates the encoder.

Work in progress :-) Maybe something more coherent will come out eventually.

Edit: Forgot about TF switching!

20 Aug 2017 7:15pm GMT