Andreas Klöckner's web page

The State of OpenCL for Scientific Computing in 2018

2018-01-16T19:39:00-06:00

As someone who maintains a fair number of software packages that build, in one way or an other, on OpenCL (and keeps creating more!), a friend recently asked me what I thought of the state of the OpenCL ecosystem in 2018. In part, I am writing this to support the notion that it is healthier than one might assume. So, with that: here is my personal, subjective view of how OpenCL is doing.

OpenCL as a Standard

First, to get one thing out of the way, independently of the state of concrete implementations (which I’ll get to in a second), the OpenCL machine model makes a great deal of sense (to me). It provides:

Two heavily overdecomposed levels of parallelism (one for cores, the other for vector lanes), with lightweight (“barely any”) synchronization.
A dependency graph between computations (“kernel invocations”) that are submitted in large batches
Shared virtual memory. This means that memory handles are just pointers in the host address space, and multiple devices and the host can access this memory with potentially strong coherency guarantees. (This is in OpenCL 2, on devices that support it.)
Just-in-time compilation

Even while OpenCL is headed for its tenth birthday, that machine model still maps very cleanly onto the multi-core CPUs and GPUs of today, and if you write scientific computing code with code generation, there is hardly a better, more coherent machine model that lets you target a broad class of machines. So even if OpenCL were to not ‘make it’, the abstraction itself is likely to continue to make sense for the foreseeable future. For instance, OpenCL and CUDA do not differ by much except aside from spelling. OCCA and ispc expose much the same model. A very similar computing environment can be assembled out of OpenMP’s core parallelism and #pragma simd.

In addition, OpenCL-the-standard provides:

A fairly complete set of transcendental functions that, typically, vectorize cleanly
Easy access from Python and interoperability with numpy (OK, this one is a shameless plug)
Compared with many custom JIT solutions, OpenCL has fairly rigorous semantics, so you’re not likely to be left guessing whether an implementation’s behavior is correct or not.

Implementations

A standard is only as good as its implementations—so what is the state of the available OpenCL implementations? OpenCL is a large and complex specification, so implementing it fully and correctly is a substantial effort.

A number of fairly recent developments have helped substantially in that regard:

LLVM’s built-in OpenCL support keeps improving
Khronos’ OpenCL conformance test suite became open-source and publicly available
Khronos’ OpenCL ICD loader has become open-source

Nvidia GPU

Nvidia’s OpenCL implementation is capable and generally works well.

Occasionally, one does get the impression that, as a vendor of a directly competing solution (CUDA), Nvidia has a vested interest in crippling OpenCL support on their devices: from lacking tool support (debugging, profiling) to reports of performance differences from CUDA (in both directions!) to lagging support for newer OpenCL versions, there is plenty that could be better.

That said, Nvidia’s leverage in this regard may be limited—surprisingly.

POCL (see below) has a proof-of-concept CUDA backend that uses Nvidia’s CUDA APIs to implement OpenCL. With some work, this could be made actually useful.
There is ongoing work in the Linux graphics stack to support computation. While there is substantial work to be done, this work has come very far in the past few years.

AMD GPU

AMD’s brand-new open-source rocm stack has just become usable for me on Debian on Radeon R9 Fury GPU, now that they’ve switched to shipping their kernel driver as a DKMS module. While I haven’t done comprehensive performance measurements, it passed most of the PyOpenCL test suite on the first attempt.

POCL CPU

POCL is an LLVM-based OpenCL implementation for (mainly) CPUs. It has just celebrated its 1.0 release. Most of all, it is conformant and correct. My group builds and tests software mainly on POCL, because it is easy to install on basically any machine (see below under ‘Shipping Software’).

Its vectorization performance varies from competitive to ‘needs work’, but in general, code executes at least as fast as reasonably written C or Fortran code. With work out of Intel on an outer-loop vectorizer for LLVM, this has the potential for substantial improvement in the near future. POCL also integrates the SLEEF vectorized special function library.

Intel Integrated GPUs

Intel’s Beignet implements CL for Linux on their integrated GPUs. It passes most of the PyOpenCL test suite and performs well.

Update June 2018: In early 2018, Intel has released an additional OpenCL compute runtime for its “Gen” family of integrated GPUs that supersedes Beignet for Broadwell (“Gen8”) and newer chips. I do not have first-hand experience with this ICD yet, but since it is reportedly based on Intel’s previously closed Windows CL runtime, I expect a mostly competent implementation.

Apple GPU/CPU

While Apple originated the OpenCL standard, the implementations they ship are borderline unusable due to bugs. For CPU work, POCL can be built on macOS with little trouble (or see below under ‘Shipping Software’). For GPU work, Linux provides a nicer environment.

Update June 2018: Apple has deprecated OpenCL in favor of their own graphics abstraction, Metal, that is unavailable on any other platform.

Intel CPU (Updated August 2019)

Intel has a (now open-source) CPU OpenCL ICD that has fairly competitive code generation. Based on file names, it appears that this is an updated version of their previous closed-source OpenCL CPU ICD. One unfortunate aspect of this implementation is that it ships with a number of shared libraries that are very likely to clash with other shared libraries already installed on your system. So, for use of this ICD, make sure to configure (e.g.) the $LD_LIBRARY_PATH environment variable so as to prefer the libraries shipped with the implementation over the system ones. This bug has discussion of this issue.

A previous version of this guide mentioned code generation bugs and crashes. I suspect that was not true, and the crashes I observed instead due to the clashes with system libraries noted above.

Shipping Software

Shipping software based on OpenCL sounds like it might be troublesome, since, in addition to the actual software, a user must have an ICD loader (libOpenCL.so) in addition to an ICD (the actual implementation).

This concern is mostly a thing of the past though. Five easy shell commands install a usable environment on Linux and macOS.

The Future

Recent news out of the OpenCL working group is that OpenCL might become implementable on top of Vulkan, a graphics API that is rapidly gaining ubiquity, by way of a single ‘shim’ ICD provided by the OpenCL working group. While this is dependent on a number of future developments, it would make OpenCL support essentially ubiquitous on all GPUs supporting Vulkan.

Conclusions

While OpenCL has been around for a while, many recent developments make me optimistic for its future. If I missed anything, do let me know, and I would be happy to add it here.

Nvidia Highlights PyCUDA/PyOpenCL Applications

2016-02-14T13:18:00-06:00

Nvidia’s Parallel Forall blog recently highlighted two applications that are using my Python HPC packages to accelerate their computational science workloads:

Increasing the Luminosity of Beam Dynamics with GPUs, a writeup of Adrian Oeftiger‘s work on CERN’s PyHEADTAIL code.
Open, Reproducible Computational Chemistry with Python and CUDA a writeup of Christopher Cooper‘s work on the PyGBe code.

Tutorial: Domain Specific Languages and Code Generation in Python

2015-11-25T18:55:00-06:00

At SC15 last week I had the opportunity to present a tutorial on how to design, build, and compile your own domain-specific language using Python. I am now releasing the tutorial material under a Creative Commons license for the community to use and build on.

Browse tutorial

Specifically, the material covers:

how to define mathematically-oriented domain-specific languages (“DSLs”) in Python
how to build transformations for your DSLs to take them from abstraction to implementation
how to generate highly efficient code from your domain-specific language
how to use just-in-time compilation with OpenCL from Python to execute generated code
a few existing design studies and use cases for domain-specific languages
how to use loopy to generate highly efficient code to work with array data targeting heterogeneous processor architectures (CPUs/GPUs)

A brief, generic introduction to Python and Numpy is also included for those coming from other languages.

Using Github’s rendering of IPython/Jupyter notebooks, it’s quite easy (and hopefully fun) to click around and browse the source repository. A virtual machine image is also available to support experimentation with the software tools being demonstrated.

I am making the material for this tutorial available under a Creative Commons license for others to use and build on.

Dr. Dobb’s article series on PyOpenCL

2013-11-14T16:10:00-06:00

Gaston Hillar has written a very nice introductory article on using PyOpenCL, to be part of a two-part series. I’m already looking forward to the second part. :)

If you’d like to try to follow along with the article, check out the OpenCL installation howto, then follow the easy installation instructions for PyOpenCL, and you’ll be in business. If you find that you need help in the process, help is available.

The Devil’s dictionary of programming

2013-11-02T13:55:00-05:00

As someone who, on occasion, has created a domain-specific-language or two, this entry of the Devil’s Dictionary of Programming hit close to home:

DSL

A domain specific language, where code is written in one language and errors are given in another.

Link (found on hacker news)

New (Py)OpenCL tutorial

2013-09-17T20:10:21-05:00

Simon McIntosh-Smith from Bristol University just let me know that he and Tom Deakin have published a new set of lecture slides and excercises (with solutions!) to teach (more generally) OpenCL and (specifically) PyOpenCL. I’ve added a link to this and a few older tutorials to

PyOpenCL’s main documentation page.

Simon requested that if you spot issues with the tutorials, you file them as issues on github.

Quote of the day

2013-05-03T09:14:38-05:00

Quote of the day:

In Python 2.6, no one can hear you scream.

(source) This reminds me of that one time when I was debugging some odd interaction between threading and contexts in PyCUDA. Never again, hopefully. :)

Job? UIUC!

2013-03-21T14:58:29-05:00

After many long weeks (months, really) spent applying for and searching for the right job, white smoke! (sorry, couldn’t resist) I will be joining Luke Olson, Bill Gropp, Mike Heath and everyone else in the Scientific Computing Group at the Computer Science Department of the University of Illinois at Urbana-Champaign. I’ll be moving out there sometime in July or August. I’m really excited.

Chambana, here I come! :)

GPUs and Python at PyData NYC 2012

2012-11-12T17:42:54-06:00

This October I had the honor of presenting my work on using Python with GPUs at PyData NYC 2012.

There was also a panel discussion on Python+Parallel that I was a part of—great fun.

Also be sure to check out all the videos of the other great talks to see what you’ve missed.

If this sounds interesting to you, also be sure to check out their next conference, PyData Silicon Valley 2013. And please (continue to!) support NumFocus and check out what Continuum are doing for big data in Python. They deserve a lot of credit for bringing the Python community together at events like PyData.

PyCUDA and PyOpenCL at Nvidia GTC 2010

2010-09-28T16:55:56-05:00

Like last year, I had the honor of being invited to present PyCUDA and PyOpenCL along with a few examples of their use to a great crowd at Nvidia‘s inaugural GPU Technology Conference 2010.

Please click the following link to view the slides: PyCUDA and PyOpenCL at Nvidia GTC 2010.

Update: Nvidia has posted a recording of the session. There’s also a full list of sessions, with many talks that are worth being watched. In particular, I’d like to recommend the ones by Bryan Catanzaro on Copperhead, which is built on top of PyCUDA, by Tim Warburton on all things GPU-based discontinuous Galerkin. Also check out the poster on Atomic Hedgehog by Cyrus Omar.

Attachments

main.pdf

PyCUDA at PyCon Quattro (Italy)

2010-05-20T08:34:32-05:00

At the recent PyCon Quattro, which took place in early May in the beautiful Tuscan city of Florence, Fabrizio Milo gave a talk on PyCUDA entitled

PyCuda: Come sfruttare la potenza delle schede video nelle applicazioni python (PyCUDA: How to make use of the power of graphics cards in Python applications)

He made a set of rather nice slides (in English), which may be of interest. They are downloadable in PDF form at the link.

Thanks Fabrizio for taking the time to talk about PyCUDA!

Putting the “Hard” in “Hardware”

2010-02-20T12:15:04-06:00

Quite often, I hear complaints that coding for GPUs is difficult. In response to such comments, I believe that, for correct perspective, the discussion needs to be framed somewhat differently.

First of all, squeezing the last drop of performance out of modern CPUs is hard, too. Here’s a nice article on cache effects by Igor Ostrovsky that explains some of the phenomena one needs to take into account and the surprising things that can happen.

It just appears to me that on the CPU, fewer people care about good performance, whereas for GPUs, you admit that you do care simply by your choice of architecture. Not caring about CPU is not entirely unreasonable—you are somewhat likely to get ‘average’ performance even without detailed analyses. On the GPU on the other hand, carelessly written code is not as likely to perform well.

So, in summary, my belief is that both CPUs and GPUs can be equally difficult to understand, it’s just that the potential payoff of caring about performance is much greater on one than on the other.

Why is GPU Computing significant?

2009-12-05T18:01:35-06:00

In my opinion, GPU computing is significant because I—as a grad student—can easily afford a machine that allows me to perform a simulation like the following in 40 minutes instead of a whole workday. That’s why.

If you’re curious, this shows the density of a vortex shedding flow behind a square obstacle at Re=100 and Ma=0.1. The attentive viewer may notice a sound wave at the beginning as the system settles from uniform flow to flow around the obstacle, as well as the passing of a gentle density “nudge” intended to throw the system off balance and accelerate the onset of shedding. This was computed using my Discontinuous Galerkin solver hedge on an Nvidia GTX 260.

This work owes a lot to Hendrik Riedmann from IAG, Uni Stuttgart who wrote the initial version of the Navier-Stokes operator in hedge.

(Btw: did you notice how the movie cleverly avoids the typical criticism of being “CFD”—colorful fluid dynamics? :-)

You are being preyed on

2009-12-02T22:22:45-06:00

Do you have a free checking account? A free credit card? Do you perhaps even get money back from your credit card company? Ever wonder who pays for all these freebies? This Frontline report (also embedded below) provides some answers and sheds light on the shady things going on in US consumer banking.

Submitted: PyCUDA: GPU Run-Time Code Generation for High-Performance Computing

2009-11-26T10:12:13-06:00

Nicolas Pinto, Yunsup Lee, Bryan Catanzaro, Paul Ivanov, Ahmed Fasih and I have recently submitted an article that explains how PyCUDA allows the user to do run-time code generation (“RTCG”), and how that is an enormous boon to implementation efforts of most high-performance codes. Among many other things, PyCUDA also underlies our efforts to bring discontinuous Galerkin PDE solvers onto the GPU.

Get it while it’s hot: Arxiv, Brown SC

Update: Fixed arXiv link.

Abstract

High-performance scientific computing has recently seen a surge of interest in heterogeneous systems, with an emphasis on modern Graphics Processing Units (GPUs). These devices offer tremendous potential for performance and efficiency in important large-scale applications of computational science. However, exploiting this potential can be challenging, as one must adapt to the specialized and rapidly evolving computing environment currently exhibited by GPUs. One way of addressing this challenge is to embrace better techniques and develop tools tailored to their needs. This article presents one simple technique, GPU run-time code generation (RTCG), and PyCUDA, an open-source toolkit that supports this technique.

In introducing PyCUDA, this article proposes the combination of a dynamic, high-level scripting language with the massive performance of a GPU as a compelling two-tiered computing platform, potentially offering significant performance and productivity advantages over conventional single-tier, static systems. It is further observed that, compared to competing techniques, the effort required to create codes using run-time code generation with PyCUDA grows more gently in response to growing needs. The concept of RTCG is simple and easily implemented using existing, robust tools. Nonetheless it is powerful enough to support (and encourage) the creation of custom application-specific tools by its users. The premise of the paper is illustrated by a wide range of examples where the technique has been applied with considerable success.

PyCUDA talk at Nvidia’s GPU Technology Conference

2009-10-04T16:20:11-05:00

This past week, I had the honor of presenting a talk on PyCUDA at Nvidia‘s inaugural GPU Technology Conference.

Please click the following link to view the slides: PyCUDA talk at Nvidia’s GPU Technology Conference.

Update: Nvidia has posted a recording of the session that you may watch or download.

Update 2: Giancarlo Colasante has transcoded the above video into just 16 MB. You may download the resulting video here.

Attachments

pycuda-nvidia.pdf

Nena vs Eminem

2009-09-05T10:27:14-05:00

Is it possible to take the instrumentals from one track, the vocals from another, and come up with something that you’d actually want to listen to? Turns out yes:

Amazing what computers can do these days. (via Holger Levsen)

Original 1, Original 2.

SciPy‘09: Advanced Tutorial on PyCUDA

2009-08-25T11:23:03-05:00

The SciPy‘09 conference ended less than a week ago. At the invitation of the SciPy‘09 organizers (especially Fernando Perez), Nicolas Pinto gave a talk in the Advanced Tutorials track on how to use PyCUDA to do GPU scripting.

First, I would like to use this opportunity to publicly thank Nicolas for all the work and time he put into making this tutorial a reality. Second, I would like to point out the video of his session, which you can watch below:

Introducing UIH: Math + Unicode, Happy Together

2009-06-10T21:04:22-05:00

As a math person, you’re often faced with the task of communicating about math. Unfortunately, most modern means of communication, be it email, the web or instant messages, aren’t really suited to typing math. Fortunately, however, many of these means do allow the use of Unicode, and Unicode allows for certain limited forms of mathematical typography.

Putting Unicode formulas together usually requires a fair amount of patience and some quality time with your favorite character map application. But now there’s an easier way: The Unicode Input Helper—or “UIH”. Here’s an image of it in action:

Using it, you may use HTML entity names with backslashes (such as \int for an integral) to put together the basics of a formula, and then use a searchable list of all known Unicode characters to add the finishing touches. The screenshot gives you an idea. Once you’ve finished your masterpiece, simply use your computer’s copy-and-paste function to get it to where it needs to be.

Download the most recent version of UIH

(Like PuDB, uih requires Ian Ward’s urwid library.)

As a side benefit, I think uih makes for a nice replacement for pretty much every character map program—but its original purpose was easy typing of math.

Debug Python in Style

2009-06-09T18:39:06-05:00

I’m happy to introduce PuDB, a full-screen, console-based visual debugger for Python that I recently cooked up.

Download it here

Or install it simply by typing

easy_install pudb

into your Unix shell. Here’s a screenshot of it in action:

Python has had decent debugging support for a while now, in the form of

Pdb, the command-line debugger that comes with the interpreter
Winpdb, a very nice GUI debugger
IDEs such as Eric

But I felt that there was a gap between these offerings—Pdb being very austere, and Winpdb and the IDEs being rather heavyweight. I wanted a comfortable debugger that’s easily usable in a shell and doesn’t require me to touch my mouse. PuDB uses Ian Ward’s excellent Urwid library for its interaction with the console.

Update: Looks like PuDB is slowly growing a community. There’s now a mailing list to host discussions.

Git Commit Timelines

2009-06-01T14:39:18-05:00

Here’s a waste of time I could just not resist: MIT‘s Project SIMILE has released a rather nice Timeline Web Widget. Combine that with GitPython and a bit of glue code in Python, and you get a neat, one-page summary of the software work I’ve done as part of working towards my PhD. Here’s a screenshot if you’re too lazy to click through:

Note that (nearly) every commit is hyperlinked to the git web frontend, where you can view what was actually changed—just click the title text in the popup bubble.

If you’d like to do this yourself, the code is available, you can get it with

git clone http://git.tiker.net/trees/commit-timeline.git

1 Billion seconds on this Earth

2009-04-13T12:45:43-05:00

Happy 10⁹-day to me!

date -d'NNNN-NN-NN + 1000000000 seconds'
Mo 13. Apr 01:46:40 EDT 2009

World Builder

2009-04-04T13:01:27-05:00

From the department of “what’s Andreas wacthing on Vimeo“: Here’s something amazing you can apparently do with one day of filming and two years of post-production.

World Builder from Bruce Branit on Vimeo.

One thing I could especially relate to about this video is the feeling of time-constrained emptiness at the beginning… If you’ve tinkered with creating video games, you probably know what I’m talking about. Research is somehow not too different.

Credit Crisis? What Credit Crisis?

2009-04-04T11:01:52-05:00

If you haven’t been living under a rock, you’ve probably heard about the global economic meltdown. You’ll probably have also listened to people explain it, maybe even understood what’s going on. Regardless, this video I recently found sums it up again very nicely:

The Crisis of Credit Visualized from Jonathan Jarvis on Vimeo.

boostmpi: Initial Release

2009-02-11T11:28:36-06:00

I’m happy to announce the initial public release of boostmpi, an MPI wrapper for Python.

boostmpi is a high-quality Python wrapper around the Message Passing Interface (MPI). MPI is a standardized interface to libraries such as OpenMPI and MPICH that provide high-performance inter-process communication for distributed-memory computing.

boostmpi uses the Boost.MPI library, which gives MPI a very usable C++ interface. This C++ interface is then made accessible to Python via the Boost.Python library.

boostmpi was originally distributed as part of the Boost C++ library. This separate distribution aims to make the software more accessible.

Disclaimer: boostmpi (born as Boost.MPI.Python) is the work of Doug Gregor. I just do maintenance and distribution work around here.

CodePy: Initial public Release

2009-02-05T00:58:50-06:00

I’m happy to announce the initial public release of CodePy, a native-code Metaprogramming toolkit for Python.

The idea is the following:

You build a string that is actually C code.
CodePy compiles it, links it, and dynamically loads it into the Python interpreter.

Ta-da: Instant native-code performance, with no explicit pre-compilation step. CodePy can also help you with the generation of the code, by letting you build a data structure and turning that into C code. CodePy works well with Boost.Python and PyCuda. In conjunction with the latter, it offers an easy way of doing GPU Metaprogramming, a very effective tuning technique for the GPU.

Talk slides: PyCuda@MIT

2009-01-28T14:14:08-06:00

Nicolas Pinto at MIT was nice enough to invite me over to give a talk in their CUDA class. I made a bunch of slides that I think are of general interest to people who are interested in PyCuda. You can find them here.

Submitted: “Nodal Discontinuous Galerkin Methods on Graphics Processors”

2009-01-27T19:19:47-06:00

Tim Warburton, Jeff Bridge, my advisor Jan Hesthaven and I have recently submitted an article detailing our efforts to accelerate Discontinuous Galerkin computations by using Nvidia CUDA GPUs. DG seems to be a good fit for these machines.

Get it while it’s hot: Arxiv, Brown SC reports, JCP

Abstract

Discontinuous Galerkin (DG) methods for the numerical solution of partial differential equations have enjoyed considerable success because they are both flexible and robust: They allow arbitrary unstructured geometries and easy control of accuracy without compromising simulation stability. Lately, another property of DG has been growing in importance: The majority of a DG operator is applied in an element-local way, with weak penalty-based element-to-element coupling.

The resulting locality in memory access is one of the factors that enables DG to run on off-the-shelf, massively parallel graphics processors (GPUs). In addition, DG’s high-order nature lets it require fewer data points per represented wavelength and hence fewer memory accesses, in exchange for higher arithmetic intensity. Both of these factors work significantly in favor of a GPU implementation of DG.

Using a single US$400 Nvidia GTX 280 GPU, we accelerate a solver for Maxwell’s equations on a general 3D unstructured grid by a factor of 40 to 60 relative to a serial computation on a current-generation CPU. In many cases, our algorithms exhibit full use of the device’s available memory bandwidth. Example computations achieve and surpass 200 gigaflops/s of net application-level floating point work.

In this article, we describe and derive the techniques used to reach this level of performance. In addition, we present comprehensive data on the accuracy and runtime behavior of the method.

Boost Numeric Bindings 20081116

2008-11-16T14:05:19-06:00

I’ve just rolled release 20080816 of the bindings, with the following stuff changed:

Fix asserts in gbsv.hpp as suggested by Vardan.
Sync Thomas’s remaining fixes.
Apply patch to gels* from Naoki at http://mathema.tician.de/node/391.
Remove duplication in lapack_names.h.
Fix line endings in Jesse’s gelsd stuff.
Add obvious improvements from current bindings svn.
Modify sync procedure even more.
Apply Thomas’s sync patch.
Update boost update procedure.
Apply Thomas’s svn sync patches.
Karl’s geev fix.
Followed Thomas’s idea of sygv->hegv.
Reduce diff with svn, step 4.
Reduce diff with svn, step 3.
Reduce diff with svn, step 2.
Reduce diff with svn, step 1.
Fix diff command.
Install to boost-numeric-bindings.
Bart Janssen’s UMFPACk bug fix.
Add missing checks for BOOST_NUMERIC_POOR_MANS_TRAITS in blas1.hpp.

I’m happy to report that the diff with the official bindings has shrunk to just Vardan’s gbsv and Jesse’s gels* stuff. I think that’s pretty great, and in no small part due to Thomas’s work on getting stuff merged.

PyCuda 0.91

2008-11-12T13:16:34-06:00

I’m happy to announce the availability of PyCuda 0.91. There is full, up-to-date documentation available.

The following exciting stuff is in PyCuda 0.91:

Support for Windows and MacOS X, in addition to Linux. (Gert Wohlgemuth, Cosmin Stejerean, Znah on the Nvidia forums, and David Gadling)
Support more arithmetic operators on pycuda.gpuarray.GPUArray. (Gert Wohlgemuth)
Add pycuda.gpuarray.arange(). (Gert Wohlgemuth)
Add pycuda.curandom. (Gert Wohlgemuth)
Add pycuda.cumath. (Gert Wohlgemuth)
Add pycuda.autoinit.
Add pycuda.tools.
Add pycuda.tools.DeviceData and pycuda.tools.OccupancyRecord. pycuda.gpuarray.
GPUArray parallelizes properly on GTX200-generation devices.
Add support for compiling on CUDA 1.1. Added version query pycuda.driver.get_version(). Updated documentation to show 2.0-only functionality.
Make pycuda.driver.Function resource usage available to the program. (See, e.g. pycuda.driver.Function.registers.) Cache kernels compiled by pycuda.driver.SourceModule.
Allow for faster, prepared kernel invocation. See pycuda.driver.Function.prepare().
Added memory pools, at pycuda.tools.DeviceMemoryPool as experimental, undocumented functionality. For some workloads, this can cure the slowness of pycuda.driver.mem_alloc().
Fix the memset family of functions.
Improve Error Reporting.

Check the docs change list for a fully hyperlinked version of the above.

Have fun, Andreas

Brown Beamer Template

2008-06-21T18:58:09-05:00

If you’re anything like me, you go to Brown University and really like Till Tantau’s Beamer LaTeX package for making slides. Unfortunately, Beamer’s theming can be a bit obstinate, and figuring out how to give your slides a Brown flavor (yum!) can be tedious.

Luckily for you, I just spent some time and put together a template that you can use for exactly that purpose. This is what it looks like:

I’ve also attached a sample PDF to this story.

Download

You can download a tarball of the most recent version or browse the source code. If you like, you can also directly get a git checkout:

git clone http://git.tiker.net/trees/brown-beamer.git

Instructions

To use this template, obtain the source tree, edit slides.tex, then change to the out directory and type ./make.sh. (If you’re on Windows, maybe you can send me a batch file that does the same thing.)

Attachments

slides.pdf

Reproducibility in Scientific Computing

2008-06-20T21:29:10-05:00

In this recent paper, Finite Volume guru Randall J. LeVeque addresses an issue that is near and dear to my heart: Computational Scientists must start thinking about the reproducibility of their results.

Scientific results obtained by experimentation are expected to be reproducible. This has long been demanded of published results in, for example, physics or chemistry. Yet, scientific computing, while often experimental, has been strangely exempt from this requirement. This ought to change.

In his paper, LeVeque formulates several demands to be placed on scientific codes:

Availability in source code form. Codes used to obtain results in published research must be available for other researchers to inspect. While LeVeque is not as clear about this point, I would like to emphasize that this availability should be of the same kind as that of the paper advertising results obtained by the code. In most cases, this means that the code should be downloadable from the same place as the paper. (There should not be access restrictions on the paper either, but that’s a different issue.)
No dependencies on for-pay software. For widest availability and reproducibility, it is preferable to not have scientific codes depend on commercial, for-pay software. That notably includes Matlab.
Readability. Codes should be written mostly to be read, not so much to be executed. We need to grow an understanding that code is a valid expression of the ideas that it contains. The more apparent these ideas are from the code, the better.

Given these constraints, he arrives at virtually the same toolset that I’ve been using for a while now, consisting of Python, numpy, matplotlib and VisIt.

I am hoping that this issue gets the attention it deserves. I’ll try to do my part.

Joint release 0.92 of PyUblas and PyUblasExt

2008-04-20T18:33:04-05:00

I’m happy to announce the availability of PyUblas and its companion package PyUblasExt in version 0.92. While PyUblasExt is available for the first time in released form, these are the changes made to PyUblas:

Sparse wrappers are now not compiled by default. You have to request them using --with-sparse-wrappers at configure time.
Provide implicit to-/from-python conversion for Ublas’s vector, bounded_vector, and matrix types.
Provide from-Python converters for Numpy’s array scalars.
Instantiate converters for a much wider range of element types.
Fixed compilation without -DNDEBUG.
Plus a slew of other cleanups and fixes.

TagPy 0.94.1

2008-04-20T12:11:53-05:00

Responding to a patch from Bill Nottingham, I’ve uploaded TagPy 0.94.1. It fixes building TagPy against TagLib 1.5 and also fixes a mis-wrap of the clear method of the Map type.

In case you were wondering, 0.94 was a dud, but by the time I realized that, I had already uploaded the tarball to the Python Package Index, which made it unchangeable. :(

Additionally, I’ve decided to only upload new releases of TagPy to its page in the Python Package Index. Easier maintenance, yay! :)

Release 0.91.2 of Synoptic

2008-04-20T11:32:36-05:00

Yay, here I go again. (Update: and again) There were a few small wrinkles in Synoptic 0.91 (and 0.91.1) for which I felt it warranted to push out a followup release, aptly called 0.91.2. It fixes the following things:

Depend only on Python 2.4 by reverting to Python Paste’s WSGI httpd.
Speed up page load by serving all javascript as one big clump.
Change JS/CSS load order to fix history slider size.
(new in 0.91.1) unbreak daemon mode in Python2.4
(new in 0.91.1) launch browser from background thread, to avoid deadlock.

Download from PyPI

First real public release of Synoptic

2008-04-19T16:18:58-05:00

I’m happy to announce the first truly public release of Synoptic, my own web-based dynamic note taking application. If you find yourself lost in a jumble of either paper notes, text files on a hierarchical file system, or an overgrown Tomboy wiki, and find yourself craving order, categories, and simplicity, Synoptic may well be for you.

So go check it out! :)

Boost Bindings 20080405

2008-04-05T20:28:54-05:00

I’ve rolled another new Boost bindings snapshot, dated 20080405.

This time around, I’ve done a bit more than update to current svn. I’ve tried to scavenge past contributions from the Ublas mailing list that got dropped on the floor. I’ve created a new version control tree where I intend on maintaining these changes:

http://git.tiker.net/?p=boost-bindings.git;a=summary

For starters, I’ve picked up:

Jesse Manning’s gels* routines.
Vardan Akopian’s gbsv routines.
Georg Baum’s non-std::complex<> fixes.
Georg Baum’s Fortran naming override.

If there’s anything I’ve missed, or if you have a contribution to the bindings sitting on your disk waiting for submission, please email me now! The bindings are a tremendous resource, let’s try to not let them wither away.

On Hybrid Scientific Codes, Part I: The Idea

2008-04-05T13:37:45-05:00

How should scientific codes be constructed? Scientific codes are often experimental in nature, and therefore need to be easy to change, and, of course, quick to develop in the first place—who wants to invest any more than necessary into an idea that might not work out after all? On the other hand, computational scientists are an impatient bunch: Once the method stabilizes, or it becomes clear that it is generally useful, more and more demands get placed on a prototype code, most of all speed. It is clearly desirable to respond to these demands in an evolutionary manner. Why rewrite an entire code, possibly in a different language, if only one small part is responsible for most of the run time?

One answer that I have found to address these problems well is what I call hybrid development. Nearly all my code these days is “hybrid”. Now what does that mean? That means, I start developing in a high-level language that is quick to modify and separates me from the more menial tasks of coding, allowing me to code and test an algorithm quickly. The cost for this quick development turnaround is that the initial prototype is often fairly slow because many small operations are performed in the (often interpreted, instead of compiled) high-level language. This is an effect that everybody who has developed larger-scale code in MATLAB or similar environments has seen. Often, the bulk of the run time is concentrated in 10% or less of the entire code. In this case, the sensible solution is to replace the slow part of the high-level code with a lower-level compiled implementation, while keeping as much of the existing high-level code intact. This is the idea of hybrid development.

MATLAB supports this idea to some extent with its MEX mechanism. There are however many more powerful tools on the market (and available for free) that make hybrid development easier and faster by orders of magnitude. In a future installment of this mini-series, I will highlight the software environment I use for my hybrid codes, and point out what I believe its advantages over other solutions are.

Getting the Most out of the Home Row

2008-02-25T21:59:18-06:00

If you’re willing to exchange a bit of memory for a lot less strain on your wrists, let alone work speed, this is your lucky day. In this post, I will describe how I’ve tuned my everyday work environment to a point where it feels completely unnatural for me to have to move my hands away from the home row. And that’s a good thing—the less often my hands have to go anywhere, the less I am at risk of having painful wrist trouble later in life.

So what’s the are the keys to all these benefits? You might laugh, but in fact they are H, J, K, and L. Namely, the cursor motion keys in Vi. It may seem strange to tell people to use H to move to the left, J to move down, K to move up and L to move right, when there are these handy arrow keys right next to where your hands typically sit. But the problem is that your hands have to move quite a ways, and if you’re like me, you also flex your wrist somewhat unnaturally just to reach those. On the other hand, HJKL are right there on the home row, if you touch type. No wrist movement required.

Now, much of what you do every day on your computer involves moving around. So why not make use of these convenient motion keys in as many places as possible? Read on for a slew of suggestions. Text Editor ======

First of all, if you aren’t already a convert, you might want to consider switching to Vim as your text editor. The learning curve may be steep at first, but with time you’ll become more of a text surgeon than you’ll realize, and many rather complex text editing tasks will commit to muscle memory. Delete a word? “dw” Change a word? “cw” Delete a line? “dd” Change a line? “cc”. If you edit text a lot, this is time well spent.

While on the topic of Vim, here are a few tricks that I find valuable:

Use * to search for the word under the cursor.
Use Control-n to search for completions of the word under the cursor. This is especially worthwhile if you write code with rather long, descriptive names.
Vim requires you to hit “Escape” a lot. That’s another nasty twist of the wrist, until you discover that Control-[ does the same thing.

Window management

If you use multiple desktops, find a way to configure your window manager to move from desktop to desktop by using Control-Alt-VI Motion Key.
One big use of the mouse is moving windows you just opened to the right place. However, with the following patch, the window manager Compiz can move (and resize) windows around using Vi motion keys.

Internet

To me, this is today’s big discovery: The Vimperator. This makes Firefox look and behave a lot like Vim. What this means is that you can use a web browser without ever moving your hand to the mouse. Important problem: How do you click on a link then? Vimperator’s solution, to me, is pure genius. By hitting F, every link on the page shows a little one- or two-letter shortcut (such as “HN”). Type that, and you’ve followed the link. Shift-H (remember, that’s Vi-Left) and Shift-L (that’s Vi-Right) allow you to navigate through history. If you know Vim, you’ll feel right at home in Vimperator.
Also, Google Mail and Google Reader support Vi motion keys, once you enable them.

Desktop environment

I use tabbed X terminals a lot. Alt-h and Alt-L make nice shortcuts for switching to the tab on the left or on the right. If you use a GTK-based terminal, you might find yourself unable to change your keyboard shortcuts until you use this trick.
Lastly, I find that Control-Alt-semicolon (at least on a US keyboard) makes a nice shortcut for opening a new terminal.

May your wrists find this information useful! :)

Barf Detergent and other Successes of Modern Marketing

2008-02-05T22:41:38-06:00

I had a good laugh off this list of unintentionally offensive product names. As the title hints, Barf detergent is up there. But there’s one name that takes the cake. It hails from Germany, as it had to. To be precise, the electronics firm Trekstor decided to name the black-colored addition to their “iBeat” line the Trekstor i.Beat Blaxx.

The name sounds convoluted enough that, no doubt, it took lots of effort-by-committee to find it. And all of that just to get the whole thing catastrophically wrong. Sounds somewhat typical of Germany to me…

Hey, Americans!

2008-02-05T08:51:09-06:00

If you’re choosing between Obama and Hillary, here is a compelling argument brought forward by someone who knows what he is doing. (transcript)

I couldn’t agree more. Go ahead, spread this! :)

Forums!

2007-12-05T13:06:34-06:00

Since I’ve received a number of calls for help with regard to my software packages, in particular MeshPy and PyLinear, I’ve decided to create a public forum for these requests for help. Here they are: the MeshPy forum and the PyLinear forum.

Here’s to years of useful discussions! ;)

Visualizing Greenhouse Gases

2007-12-02T10:16:32-06:00

Have a friend who does not believe CO₂ emissions are real and may be harmful? Show them this:

TagPy 0.93

2007-10-24T23:53:57-05:00

I’ve pushed out yet another version of TagPy. This time it is to mop up after some rather unhelpful default in TagLib. That is, unless you have the magic lines:

import tagpy.id3v2 as id3v2
id3v2.FrameFactory.instance().setDefaultTextEncoding(tagpy.StringType.UTF8)

in your code, the generic API tag.artist = u"*some funky unicode string*" will write Unicode tags as Latin1 if the previous ID3v2 tag was also Latin1-encoded. This will of course destroy your carefully crafted unicode string. Therefore I urge all users to include these lines in their code, in order to avoid such trouble.

All that’s new in 0.93 is the wrappers that make this call possible.

I tought about making the above code the default for TagPy, but I wasn’t sure whether that was the right thing to do, seeing as how this default would deviate from TagLib.

Yuck!

2007-10-11T21:08:57-05:00

1,662,780How Many Germs Live On Your Keyboard?

(xkcd is great.)

Möbius Transforms

2007-10-02T23:10:04-05:00

Complex analysis on YouTube:

Yup, too cool not to share. :)

TagPy 0.92 released

2007-10-02T10:10:34-05:00

In order to fix Debian bug #438556 (a crasher that made almost all of the ID3v2 infrastructure useless), I have just release version 0.92 of tagpy. Go get it while it’s hot.

Not enough Flash video yet?

2007-09-26T22:12:42-05:00

Then go watch something sad and then laugh.

PSI Smilies in Gaim/Pidgin

2007-09-25T16:45:06-05:00

Thanks to Debian, I got pidgin 2.2.0 today, and with it, its new atrocious smiley theme—eyebrows and all. This prompted me to make a smiley theme out of Psi‘s old smileys, which I still remember liking the most.

You can download that theme here: PSI Smilies in Gaim/Pidgin

And this is what they look like:

Attachments

Fermat’s Last Python Script

2007-09-18T10:58:45-05:00

def fermat(n):
    """Returns triplets of the form x^n + y^n = z^n.
    Warning! Untested with n > 2.
    """
    from itertools import count
    for x in count(1):
        for y in range(1, x+1):
            for z in range(1, x**n+y**n + 1):
                if x**n + y**n == z**n:
                    yield x, y, z

Nicked from the earthbound kid.

What they don’t tell you about VTK XML binary formats

2007-09-18T09:52:10-05:00

You might have been trying to implement the new-style VTK XML file formats. And you might have realized that what they tell you in the file format “specification” is just incomplete. Well, here are a few bits of information missing from there:

Binary Blob Header: In front of every binary blob, base64 or raw-binary, appended or not, there is an UInt32 length indicator. If you do not have this length indicator, you might get error messages like

Cannot read cell offsets from XXXX in piece 0 because the "offsets" array is not long enough.

Note that if you are encoding in base64, that length header must be encoded separately, so that the end result looks like XXXXXX==XXXX… (note the two equals signs indicating the early end of the length header block).

Compression Header: If you use compression, the following header applies instead:

struct {
  uint32 blocks; 
  uint32 blocksize; 
  uint32 last_blocksize; 
  uint32 compressed_blocksizes[];
};

Again, this header is encoded separately from the data. This was figured out by Thomas Svedberg. Thomas adds that the following works for him:

struct {
  uint32 blocks = 1
  uint32 blocksize = total data size
  uint32 last_blocksize = total data size
  uint32 compressed_blocksizes[] = compressed data size
};

Offset Field Meaning for base64: The “offset” field on DataArrays with format=”appended” is not a binary offset, it is the base64 character count from the underscore that starts the AppendedData section.

Vector Element Order: If you are encoding vectors (i.e. anything with NumberOfComponents=”3”), and suppose your vectors are X,Y,Z,…, then the encoded data stream has the ordering X[0], X[1], X[2], Y[0], Y[1], Y[2], Z[0]…

Books, Books and more Books

2007-09-07T22:55:55-05:00

I’ve found it sad for a long time that I’ve read many books in my life, but my memory of them is slowly fading, and I would at least like to keep a list around. As you have probably guessed, you would not be reading this if I hadn’t stumbled on a cool site that lets me easily put together a catalog of books that own or have read.

And that site is: http://www.LibraryThing.com. And if you care what kind of books I read and own, here’s my library. That list is still pretty incomplete, but I hope to add more over time.

Singing the praises of

2007-08-29T10:07:13-05:00

Here’s a chicken-and-egg problem for you. And you can be part of the egg. I’ll show you how in a minute.

I bet you’ve noticed that bajillions of websites have little links that say “next” and “previous”, right? Like previous page, previous item, previous whatever. And sometimes there’s “up”, too. As in “up” to the parent directory, “up” to the category containing this item, and so on. And every time you come upon a web page like that, you ask yourself, “so where is the link that takes me to the ‘next’ page?” Wouldn’t it be nice if that button was always in the same spot, preferably on your browser user interface?

You might be surprised to learn that a solution to this already exists. The HTML standard specifies the <link rel=”next”> and <link rel=”prev”> header tags. But how can you make them show up in your browser? Easy: Use the Link Widgets Firefox extension. By installing and using that, you’ve become part of the egg. Now the chicken is that more sites need to start using the <link rel=”…”> tags. There are already enough of them to make the extension worthwhile, for example Google, or any Drupal site (such as this one :) ), or anything generated by latex2html, which includes surprisingly large swaths of web documentation. But the more people have the UI to use these links, the more sites will include the markup, making the whole thing more and more useful over time.

And making the web more useful is not a bad thing, is it? ;)

New release of Boost.Bindings tracking Boost 1.35 CVS

2007-08-26T13:11:02-05:00

I’ve rolled a new release (20070826) of the Boost.Bindings library to track what’s currently going on in the Boost CVS tree leading up to release 1.35. What’s released works with current CVS as of 8/26/2007, I assume it will also work with the final release. I suspect that it will not work with boost 1.34.1 and older.

Additionally, this release has a configure script and a Makefile to enable it to be installed with the familiar ./configure;make;make install procedure. Using this is completely optional, the library will also work without installation, as before.

Dmn spmmers

2007-08-26T13:05:39-05:00

Ok, the spammers yet again figured out how to circumvent my anti-spam measures. So yet again, I’ve disabled anonymous commenting. Only this time it’s final. And I have a mess of seven bajillion comments to clean up.

I’m just so sick of these people: :barf:

Trip to the Northwest

2007-08-20T15:21:04-05:00

For the first two weeks of August this summer, Frank and I went on a two-week trip through the northwestern United States. The trip was nothing short of amazing to me, and I’m having a hard time adjusting back to the dull (by comparison) life in Providence. We saw some really amazing nature, and had tons of fun. Naturally, there are pictures available. Thanks to Google Maps, you can also take a pretty precise look at our route: (You have to zoom out a couple times to see the whole thing.)

View Larger Map

Yes, we did indeed drive all that. 3300 miles altogether.

Here’s a brief summary what we did each day:

August 1

Arrive at PDX. To me, this was a flight that visited every corner of the US. (by going through PVD->BNA->LAS->PDX. Thanks, Southwest. :P Anyone wondering why I arrived with a headache?)
Shop at Target.
Find camping in Scappoose, WA.

August 2

Shop at Fred Meyer in Scappoose.
Establish that coffee with each breakfast is a good idea.
Establish that dry milk with cereal is not a good idea, but a necessary evil.
Repack the car to make stuff more accessible.
Drive along US-30 to the WA coast, including Cape Disappointment. Cross the Astoria bridge.
See Mount Saint Helens from a distance.
Return inland and crash at Mt. Rainier for the night, seeing a photo presentation at the campground in the evening.

August 3

Go hike the Skyline trail at Mt. Rainier in total fog. At first, that was fairy-tale lovely (while there was still vegetation and animal life to gawk at), but as soon as there were only snow fields and dirt, that was pretty boring.
Drive the National Park road, and actually see Mt. Rainier.

August 4

Leave Mt. Rainier for Seattle.
Stop at a suburban McDonald’s to score some wireless. Book a kayak trip for 8/6. Book a hotel for Seattle.
Check into said hotel, take well-deserved showers.
Head into Seattle, find parking way outside town, about 15 minutes walk from the Space Needle.
Ride the Seattle Monorail, see Pike Place Market, find it underwhelming. Walk underneath WA-99 to find Seattle’s “waterfront”. Ride the ferry to Bainbridge Island and back, see Mt. Rainier in the distance.
Visit Pioneer Square, find it underwhelming. Go eat African food.
Return to the hotel, getting stuck in traffic for about an hour (at 11pm, no less).

August 5

My birthday, yay. :)
Get up bright and early, shower! ;)
Leave hotel room, taking parts of door frame with us. (Well no, the door had a latch that was supposed to keep the bad guys from coming in. But when I opened the door without opening the latch, it simply pulled the door frame off the wall. Real secure, that.)
Figure Seattle’s got not too much more to see. Move kayak trip from 8/6 to 1pm today, leave Seattle via I-90 East at 9am.
Arrive in Leavenworth, WA at River Rider, ready for a cool kayaking trip by eating the provided (yummy) BBQ.
Have tons of fun kayaking down the Wenatchee in inflatable duckies.
Drive onward down the Wenatchee river, pretty. ;)
Reach the Spokane KOA by nightfall.

August 6

Leave Spokane, WA.
Go shopping at Safeway in Cœur d’Alene, ID. God, Safeway is awful.
Eat lunch at Cœur d’Alene lake, pretty.
See wildfires somewhere in Montana.
Reach the Fish Creek campground at Glacier NP around 6pm.
Go swimming in Lake Dogulas.

August 7

Hike the Loop Trail up to Granite Park Chalet. Be deadly afraid of bears, yell “Hallo” all the time.
Pause at Granite Park, continue on to Logan Pass.
See some white mountain goats (?) and a bighorn sheep (yes, they’re called that for a reason.). Be even more scared of bears. :)
Arrive at Logan Pass, take Glacier’s cool new shuttle back to the Loop.
Very cool, but very exhausting hike. I was longing for a shower and convinced Frank to go find pay for showers at the West Glacier KOA. Also illegally used the hot tub there for a while. (So sue us. ;))
Drop into bed nice and early.

August 8

Drive the Going-to-the-Sun Road to the eastern exit of Glacier, shoot plenty of pictures.
Leave Glacier NP on US-89, follow that to the entrance of Yellowstone, seeing endless prairies, Great Falls, MT and Lewis & Clark National Forest, which itself is really pretty.
Arrive at Yellowstone by nightfall, only to be told that all park campgrounds are full.
There’s a commercial campground with all the bells and whistles, and a government campground with the bare necessities (vault toilets) to choose from. Disagreement arises over which to pick. Frank gets his way, and we live comfortably with showers, running water, laundry, electricity at the site and gasp wireless internet.
Make noodles in celebration.
Tease Frank for being a capitalist.

August 9

First day in Yellowstone. See Mammoth Hot Springs, find them stinky and unremarkable. While they might’ve looked way awesome int the past, they don’t so much when there’s not much water.
See the Golden Gate, Sheepeater Cliff and Norris Geyser Basin. Weather turns to the usual National Park Weather in Norris.
Head across to Canyon Village, hiking down to every viewpoint at the (very cool) Grand Canyon of the Yellowstone. The last one was best, but also the most strenuous, involving stairs: Uncle Tom’s viewpoint.
Ate dinner in the park, return to campsite via Tower in pitch darkness. Think about running over grizzlies in the car.

August 10

Second day in Yellowstone.
Drove into the park via Tower and Canyon village, seeing Tower falls on the way in. A bit further down the road, there was the Mud Volcano area, and much to our surprise, there were bisons all over it. One crossed the walkway about 10 meters ahead of us. Lots of signs that say, “more people get hurt by bisons than grizzlies.” But bisons just look like big cows. They must be harmless. :)
Wildlife jam a little bit down the road, bison walking across one car ahead of us.
Lake Yellowstone’s pretty, too. A crow (raven maybe?) came really close to us while we ate lunch and showed no fear of people at all.
Finally arrived at Old Faithful, saw it go squirt (which is actually more impressive than it sounds here, definitely worth seeing—even though that’s where the crowds are).
Walked around the geyser basin at Old Faithful. More of the same, but still worth seeing.
Ate dinner somewhere at Firehole creek

August 11

Saw Midway Geyser Basin on the way out of Yellowstone to the south. That one is actually not more of the same, it’s really impressive, big and blue. And not stinky.
Drove through Grand Teton National Park. Went swimming in Jackson Lake.
Drove through Jackson, WY. Tries to look like an old western town, doesn’t really fail miserably. Somewhat Touri-Nepp. :)
Drove through more of WY and ID. Saw Bear Lake by nightfall, ate a cold can of ravioli for dinner.
Reached UT, got gas, saw wildfires over Salt Lake City. Hit I-80 west towards Tahoe.
Played Battle for Wesnoth, Gianna Sisters and many other things while driving.
Found a parking lot, slept in the car at 4am.
900 miles in a day. Dude. Far.

August 12

Licked our wounds from the long overnight drive. Found the Tahoe State Recreation Area campground, way cool—with hot (pay) showers and a cool attendant.
Drove around Lake Tahoe.
Saw the Stream Profile Chamber at the Taylor Creek visitor center just south of pretty Emerald Bay.
South Lake Tahoe’s not so pretty.
And the casinos start right when you get to the NV state line—in a town sensibly called Stateline, NV. Yay for inventive city naming. :)

August 13

Mountain biked the Flume trail from Incline Village. Very cool. Very exhausting. :D
Ate ribs and steak at a local restaurant for dinner. Actually, the ribs were too much. Took them home in a box.
Fell into bed dead-tired.

August 14

Kayaked around for four hours on Lake Tahoe starting from Tahoe City. Covered quite a bit of distance, and ate lunch out on the lake—ribs on the water taste twice as good as ribs on land, even if they’re cold. Yum. :) Also got horribly sunburned, because I didn’t put sunscreen on my legs. They’re still peeling as of 8/29. Bad. (But still worth the fun.)
Drove to Mountain View and ate dinner with Hendrik at the Google cafeteria.
Hung out with Hendrik, played DDR and looked at our photos on his bedroom wall—using his projector.

August 15

Repacked the car in preparation of my flight out that night.
Breakfast again thanks to Google, Inc. :)
Drove south on 101 to Santa Clara to see a bunch of logos on building walls, including Intel, Yahoo, Sun.
Drove back north to see Stanford.
Continued towards the Golden Gate on I-280, drove to the Marin vista point and back, walked halfway across the bridge. Foggy, cold and windy. And very cool, yet again. :)
Chilled in Golden Gate Park.
Checked Frank into his hotel.
Met up with Hendrik, saw Coit Tower, drank coffee (except Frank, whose coffee got pooped into), ate pizza. Headed towards Twin Peaks, where the wind and the evening sun provided plenty of entertainment. Ate quesadillas for dinner at Taqueria Can-Cun.
Headed to SFO, and back home via ATL into BOS.

Back in Debian

2007-07-12T23:28:49-05:00

So after madman is currently orphaned in Debian and might get kicked out of the archive along with xmms kicking the bucket, I just noticed that one of my newer pieces of code made it into Debian: tagpy! :) That’s awesome.

To everybody who is holding their breath waiting for a madman update: I still use madman for my own music managing needs, as it still (surprise) is exactly what I want in a music manager. But I simply don’t have enough time for serious maintenance. I might switch it to audacious in the near future to keep it working for me, but I’m not promising extensive new features…

Trust

2007-03-11T14:59:49-05:00

Watch this:

You won’t regret it. :) (PS: Yes, I know it was made by a bunch of Germans, and yes, they misspelled “belief”, but that doesn’t kill it, in my opinion.)

Embarassing Update: I figured out how to embed youtube videos. :)

New release 0.91 of lircd-xpc

2007-02-26T19:35:34-06:00

I have just uploaded version 0.91 of lircd-xpc to the site. This should make the daemon a bit more resilient against a state where one endpoint thread crashes, and the other lives on, resulting in an unresponsive daemon. If you’ve encountered this, you might want to update. This requires the latest libhid (and Python wrappers) including a patch that I’ve just sent to the list. That patch should make it into libhid svn pretty soon. Until then, you may grab my custom libhid Debian packages from the download directory.

Merry Christmas

2006-12-26T21:33:43-06:00

Dearest readership! It is with great pleasure that on this second day of christmas I offer you my best wishes for the feast of love—with this picture that I have, in true christmas spirit, lifted from elsewhere:

Merry Christmas to all anyway… :)

Attachments

bethlehem-400.jpeg

Why does the Web have a Problem?

2006-11-19T12:28:21-06:00

Because when you’re looking for Runge-Kutta methods, you find neat stuff and end up playing with it. That’s why.

(How’s that for a first journal post after fifteen years, huh?) :)

GCC 4.1 breaks Pylinear

2006-07-01T22:49:17-05:00

If you get errors like

multiple definition of 
`_ZN5boost7numeric5ublas21scalar_divides_assignIT_T0_E8computedE'

when compiling PyLinear, it is because of this GCC bug, and it’s quite likely that you are using GCC 4.1. Supposedly, this will be fixed in GCC 4.2. This bug will likely also affect other UBLAS applications.

The suggested workaround is to use GCC 4.0.

UPDATE: Everything seems fine in at least Debian’s gcc package of version 4.1.1-11.

Comcast, DHCP and MTU

2006-06-26T23:50:33-05:00

If you are a Comcast customer (particularly in Illinois) and your Voice-over-IP (or other large-UDP-packet-using) application does not work as it should, don’t worry, it’s their fault. :P

For some reason, they include

...
OPTION:  26 (  2) Interface MTU             576
...

in the replies from their DHCP servers, which leads your computer to throw away large UDP packets (such as SIP INVITE replies). I don’t know why they’d do this since a) it breaks their customers’ applications (on purpose?!) and b) it puts more load on their network.

In any case, here’s the antidote: In dhclient.conf, say

# F!$#$ stupid comcast.
supersede interface-mtu 1500;

After that, everything should be working fine. If you’re violating their terms of service in doing so, it’s your own fault. I didn’t tell you. 8)

It only cost me like two hours of my life to find and fix this. Quote from the reply from the customer service rep:

I’m sorry, but we do not have any information as to why the MTU is configured for the level that it is. That decision is likely made by the Network Department and something we would not be able to obtain or discuss.

I love you, too.

Version 0.91 of TagPy is out

2006-06-26T23:33:17-05:00

I’ve just released version 0.91 of TagPy. Besides some minor fixes, the biggest change is that you don’t say any more

tag.setArtist("Fat Boy Slim")

Instead, the (IMO) much more pythonic

tag.artist = "Fat Boy Slim"

is now where it’s at. Reading these attributes (and just these) is also changed from tag.artist() to tag.artist. I repeat, this only affects

Tag.title
Tag.artist
Tag.album
Tag.genre
Tag.year
Tag.track
AudioProperties.length
AudioProperties.bitrate
AudioProperties.sampleRate
AudioProperties.channels

Since these are so frequent, I saw a compelling reason to change them. Everything else in the API will remain as close as possible to TagLib’s C++ conventions.

Sorry for the source-incompatible change, it will not happen again.

Enjoy.

Clockwork now shows statistics on the current day

2006-06-23T00:57:49-05:00

I’ve made a minor update to clockwork that now allows you to immediately see how much longer the current workday is supposed to last, from the output of clockwork howlong. This update came out of a particular need that I had at my new summer job at the Math and Computer Science divsion at Argonne National Labs. I’ve also released Version 1.1, with that (and another minor) change.

Rattle back, Rattleback!

2006-05-13T14:20:46-05:00

As part of the course requirements for the mechanics class I took, a team of which I was a part designed a Rattleback. A rattleback is a top that appears to have a preferred direction of rotation, i.e. if you start spinning it the wrong way, it (unintuitively) will turn around and spin “the right way”. There’s quite a bunch more to know about the physics [1] of it, if you care. It might also be an interesting dynamical system to study, as little is known about it past what can be found by numerical simulations. Speaking of numerical simulations, if you are a student of EN137 after me and would like to take a peek at our simulation code (which was written in Octave, with a bit of PyLinear sprinkled in), be my guest.

In any case, you may view the Rattle back, Rattleback! of our rattleback to see for yourself!

PS: I blatantly stole the title of this entry from one of the other teams. :)

[1] A. Garcia, M. Hubbard Spin Reversal of the Rattleback: Theory and Experiment Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, Vol. 418, No. 1854 (Jul. 8, 1988) , pp. 165-197

Attachments

rattleback.avi

Boost Bindings update for UFsparse

2006-04-30T00:07:48-05:00

I have released a new version of the Boost Bindings that fixes a problem with UMFPACK now typically being packaged as part of UFsparse. While I don’t think that this kind of packaging is a smart idea, Debian has recently jumped on the bandwagon, and thus this fix is necessary. The only change is from

#include <umfpack/umfpack.h>

#include <umfpack.h>

to let the code find an umfpack.h that is now under a directory called ufsparse.

Note: This version diverges from boost-bindings CVS, by this one change.

Think of those out there in the cold.

2006-03-26T17:34:53-06:00

While I don’t typically propagate knitting or crochet content, this one was too good to pass up. :)

So, if you have a pretty pitiful-looking tree outside your window, why don’t you do it a favor, and…

…KNIT IT A SWEATER? :D

Oh, and Josie (whose friend Jay pointed her to this story) will claim that she posted first, which is ~~untrue~~ almost untrue. But my post has a picture. And more original commentary. So, there. Me = better. :)

Audio player problems

2006-03-26T17:17:40-06:00

I have generally been very happy with my new Samsung YP-MT6Z, until I hit a snag today. From my tests, it seems that the total lengh of all folder names on the device cannot exceed a certain number, on the order of a kilobyte or so. That boils down to, say, 50 decently-named folders. That’s it. No more. Oddly enough, that kind of kills the advertised second use of the device as a means of data transport, seeing how it has more than enough storage, a gigabyte to be precise.

And the best part is, if you have more than that, the player will fail unobviously and gracelessly. It will simply display NO FILE in the LCD and refuse to find or play all my audio files. There is no error message, alerting me to the fact that I might have hit an internal limitation. It also won’t just ignore any extra folders. No, it insists on just ceasing to perform its main function altogether. Great work there, Samsung software engineers.

To top it off, I had the pleasure of speaking with Samsung Technical Support, who, after lenghty “research”, claimed that the device would not support more than one folder. I got the sneaking suspicion that the guy didn’t even know what I meant by “folder”, and terms such as “nesting” or “hierarchy” seemed to give him endless trouble. Yay!

Buyer, beware. It’s a good device if you can live with this limitation, which I will have to.

Rain Damage, Rain Damage, Go Away

2006-01-28T20:05:39-06:00

As I previously reported, my old laptop, aramis, got rained on and sustained water damage resulting in—lo and behold—not a dead computer. In fact, most of it was fine, it even booted into X, and all of its components were usable, it would just always crash after around 15 minutes. That led me to believe that maybe, by switching out the mainboard, I might obtain a usable machine again.

Putting that plan into practice, I downloaded IBM’s maintenance manual for the laptop, which, for the record, was an IBM Thinkpad A-Series A21m, and looked up the “FRU” (Factory Replacible Unit) number of the mainboard, which, in my case, was supposed to be 08K3264. These boards were hardly to be found on ebay and were hideously expensive, whereas a good bunch of cheaper boards labeled with FRU 12P3156, which suppposedly are compatible with ThinkPad A20 and A21 models, were readily available. I was unsure, but decided to take the risk—given that the 12P3156 board cost only $44.

After two nights of rather nerve-wrecking work on the laptop, I can announce that it is functioning again. :D The difference between the two boards types seems to be only that the one I had (which turned out to be FRU 12P3288) had a header for TV-Out, whereas the new one didn’t. Big deal. I can use my 800 MHz CPU on it, and as far as I can tell, it is being run at full speed.

It turns out that my old board had some visible corrosion damage that probably caused the trouble I described above. If I had taken my laptop apart right after the water came in, it might have dried and thus been rescued right away. May this be a lesson to whoever has to do this next. :)

Summary: When buying laptops, buy IBM! Why?

Sturdily built
Maintenance information available
Existence of a sizable components aftermarket.

That said, I’m not sure Lenovo is going to do as well, but I hope so.

Taking notes like Germans do

2006-01-24T22:09:41-06:00

Since the semester at Brown starts back tomorrow, I was a little concerned about my total lack of writing material. I still have my non-standard ruled paper from last year, but since a) it’s not a standard size, not even US-Letter, much less A4 and b) I don’t like ruled paper, I figured, hey, I’m going to take notes in style like I used to.

Tadaa: Good old German Taking notes like Germans do, letter size, 5mm squares on it.

Courtesy of Incompetech. :)

Sports & The obligatory Aftermath

2006-01-20T16:29:43-06:00

I spent today doing something very laudable: :yes: I went on an extended bike ride. Of course, riding your bike is not half as fun if you can’t waste some time letting your computer process the greatness of what was achieved. Fortunately, the Google Maps Pedometer comes to the rescue! If you follow that link, you can even see a height profile of my ride, and you can witness the fact that I fulfilled my private goal of riding into Massachusetts. Not that I knew that I did, since there were no big “Welcome to Massachusetts” signs around. It was just when my legs started hurting badly enough that I thought the return trip might not be fun if I didn’t turn around now.

Well. Let’s think about that for a second, and make it into a general law of cycling:

Don’t wait till your legs hurt to turn around!

Otherwise, the return trip might hurt. Badly. Not that this reflects my experience today or something. (Owww… my legs…)

(As an added bonus, and since the GMaps-pedometer is so fun, I also put in my way to school.)

2005 in Review

2006-01-09T07:49:47-06:00

Since it’s been relatively (cough) quiet around here, despite promises to the contrary, I’ll add a summary of my highlights of 2005, and hope that that will shed a bit of light on the latter half of the year, where many of you might not have heard as much as promised from me.

January

After my botched oral exam, I got my Diplom in Technomathematik from Uni Karlsruhe.
Went to Amsterdam with Josie for some fun. Crappier hostels along the year still make me yearn for the Flying Pig. :)

February

Started work at Prof. DÃƒÂ¶rfler’s institute, teaching and generally having fun with Jan, Kim and all the others.
Finished the month with a week-long snowboard vacation in Gleiming with Robert, Kiki, Frank and Lippe. Des war cool.
Got admitted to Brown Graduate School.

March

Work in Karlsruhe is in full force before the semester has begun, with bits of research and administrative software.
At the end of the month, I fly out to New York, then Providence, then DC and finally Raleigh to take a brief look at the graduate school from which I had

Initial fixes

2006-01-08T12:56:40-06:00

I have just released Version 0.90.1 of TagPy to fix a few build hiccups that were reported. Keep up the good work! Additionally, and as a cool benefit of me using Drupal for these pages, you can subscribe to the TagPy RSS feed. :)

And of course you may download the new version from the TagPy home page.

A Stapler for PDFs

2005-12-11T20:39:15-06:00

pdftk is a piece of GPL software that I’ve been looking for for a long time: It can split, join and otherwise mangle PDF files, which in my specific case helped me submit one of my (many) take-home final exams as a single file instead of three separate ones, reducing the margin for error on the professor’s part. :)

Oh, and of course, Debian’s already been there:

aptitude install pdftk

madman added to FreeBSD ports

2005-10-24T00:18:57-05:00

Daniel Roethlisberger wrote in to note that madman is now in the FreeBSD ports collection, which should make it a lot easier for you BSD’ers out there to try madman. A corresponding note has also been added to the download page.

The beagle has landed

2005-08-22T15:16:56-05:00

We made it! Since last Friday (pretty late) night, we’re officially in Rhode Island. For the time being, we’re staying at Ginny-B campground, and they are a pretty decent place. I mean, they didn’t prevent the rain yesterday, but you don’t get that at many campsites anyway.

Even better, we’re looking at two apartments that we might actually want to move in to. We just need to figure out which, since, how else could it be, Josie likes one, and I like the other. 8)

More later. Wish us luck. Things are great. :)

Selling my stuff

2005-08-12T11:51:27-05:00

While I’m in the process of packing up all my stuff, why don’t you help me out by buying some? I’d, you know, really appreciate that. And, knowing that it was mine, you know that it’s some quality junk. ;) And I do take Paypal.

You can also find us at the flea market in Singen tomorrow (i.e. Saturday). Come and give us your money. Oh, and take our stuff, too.

Linux on an IBM/Lenovo Thinkpad R52

2005-08-07T09:31:52-05:00

This post summarizes my experience of making the hardware in my shiny new IBM ThinkPad R52 work with Linux.

I initially tried Debian on this computer, but later on headed over to Ubuntu with a home-built kernel for a bunch of reasons on which I won’t elaborate here. Update: I’ve been back on Debian for a good while now.

The following remarks relate to Linux version 2.6.12.3, if not otherwise specified.

Serial ATA: Intel ICH6

The Debian Sarge installer, using the “kernel26” option, fails to recognize the hard disk. The default 2.4 kernel works ok.

After building my own kernel, using the Intel PIIX/ICH SATA support (SCSI_ATA_PIIX) driver worked fine for the hard disk, but the PATA DVD/CDRW/CDROM was not found. The first attempt was to compile support for PATA into the kernel as well (not just as a module), as the driver’s help states:

      This option enables support for ICH5 Serial ATA.
      If PATA support was enabled previously, this enables
      support for select Intel PIIX/ICH PATA host controllers.

However, this is only one part of the way there. The next part is to change

#undef ATA_ENABLE_ATAPI         /* define to enable ATAPI support */
#undef ATA_ENABLE_PATA          /* define to enable PATA support in some
                                 * low-level drivers */

into

#define ATA_ENABLE_ATAPI        /* define to enable ATAPI support */
#define ATA_ENABLE_PATA         /* define to enable PATA support in some
                                 * low-level drivers */

in include/linux/libata.h. I’m really puzzled why this is not on by default.

Update: In Kernel 2.6.14, this patch is no longer necessary (or possible, for that matter). Instead, append libata.atapi_enabled=1 to your boot command line.

Update: In Kernel 2.6.16, there is a bad interaction of early versions of suspend2 and libata, which prevents ATAPI devices from working. The fix is to upgrade suspend2.

Modem: A Conexant HCF WinModem

You need to buy the driver from Linuxant to actually make this work. The snd-intel8x0m module loads, but does not work, so the Smartlink modem daemon can’t work with this, at least not yet.

Graphics: Intel i915

Intel i915, supported out of the box by X.Org, using the 2.6.12.3 stock DRI module for that card. Two things made life with this chip a lot better:

Section "Device"
    Identifier      "Intel Corporation Intel Default Card"
    Driver          "i810"
    BusID           "PCI:0:2:0"
    Option          "MonitorLayout" "CRT,LFP"

    # NOTE THESE TWO
    VideoRam        65536
    Option          "DDC"   "false"
EndSection

The VideoRam option makes the card use more main memory during 3D operation, which makes many GL applications lots faster, because there’s less texture thrashing going on.

The DDC option allowed me to use my NEC Multisync LCD 1860NX with the card; when the option was not specified, it refused to come up with anything higher than 640x480.

Update: I have recently gotten multihead to work, so that I am now using an LCD and the laptop panel side-by-side. Pretty wicked cool. I’ve attached a working xorg.conf.

Network: Broadcom Tigon3

Works great using the tg3 driver.

The only stumbling block is that Debian refuses to include this driver in its stock 2.6 kernels, since it has a builtin firmware blob. Nathanael Nerode, however, recently made a patch that splits this out into a separate file, which would allow distribution with Debian in the future.

Update: Debian seems to have come to its senses; the tg3 driver is back in the stock kernels as of 2.6.14, at least.

WLAN: Intel IPW2200BG

Theoretically supported through the ipw2200 driver, but this is really flaky right now. Early versions (0.19, IIRC) drop the connection every once in a while and need to be reloaded to work again. Newer versions (>1.0.4) are bitten by “Firmware error” messages, as detailed in this bug log. No solution as yet.

Update: Version 1.0.3 is reported working somewhere, but it apparently needs a patch to compile against 2.6.12.3. I haven’t tried this yet.

Update: Version 1.0.0, as included in Kernel 2.6.14, works like a charm. Get firmware version 2.2 to go with it. Better yet, this version is also in the Debian stock kernels.

Sound: Intel AC‘97

No problems using the ALSA snd-intel8x0 driver.

Power management

SpeedStep and ACPI work out of the box with the appropriate drivers. Suspend-to-RAM works, just like -to-disk, even though currently, 3D (DRI, that is) can’t be used after wakeup.

Other things

The display is very bright, even a bit too bright for my taste. The keyboard is excellent, as is usual for IBM machines. The fan is louder than on my old ThinkPad A21m, but with the processing power of a Pentium-M running at 1.7 GHz, I am supposing that this was to be expected. I’m a bit concerned that the ACPI Thermal Zone 0 reports temperatures of 76 degrees Celsius under sustained peak load, which sounds a bit high. But then, the trip point is at 90 degrees, so I guess it’s fine.

Also, I’m currently trying to return the copy of Windows XP that came with the machine, and on the upside, my complaint has not yet been smacked down by Lenovo. So there is some hope. I’ll write a detailed report once this is over.

I’ll keep this report updated as I continue to tame the hardware. Martin Schwenke’s page on the T43, quoted above, was of great help in getting everything to run.

Suspend-to-Disk/Suspend-to-Ram

Suspend to Disk is straightforward, for Suspend to RAM to work you need one little non-obvious trick: specify acpi_sleep=s3_bios on the command line. Also make sure that vbetool does not mess with your video card (i.e. no repost, no state saving). Thanks to ThinkWiki for the hint.

Conclusion

I’m perfectly happy with the machine. :)

Attachments

dualhead-xorg.conf

Life in London

2005-08-04T12:30:08-05:00

The pictures from our trip to London are up. Enjoy this one, as well.

I’ll update this post with an account of the trip soon-ish. Before then, let me just add that you should not stay at the “Smart Backpackers Hyde Park Inn Hostel”, even though they tend to be dirt cheap. You’ll regret it…

Life and the Office, and Work, too

2005-08-04T08:00:31-05:00

Did you ever wonder how much the terms “guilt” and “office” are connected? Did you ever wish to just relax while working (and have your work benefit from it), but didn’t actually do so, since it looks bad? Wish you could work at home, knowing you’d be more effective there? Did nine-to-five ever feel like a nice corset, like an excuse for not getting work done? Do you want to know why there’s no way past proper discipline? I felt many of these things during my days at Prof. Dörfler’s institute, as a consequence of having a proper office, colleagues and all, but I was never really able to put them into context or come to a conclusion on what to do about them.

Paul Graham, of whose writing I’ve been a fanboy for a while, says it straight and to the point, and many of his points seem like appropriate endpoints of thoughts I’ve had.

I’m glad I found this before starting grad school for several reasons:

Now I know I’m not alone in feeling like this.
Feeling guilty about stuff is nearly always a bad idea.
It’s almost always a better idea to just do what you think is reasonable, even if it looks bad.

Duh? Maybe. But I challenge you to live by these principles every day. (And myself, too.)

PS: Sorry that this is propagated straight from Slashdot. But then, I think it’s too important to miss.

Trip to Augsburg

2005-07-21T03:39:45-05:00

The pictures from our trip to Augsburg are up. Enjoy.

Losing my Faith in Debian

2005-07-18T06:09:48-05:00

I really do not like the way things are going in recent times within the Debian project. I feel that a group of people has gained way too much control over how things are run, and those people do not seem to care about users, they seem to care about enforcing the DFSG (the Debian Free Software Guidelines) to the letter. This manifests itself in several changes that have the potential to harm Debian in the short and in the long term:

Drivers containing binary firmware has been removed from the stock kernels, on the argument that they are “non-free”. This is, of course, true. Yet, I don’t see the difference between storing the firmware in an EEPROM in the hardware, or in a binary blob in the driver. Also, demanding that the source to these binaries be open is pure utopia: manufacturers are already struggling to keep the programming info for their chips secret. Forcing them to open up their firmware is a) going to really reveal the company’s IP, somewhat unlike chip datasheets, making this completely unjustifiable from a business standpoint b) mostly useless technically. While I see the point of being able to hack your firmware, the toolchains required to do so are often also not open, most users are not going to care and would rather prefer their device to just work, and, lastly, the binary blobs do not pose any obstacle to kernel development (be it from a portability, maintainability, or clarity point of view).

The most telling part of the story is that while these people (I am tempted to say “zealots”.) were quick to remove the corresponding firmware, but did not care enough to make it available through non-free to be able to cater to the users’ very real interest to use their hardware. Even worse, for the Broadcom Tigon3 Gigabit ethernet chip

These are turbulent times.

2005-07-13T18:52:19-05:00

To follow up on the previous post, my laptop did find its way back to life. But unfortunately, only briefly so. It started all the way into X just once, and now seems to be dead forever. In the light of this, I’ve given in and ordered aramis‘s successor, tentatively named ~~little-bear~~ grizzly. (I was looking for something a bit more manly than little-bear.) :)

In the meantime, the only reliable computer I’ve had is at work, while at home, I first had the apartment server, then my mom’s PC, which I had fortunately begun to build before my laptop died, then my parents’ old PC (a K6-300, yuck!) and now the apartment-server-turned-my-dad’s-PC once again. All these computers have been blessed with kubuntu, which I can recommend as a really sweet piece of engineering. It makes installing a Debian system with non-server-capabilities easier than ever before.

Work is promising to be busy for the rest of my time there, with a bunch of trivial things still to be done, and other, less trivial, but more fun things expected (like a few good pointers on where to take my past thesis work, and someone else in the physics department to bounce ideas off of). Private life will be equally as busy, with

a trip to France
a trip to Great Britain
a farewell party for myself
packing & moving to Providence

all lined up.

Unfortunately, all this means that most of my private projects are on hold until everything settles again, which I pretty much do not expect to happen until October or so. Indeed, these are turbulent times.

Update: grizzly has landed on my desk, and thus far, I’m really happy with it. Except for one dead (red) pixel at the top right of my screen. Which doesn’t really matter. :yes:

Uh, yay.

2005-07-04T18:48:00-05:00

This July 4th could have been better. While it was a very appropriate day to get my visa (yay!), the evening was actually fairly bad. First, I discovered that a heavy rainstorm this afternoon has found its way into my laptop through an opened window. As yet, the laptop is drying, and as far as I can tell, the hard drive is dry. So this might be just monetary damage.

Then, I find out that Drupal has a few severe security bugs. I upgraded my (recent-version) site and yanked the plug from Josie’s. I can almost already hear her complain. I’ve also yanked the gallery site, as I’m not sure whether its software is affected.

Altogether, the good thing is that tomorrow can’t be much worse.

Update: How appropriate. At present, the house does not have running water. What next?

Hendrik on Spiegel Online

2005-06-24T12:35:32-05:00

A friend of mine, Hendrik, is on Spiegel Online with his pet project, the Stanford entry into the DARPA grand challenge. His team also made Slashdot a while ago. :D

Go, Hendrik! Will you ever talk to me again, now that you’re famous? ;)

ARPACK/LAPACK Symbol Duplication

2005-06-22T19:26:29-05:00

Update: The packages distributed here have been superseded by the new autotoolized ARPACK.

This evening, I ran into a pretty frustrating problem with ARPACK. It aborted without being able to compute an Arnoldi factorization from within one program, but the same snippet of code (with the same matrix!) would work just fine in another.

To make a long story short, I eventually tracked down the root cause: Both LAPACK and ARPACK define a routine called SECOND, but with different call signatures. LAPACK’s copy is a function, whereas ARPACK’s copy is a subroutine. Usually, all goes well, and the linker is intelligent enough to pick the right copy in all instances. I was unfortunate enough to be in a situation where that was not the case. The call to the wrong SECOND messed up the stack, and things got really strange from there. (For example, DLARNV returned a NaN instead of a proper number.)

Hence, the ARPACK-supplied SECOND routine needs to be renamed. The attached ARPACK/LAPACK Symbol Duplication against ARPACK version 96 (a revised version of a quick sed job) realizes this suggestion by renaming the routine to ARSCND (as in “ARpack SeCoND”), which fixed the problem for me. Apparently, somebody else on the Octave team ran into a similar problem, but appeared to get only unhelpful replies.

Hopefully, this posting will save somebody some trouble.

Update: This Debian bug tracks progress on the issue.

Update 2: ARPACK/LAPACK Symbol Duplication of the patch works against ARPACK96 with patch.tar.gz applied. It’s also slightly cleaner: It renames second.f to arscnd.f.

Update 3: I got a reply from Rich Lehoucq acknowledging the bug two years ago, but he never got around to posting my patch. Thus I’m taking the liberty to redistribute an updated version of ARPACK here. The build system has also been updated so that it should work on most modern flavors of Linux without any tweaking.

Attachments

Burning you were to know…

2005-06-20T06:39:42-05:00

…the true origin of Darth Tater. Wait no longer.

While pretty unnecessary, …

2005-06-15T17:31:51-05:00

…I think it’s cool that somebody at Radio Bremen translates each month’s news into Plattdüütsch and Latin. Such things only happen when commercial interest is not at the forefront of everyone’s mind.

Find the word

2005-06-04T13:49:14-05:00

Josie’s doing something with kids tomorrow, and she needed a word puzzle. I made her one. No. I made her a program to make any number of them. :) Like this one:

KJAPANINAW
ACOUSLAPDO
NIUSBELEUR
GPRDIPPAPC
ASOUDEHCTE
RBCBHGAERS
OEKYMOUSET
OOEALEHLIE
WIRIJSELAR
CATNGMHIDU

I’ve attached the Python Find the word to produce stuff like this to this story, should you happen to need something like that. It’s the stupidest possible algorithm, purely random. It works well enough, though. You also need a word file, I’ve attached an Find the word. Then you call it with

python find-the-word.py words

and it spits out what you want. Golly. :P

Attachments

Sieve of Erathostenes

2005-06-02T15:41:00-05:00

I challenge you to come up with a shorter/simpler piece of code that executes the Sieve of Erathostenes than this Python snippet:

N = 1000
candidates = range(2, N+1)
i = 0

while i < len(candidates):
    candidates = [n for n in candidates
        if n % candidates[i] != 0
        or n <= candidates[i]]
    i += 1

print candidates

Of course, renaming variables or such silliness does not count. I know that a stopping condition of candidates[i] < math.sqrt(N) would make the whole thing faster, but that’s beside the point. This is about simplicity.

So, go ahead and humiliate me. I know you want to. }:)

LinuxTag is not free any more (Update: or is it?)

2005-05-31T14:24:42-05:00

Just now, I was thinking about which weekend I’d have to reserve to go to LinuxTag, which will be taking place in my current home town, Karlsruhe. Turns out that I don’t even need to decide that: They will be charging 35€ to get in this time around. The show was good in 2003, but yaaaawningly boring in 2004. (Yes, I can hear you yawn now, too.) Even though I am officially employed now and the entrance fee would not be a major problem, I will not pay, and consequently not go. I don’t think charging money is a good idea, and I firmly believe that most of the community will stay away as well. They have a figleaf “community ticket” program that depends on some sort of viral scheme, but I really doubt that it will work out. Look at the people who have made Linux what it is, and who matter the most: the developers. In general, they don’t have much money to just spend on random stuff. Plus, they are already contributing valuable time for free. If LinuxTag wants to turn them away, so be it. In my case, they’ve succeeded. But who else would be the target audience? Commercial users? Maybe. Private users? I doubt they’ll pay. Then who? I don’t know.

Back in 2003, I made a donation to them to keep up the good work. I’d be willing to donate more this year (and yes, more than the 35Ã¢Â‚Â¬ they’re demanding), if they told me that they really needed the support and the whole thing was still free. If they’re forcibly trying to separate me from my money, I’ll really rather stay at home. I’m stubborn like that. Screw them.

They say that Linux has “grown up”, and so, LinuxTag needs to do the same. If they are trying to become a commercial fair, fine. Let them be another BuzzwordWorldExpo(tm). Other fairs are giving away exposition space to OSS projects as well, and have generally more to offer. Let them compete and see how they fare.

Update: The “viral scheme” seems to have worked out at least to a point where I got a free ticket to LinuxTag. I guess that public whining is a good way to score one. So, now I know what to do that weekend. :) Based on how things pan out, I might have to take back what I said above, in which case I’ll do so on this journal.

By the rules, I may now be able to invite more people to LinuxTag. I’m certainly willing to do my part to get more people invited, should they want to. Just drop me an email.

“Sith” is a four-letter word

2005-05-31T01:30:49-05:00

Think anagrams. ‘Nuff said. :)

Don’t shop at K&M

2005-05-27T08:22:11-05:00

If you ever thought about shopping at K&M Elektronik, don’t do it. They suck. Here’s why:

I bought a wireless card from them a while ago, which, over time, ~~bit the bullet~~kicked the bucket. I am supposing that I was seeing thermal problems, since the card would always start out fine, but then degrade and finally cut the connection as it “warmed up”. So far, so good. That can happen. Understandably, I wanted the broken card exchanged. They gave me three options:

Get my money back, minus an apparently time-dependent “fee” of 2,70 Euros. (Hello? The card is broken. You sold me broken stuff.) :o
Exchange it for another card with unknown Linux support right now and let me pay the price difference, without the chance to return it if it doesn’t work in Linux. (Errr, no.)
Send it in and give me a replacement, but even then I can’t be sure that I get the same type of card. (Not even close.)

The rep’s explanation was that I bought the card as “Markenfabrikat” (=”brand name”) without being guaranteed a specific brand of card. Riiiiight. Now I’m really not so sure I bought anything except abysmal customer service. :barf:

Needless to say, I went for the money option and took what little money was left straight to a competitor, which has this far always had good service, even in borderline cases.

Drupal 4.6

2005-05-25T17:47:44-05:00

I’ve just finished updating this site to Drupal 4.6. It went pretty smoothly, except for some hiccups with the image module. I like the new version, even though not much is different. The slightly changed look of the site might give this away, too. If there are any problems with the upgrade, let me know.

Update: The new image module is a dog. I’ll be using gallery2 integration instead. The Drupal module project has more info.

Update 2: Screw it. Not even that works right. My galleries and journal stay nice and separate. Period. :no:

Check…

2005-05-18T14:42:22-05:00

…this out. :)

Misspelling my Domain

2005-05-17T15:48:46-05:00

In recent times, I’ve actually seen more misspellings of the domain which hosts this journal than accurate ones. I’ve seen

tinker.net—which, in reality, is a site on how to catch reel big fish.
ticker.net—which, in reality, is held by some domain squatter.
fiker.net—which, in reality, does not exist, but bears a close phonetic relationship to a pretty vulgar German word. Worst of all, I caught this misspelling from somebody in my tutorial session. I really don’t know what she thinks of me, but I’m fairly sure I’m not like that.

This makes me feel the accomplishment of someone who’s found the right domain name, since people can’t seem to spell my last name either. :) Of course, I can’t really judge how many people just think, “Awww man, learn to spell.” To which I could grinningly say, with the same feeling of accomplishment, “Hah! The pun you’re missing is in German. The equivalent pun in English can be pulled off using tician.net (which you may grab while it’s not taken).” Unfortunately, I might not be there to enjoy the occasion. ;)

Of course, my desire to add one pun on top of another by calling my journal “news.tiker” has only aggravated the problem. Oh well. }:)

madman meets musepack

2005-05-17T15:43:52-05:00

Thanks to Scott Wheeler’s tireless work on Taglib, madman (the development version in Arch, really) now has support for Musepack (or “MPC”) files. This was half-way to being added once before, but support in Taglib makes this a no-brainer. Seems like tag writing in Taglib is still not 100%, but I assume it’ll get there. I hope you’re happy. :)

To celebrate this, I also fixed Arch downloads. If you got complaints about 2004 last time you tried to get something from Arch, try again.

Debian packages of the development version

2005-05-16T16:41:35-05:00

Ben Montgomery made some Debian packages of a recent development version of madman. Try them out, and bother Ben or the list if you have trouble. Yes, I do agree that this is a strong sign that I should release 0.94 ASAP. It’s coming, don’t worry. ;)

Update: Ben just told me that he doesn’t have time to keep up with development any more, so he’s retired his packages. Any volunteers to take his place?

Semolina Pudding / Grießpudding

2005-05-15T13:41:50-05:00

Yesterday, Josie and I were shopping at Penny and found a pack of semolina that had a really gorgeous-looking pudding on the front. I mean gorgeous, like, just the right reflectiveness, wobbliness, everything. And the pack, whose side conveniently showed the recipe, was only 35 cents. Arrr. Soo good. In any case, if you need dessert and are not afraid of converting from metric, here goes:

125g of semolina (Weichweizengrieß in German),
1 liter of milk,
4 heaping tbsp of sugar,
a pinch of salt,
peels scraped off a lemon with a fine cheesegrater (Josie says that’s called “zest”),
and an egg.

Now do the following:

Bring the milk to a boil.
Separate the egg white from the yolk, stir the yolk into 2 tbsp of water until smooth.
Beat the egg white until it’s pretty stiff.
Once the milk is boiling, take it off the heat and stir in sugar, salt, zest, and watery yolk. Bring to a brief boil, and take it off the stove.
Fold the egg white into the mixture, pour into serving bowls and let it get cold.

Hmmmm. Oh, and while you’re at it, make lemonade from what’s left of the lemon. Don’t you remember? If life gives you lemons, … ;)

snort

2005-05-14T05:08:00-05:00

Courtesy of Andrea (who’s giving Josie socks), I give you Darth Tater:

Attachments

darthtater.jpg

Who’s gonna pay for that content?

2005-05-07T06:59:06-05:00

If you are like most everybody else, you keep music on your computer, be it in MP3 or some other format. Then, most likely, not all of that music got to your hard drive in ways approved by the recording industry. You feel bad for that, don’t you? You probably don’t care about rewarding the record company, but you feel that the artist would deserve at least what they get if you buy their CD. What to do? Send them money on Paypal, saying, “whoops, I copied your music, but no hard feelings—here’s a dollar to ease your pain?” Doesn’t sound good, does it? Some have suggested a culture flat rate, a government-administered fund that pays content creators. But why involve the government in something that it knows nothing about? How would the fund distribute the money? Unsolved problems that, IMHO, prevent this from being viable. So there’s a problem here: You want music. The artist wants to give it to you. But he needs to eat, too.

Actually, it’s worse than that. Movies, music, texts, software, all this can generally be seen as content. Much of today’s content is put out by an industry dedicated to its creation. There’s the movie industry, the software industry, the publishing industry and the recording industry, which, despite some differences, are all suffering from similar problems these days.

The purpose of this essay is to propose a solution to this problem, and the key term will be a “college of creators.” But before I go into any detail about that, I would like to state some truths about creation and dissemination of content that I think are invariable:

Copying, legitimate or not, of content is here to stay, and cannot be prevented. The industry’s Digital Rights Management (DRM) technology will be both expensive and ineffective. Corny and dated as it may sound, “information wants to be free.”
The widest possible dissemination of content is not only inevitable, it is socially desirable. The public domain should be as large as possible.
If a creator desires to be rewarded for (or even make a living off) his creations, there should a way for him to receive a tangible (i.e. monetary) reward for a substantial contribution to the public domain.
Publicity is unnecessary, word-of-mouth is sufficient. Advertisements only serve to skew the public opinion to further business interests. Diversity is better.
The cost of publishing is rapidly going down—Internet
There is too much content on earth.

We wash so much whiter…

2005-05-02T03:28:59-05:00

…and that’s guaranteed! :)

Summer is here!

2005-05-01T08:14:16-05:00

So, how come I finally noticed? We had a phat BBQ yesterday, starring some salad, grilled veggies, steaks, and finishing with Josie’s berries’n’cream dessert. While your mouth is watering, I’ll also tell you that to enjoy this, I spent much of yesterday cleaning our balcony. Sounds un-fun, but actually it was. I like splashing around with water and removing really dirty dirt. :)

Being outside also rocks since I can’t do most fun things for another week because of some stitches I got on my back. But they’re healing, and come next Sunday, I’ll be back to being allowed to do what I want! Among the first things I’ll be doing will be: Take a normal shower. Washing yourself with just a washcloth on the sink is so slow, it stinks. (No, not I stink after washing like that. It stinks. }:)) Go skate! * Go to the pool! :yes:

The clean balcony also enables me to sit outside with my laptop, using my newly-gained wireless networking. Wireless (or nocat—see the Einstein quote on the left ;)) rocks, I must say. At this point, I’m using NDISWrapper to drive the card, since the open-source driver is, umm, not-quite-there-yet(tm). I hope that’ll change, though.

Update: My card was broken. That’s why rt2500 didn’t work. Now that I have a non-broken card, everything is fine.

Now, back to our regularly scheduled lemonade slurping. :)

Harmful and privileged

2005-04-25T17:00:12-05:00

Imagine that you have a habit of harming other people. Imagine that other people feel nauseous in your presence and can’t stand your stench. And it’s not like you couldn’t be like everybody else. You are choosing to be toxic. What would you expect society to do to you? Lock you up? Maybe.

Actually, far from it. The German society—well—Deutsche Bahn, the German railway company, will gladly grant you more space to sit (with others like you), give you your own bathroom and many other perks. As you might have figured out by now, this is about smoking. In recent times, as German trains seem to be filling up for some reason or another, people in non-smoking compartments have a tendency to sit packed, filling up every seat, while smokers are often seen leisurely lounging in the pair of seats they have to themselves. I’m a bit mad about this because a harmful habit is what grants them this extra degree of freedom.

I am claiming that there is also no business argument in favor of having smoking compartments. Consider what happened after the recent smoking bans all over the world (and in Germany!): Bars and pubs ended up selling more. But I imagine that no-one is listening and that Deutsche Bahn, as so often, is asleep at the wheel. (As you might know, this only refers to the long-distance trains—Smoking was prohibited on regional trains a while ago.)

And white smoke has finally risen…

2005-04-17T17:00:23-05:00

After a 10-day visit to the US, much pondering with friends, and a late night, a decision has finally been reached. And no, I’m not talking about the pending election of the new pope. Instead, I have (in cooperation with Josie) chosen where to spend the next few years of my life.

The winner is…

…the Division of Applied Mathematics at Brown University, where I will be pursuing a PhD.

I ended up with offers from there, the University of Maryland and North Carolina State, both of which I (sadly) had to turn down; they also looked like very decent places to go for an applied math PhD. Alas, I can’t tear myself apart. :o What made Brown my top choice was a really wild mix of feelings—in the end, there were many rational arguments for or against each place: public vs. private, big vs. small, heavily grant-oriented vs. fairly independent, sociable vs. individual, rather European-looking surroundings vs. the South (which I like, don’t get me wrong)… A few days after having given up the liberty of choice, I’m still comfortable with my decision. I guess that’s a good sign. :)

I’m still trying to recover from jet lag. I don’t know, the eastward trip always bothers me more than the westward one. I can’t seem to get over the time difference in less than about 1.5 weeks. :? I was lucky enough to sit next to Till on the way out, which made that flight a whole lot more fun than the return trip. My initial three “vacation” days in New York left me sick and were not enjoyable at all. NYC is, from what I can tell, just too friggin’ big for me. Just seeing nature after not having any for three days was such an enjoyment… From there, Greyhound and Amtrak took me around the country to all those three schools, which was fun. (Including the overnight train ride to DC.) I flew out of RDU last Sunday, after having seen Josie’s parents on Saturday.

The last week, my first week back in Germany, also marked the start of the semester at Karlsruhe, providing for a lot of work. I’ll be taking courses on differential geometry and multigrid methods, and I’ll be teaching half of a tutorial section on numerical methods for PDEs. My first appearance in that function is due Apr 26. 8)

Free Software-wise, I know from the emails that I’ve been getting that I really need to spend some time on madman (which will happen in due course). I’m also in the process of preparing a new release of PyLinear. Hang in there. :D

Trolltech broke my madman

2005-03-27T06:09:17-06:00

Apparently, Trolltech made some incompatible change to QTable’s code, and by “virtue” of this change, in-line retagging doesn’t work for me any more. I’ll endeavor to have this fixed soon, but I’ll be out to the US for about a week and a half at the end of March, so I suppose it’ll be a while. Hang in there…

Update: The problem appears to have vanished. So, I assume it was a temporarily broken Debian package. Or a fluke in the transjector matrix. Or whatever. :)

Äl Jawala

2005-03-27T05:58:56-06:00

The sole purpose of this entry is to provide some Google Juice to Äl Jawala, a two-sax, two-drummer band. We saw them outside Karstadt yesterday, and apparently they were at Sauschdall in Ulm last week. They rock. :yes: We bought their live CD “Balkan Big Beatz”. Dude. Kewl. 8)

If you are the US government reading this for purposes of granting my study-abroad visa: They’re a jazz band, not an Islamic terrorist group.

Buying pants in Germany

2005-03-27T05:36:36-06:00

Let’s start by stating a fact: Pants sizes are an at least two-dimensional affair. You have the waist and the length. Both are correlated, but not so strongly that it would be effective to make a one-dimensional measurement out of the two. Obviously, some are fatter than others. (For example, unfortunately, my waist length has gone from 34 two years ago to 38 these days. :( My legs, however, have stayed the same 36 inches they’ve always been.)

Some parts of Germany’s textile industry, however, seem to think that one size number for pants is enough. When the above fact struck them (and I suppose it must have struck them pretty hard), they apparently panicked and thought of a very ad-hoc fix for the situation: Take the pants sizes as they are and double them to mean “longer legs and thinner”, and halve them to mean “shorter legs and fatter”. That’s one way to conceal two dimensions in one. And it yields a very amusing nonlinearity: I can typically wear pants sizes between 54 and 106. Minus the range between 56 and 102. :)

But it gets worse. Apparently, C&A figured out the above fact, but too late. Within the same stack of pants (from the same manufacturer!), we found:

Pants marked two-dimensionally in inches. Good.
Pants marked two-dimensionally in some strange unit, by which I had 54x48. Very apparently not centimeters. :)
Pants marked one-dimensionally in inches.
Pants marked one-dimensionally in the German system.

People, get a grip. I know that nobody’s listening, but for once I suggest Germany just adopt the American 2D inch sizing system. It works. It’s well-defined. Lord knows I’m no big friend of the Imperial system of units. But this is one instance where Germany’s otherwise happily-metric system is severely screwed up.

If you wanna see what I bought, look here. (I took those pictures mainly for my parents’ benefit. Not for yours. Did you really think I was that vain? ;))

We’re not creators of ourselves

2005-03-24T01:35:21-06:00

We’re not creators of ourselves
We are only skin and bones
Our universe immersed in glass,
and now it’s raining stones…

Lyrics by De Phazz, “Keep it simple” on “Natural Fake”

De Phazz - Natural Fake

2005-03-23T15:42:17-06:00

After a little more than a week of owning De Phazz‘s new album Natural Fake, I’m not as excited as I was at first. Maybe that has to do with how much I listened to it. But altogether, I have noticed that many songs are much more repetitive than their previous work. And repetitive = boring. The CD has been fairly successful at putting me to sleep on several occasions, which may have some merit. :) But listening to it today right before listening to some songs by Cargo Cult reminded me that there’s sometimes more to a good song than what Natural Fake delivers. Hear you me, the album doesn’t suck. I like it. It’s just a bit short of the mark that was set by, say, Death by Chocolate.

Anyway—keep on rockin’. 8) ;)

When I grow up, …

2005-03-23T15:41:19-06:00

… I want to be just like him.

Photo trip through Ulm

2005-03-16T14:51:34-06:00

Josie and I went on a brief end-of-winter-walk through Ulm on Sunday, and even though it was pretty cold, we got a hold of a few shots. Enjoy.

I, for one, am glad that spring has finally sprung in Germany. Just imagine. I was out in a T-Shirt today. And I officially opened my inline skating season. Wow. ;)

It feels good…

2005-03-09T14:32:04-06:00

…to have a door again.

That will be all.

The Battery Rant

2005-03-07T18:41:05-06:00

No matter what you do, do not ever buy the 2000 mAh rechargeables from Hama. They seem to have a pretty high defect rate. Today I finally got a refund for my set of four, which I had exchanged no less than twice before. Each time I came home with a new shiny set, I put it in my charger (also from Hama), only to find out that at least one of the batteries was broken. (Broken as in zero voltage between the ports.) At least the refund from Saturn was fairly painless.

Saturn, however, attempts to trick you when you buy stuff. They have two separate battery shelves at different locations in the store, one overpriced and no-name, the other reputable-brand and reasonably-priced. Inexplicably, the section with the reasonably-priced batteries says “Inkjet paper” at the top. It makes you feel like they don’t want you to find that stuff. This is true at least of the branch in Karlsruhe. (Note that my batteries came from the reasonably-priced shelf. :P)

The State of 802.11g Wireless Networking in Linux

2005-03-07T18:40:34-06:00

This entry describes my way to wireless LAN in Linux. It is an update of an earlier article. All I wanted was an 802.11g-compatible PCMCIA card with a reasonably open and functional Linux driver. I find it ok if the card’s firmware is not open source, but I hate dealing with binary-only blobs in kernel modules. NDISWrapper is (IMHO) a cute hack, but even uglier than binary blobs and consequently not an option. Under these constraints, there weren’t many choices in terms of the chipsets that I might use:

The RaLink RT2[45]00 chips have GPL drivers released for them by the manufacturer. My new WLAN card, a Conceptronic C54RC, has this chipset. There are a ton of others that have it, too. The rt2500 driver, checked out from CVS as of 2005-03-07, works for me, even though it takes a while until the card associates with an AP. The drivers’ web site could also stand some improvement, but I hear that this is being worked on. Also, the card is supposed to be tad slow, probably because of a bottleneck in the current driver. But overall, I’m pretty happy with my purchase.
The Intersil PrismGT was a hopeful contender, until the card manufacturers discovered that they can save a few cents on each card if they stick the MAC-layer processing into the host driver instead of the card’s firmware. (This cheaper type of card is called SoftMAC and will make the prism54 driver complain about “no 'reset complete' IRQ seen - retrying” and “prism54: Your card/socket may be faulty, or IRQ line too busy“). The driver is complete, functional and open, but it doesn’t support the SoftMAC cards, and the earlier hardware versions that aren’t SoftMAC are nowhere to be found, not even on ebay. I find it a shame that a good piece of software is ruined by the manufacturers’ greed. Honestly, I’d gladly pay more for a card that just works with an open driver, and I believe that I’m not alone. What’s worse: There is no sure-fire way to tell the bad cards from the good ones. In some instances (like the Netgear WG511, which I bought and returned), both types are almost indistinguishable visually. (In fact, for the Netgear cards, the label MADE IN CHINA is the only feature that seems to suggest unsupportedness.)
Atheros-based cards seem to be an ok (but partly non-open) choice. It seems that there is no (or very little) on-card processing power, no firmware, but a binary blob called “HAL” (as in Hardware Abstraction Layer). This HAL shields the card’s apparently sensitive and protection-worthy programming interface from the prying eyes of evil source code readers. (Not that the whole thing hasn’t been reverse-engineered yet.) It seems that the restriction is partly the FCC’s fault, which doesn’t allow freely programmable software radios like the Atheros chipset in a consumer’s unqualified fingers. Since there is so little on-card processing power, the driver needs to reimplement nearly all of 802.11. This implementation is, apparently, fairly complete and was ported from OpenBSD, but I’ve heard that it lacks a properly working Ad-hoc mode, which is no big deal to me.
Intel’s ipw2200 (aka Centrino) looks good, but is not available in PCMCIA form factor.
TI’s ACX100 is a relatively new contender driver-wise, and, according to its developer, it basically works, is currently maintained, but doesn’t do WEP reliably. TI is also notorious for not giving out hardware specifications to open source developers.

If you are a wireless card manufacturer reading this, please try to follow these guidelines:

If you change the chipset, change the name of the card. I hate nothing more than thinking that I’m buying something that’s supported, when, really, the hardware has changed and I’m buying an expensive paperweight.
If there’s a working open-source driver, don’t stop selling the card. If it’s causing cost trouble, up the price. I’ll gladly pay for something if I know that it’ll work.
Release full-GPL Linux drivers, or at least release hardware specifications (possibly under NDA) and allow the development of GPL drivers.

If you are a customer looking for a card and reading this, consider not buying from manufacturers that do not follow these guidelines. At this point, this includes D-Link and Netgear, among others. Instead, consider buying RaLink-based cards. They’re inexpensive, and they have full-GPL drivers that work (minus a few bugs, which I assume will be ironed out).

For more information, see the exhaustive NDISWrapper card list and Jean Tourrilhes’ Wireless HOWTO. I have tried to combine its information content with driver-specific bits’n’pieces I scavenged for my own purchasing decision. I hope this helps somebody.

AutoDJ from Source

2005-03-06T18:43:15-06:00

I just committed a first cut of the much-requested “AutoDJ from Source” functionality. I’d appreciate feedback on how well it does for everybody. 0.94 is well on the way, but as you may have seen from my journal, my life has sped up rather than slowed down since I finished. However, I’m fighting to actually spend some time on madman maintenance. Wish me luck. :)

I won’t believe it until I see it with my own eyes, but…

2005-03-03T14:05:06-06:00

Good news, everyone! :D My parents just got a letter from Brown University stating my admission to the PhD program in the Division of Applied Mathematics there. Yeeeee-haw!!!! :D 8) Oh, yeah, there’s plenty of financial aid to go around, as well, including a fellowship for the first year. How cool is that? :D I guess I’ll have to have a party some time soon…

Update: Here are some pictures of what it’s supposed to be like. :)

Update 2: I’ve decided to stop talking about matters of grad school admissions on this site until I know where I will be going, since it has come to my attention that at least one person involved has actually read what I wrote. I’d find it poor style if I just kept blabbing. I’ve also removed the link from tiker.net to this site.

Back from Snowboarding

2005-02-28T17:45:33-06:00

I’m back from snowboarding! It was great. :D Snowboarding rocks. Even the one ski person we had on the trip had to admit that. }:) We had seriously great weather. Sun during the day, snow during the night, fairly warm, good snow conditions. I don’t know what could’ve been better. 8)

We were at Reiteralm, which is part of a group of four mountains, all of which we could use on our lift ticket. We didn’t even have enough time to explore the area in its entirety. :P There were great spots for boarding through pristine powder, enough forest to go around, a good mixture of runs, with the black (i.e. “difficult”) ones typically being best. Our accomodation was excellent, too, as you can see from the photos. We spent most (but not all) of our nights roleplaying on Kiki’s adventure. My character, Lenelun Lender Le Lenwed (a gnome paladin) almost got killed twice, but he also caused enough damage to have been worth the trouble it took to create him. :)

I made an online photo gallery.

Digital Camera Kung-Fu

2005-02-18T11:47:28-06:00

Since the beginning of this month, I’m the proud owner of a Canon PowerShot A95 digital camera. (Note that the new category “Digital Photography” has sprung up on this site.) I’m very excited about the camera. The battery life is just incredible. The only hair in the soup is that the tripod mount is plastic. But if it lasts, I guess I don’t care.

In order to protect this investment (gift, actually), I’ve thought of a somewhat creative use of the MyCamera(tm) feature of this camera, which allows you to show a custom picture during power-up. This is what I put in:

If you want to do this, too, I’ve attached the Inkscape Digital Camera Kung-Fu. (Oops. Now you know my phone number and email. Oh well.)

To get the best possible use out of the pictures you take, I can whole-heartedly recommend DigiKam, which is the best picture archival/viewer program for KDE that I’ve encountered until now. It’s been dealing well with my 2000+ pictures.

You may now spend the coming week in awe awaiting tons of snowboarding pictures week after next. }:)

Attachments

Eat that, C++!

2005-02-18T10:57:08-06:00

You didn’t know that Python can actually have custom infix operators, like vector1 |cross_product| vector2, did you? From where I’m sitting, I can see C++ turning green with envy. :) Of course, this is a hack, but it’s actually not so horrible that it wouldn’t be usable in day-to-day code. The operators obey module/package qualifications, too.

How neat is that? :jawdrop:

Update: Very little actual thinking reveals that the same thing is possible in C++ as well. It remains as a last consolation that pulling it off takes about three times as much effort as in Python. :)

Out for snowboarding

2005-02-18T09:10:56-06:00

If you’re trying to contact me, and it’s the week between Feb 19 and Feb 26, you might as well give up now. I’m enjoying myself snowboarding somewhere in Austria. Wish me sunny weather, good snow and all that.

See you week after next! 8)

Someone is trying to fool us…

2005-02-17T02:44:51-06:00

…us as in the Germans, and, sadly enough, it’s working. When I first saw the brand of bags called “Down & Out” (Pleite & keine Hoffnung), I thought them to be an oddity, something that no-one would buy. Then I saw them in the wild, and even better, on my own friends. Then, I really thought, hey, brush up on your English, dude. Know what you buy.

Today, it all got topped off, when, on my way to work, I saw a fairly cute, curly-haired girl with a bag branded “Flaming Sideburns” (Flammende Koteletten). Riiiight. Girl. Sideburns. Ahem. Is some cruel bag manufacturer out there trying to find out what would be the dumbest brand of bags that Germans still buy? If so, stop it! You’ve had your fun, now go produce serious stuff. ;)

I’m sort of glad that my own backpack says “Bach/Ireland” and nothing else. 8)

Update: Christian Plagemann pointed out that “Flaming Sideburns” most likely doesn’t refer to brand of the bag, but rather a band. Note to self: Must not fall for random conspiracy theories, no matter how attractive. :P

Every Mathematician should have a Coq

2005-02-08T18:23:41-06:00

(No, not what you think.)

For work today, I had to proofread some lecture notes for my boss. After work, I started wondering: What if we could leave the tedious verification of our math to computers? To me, math is about understanding and constructing, and not so much about meticulously verifying “old” proofs—an activity that I believe should be automated.

Coq is one system that achieves this, even though it’s geared more towards constructing rather than checking proofs. It’s written in Objective Caml and will be consuming a bit of my research time to come.

As an aside: While surfing the web surrounding Coq’s site, I discovered AEI’s Living Reviews, with that being journals of review articles that come with a promise of maintenance and actuality. Currently, this kind of journal only exists for solar physics and relativity. While the “current” system of just republishing an old review article is not bad, I think that making the whole maintenance process explicit and giving it a shorter turnaround time is actually a pretty good idea.

The Lowdown on Amsterdam, Day 4 (Saturday)

2005-02-08T11:01:05-06:00

(You might want to read day 3 and all the days before this one to avoid getting confused.)

Day Four was very much reserved for the van Gogh Museum. While “museum” always has that dusty, old sound to it, this is certainly untrue of this particular one. After grabbing lunch (ham’n’cheese croissants—mmmh, greasy fish— :barf:, cold chicken— :barf:, smoothies—mmmh), we trodded over to the Museumsplein and forked over the 10€/person to get in and 4€ more for the (good) audio tour. To start with, I sent Josie a video message from the museum while she was on the pot. Fun…

The museum starts out on the ground floor with a roomful of what was before and around Vincent at the outset of his life as an artist. This is the door-opener for the first floor, where you stumble into a narration of Van Gogh’s (short) life, visualized and accompanied by his pictures. It is amazing to see how his life’s events and mental self reflect into his pictures and vice versa. I would honestly have missed much of what can be discovered without the audio tour, and plus, it gives you the opportunity to just sit down and listen when your feet grow weary of all the standing around. They also have a PDA-based multimedia tour available, but I wonder if it would be worth the money, since I found it just perfect to be able to listen and look at the paintings, keeping both senses busy, but not overloaded.

We stayed and explored the museum until closing time. Among the things we found were (well-taken) artistic photos of the sites in Vincent van Gogh’s life as they are today and a few computers where you can explore the museum’s comprehensive web site, linked above.

Afterwards, we bought dinner, cooked, and headed out for a night on the town. Our first step was into some “we have free live music” place. It was good. Until the free live music, courtesy of “The Invisible Orchestra” started. We downed our drinks and left, looking for other entertainment. “The Australian” had bad coffee, expensive chocolates and girly music, all delivered by the most boygroup-y band of servers I’ve ever met. Making rounds through a few other bars, we ended up back at Boom to drink to the end of our trip.

(That’s it. You can stop reading now. Day 5 was the ride home—and fairly uneventful, except for what happened around Frankfurt. Yeah, we saw the Begijnhof before we left. Go there, it’s neat. But it doesn’t warrant another entry.)

Update: I realize that I did a shoddy job proofreading an earlier version of this entry. Fixed.

SomaFM/Groove Salad rocks the socks off Internet Radio

2005-02-08T08:31:54-06:00

I’ve been listening to Groove Salad at work for a while now, and their mix has rocked my boat continually. Highly recommended. I have yet to try their other stations. 8)

Update: Embarassing typo in the previous Title line: Does it really rock the socks of Internet radio? I’m curious what that would look (or sound?) like. :O

The Lowdown on Amsterdam, Day 3 (Friday)

2005-02-07T15:13:05-06:00

(You might want to read day 1 and day 2 first.)

Friday was most likely the most serious sightseeing day we had during our time in Amsterdam. We started out in a wool store for Josie, had lunch that gave me heartburn for the rest of the week, and then continued on to the Anne Frank huis. (BTW, for all you non-Dutchies, “ui” is apparently pronounced as a weird mixture of “oi” and “ow”. That’s at least what we inferred from the tram pronouncing its stops, in this case, notably “Het Spui”.)

The Anne Frank museum is worth seeing, and more than once I felt a chill moving down on my body, for example when we came to the part of the museum that covered how the Frank family got caught, deported and executed. It is a truly moving museum, and they have made much of the fairly little that there is to see. I left there with a wish to actually read Anne Frank’s diary, maybe I’ll make that my next book. There are some “extra” sections to the museum, like one room where an audience votes on current matters of “freedom of expression” vs. “restriction and persecution of certain forms of expression”. This was enlightening, often surprising, and technically well-realized.

After leaving the Anne Frank huis around five, we headed for Amsterdam’s smallest gallery, and from there past the Westerkerk to the “Homo-Monument”, some large sculpture consisting of three interconnected triangles commemorating the suffering of homosexuals in our societies, present and past. (BTW, I challenge you to take a picture that a) does it justice and b) contains all of it.) From there to Bagels and Beans for a honey-walnut bagel and some refreshing juice. Hmmmmm. Nice ambience, too.

After a bit of shopping, we decided to go see the red light district. That included popping into one of the novelty-filled sex shops. I wonder what people do with all the things you can buy there. :? While we were in there, Josie hit me repeatedly to make me stop giggling. }:) I am amazed by how well the “district” integrates into the rest of the city life—there are no high concrete walls like in Germany, and really, there’s a huge vegetable market just one block away from it. I admire the Dutch for their easy-going-ness. I wanna be like that. :)

Later on, we went to Local Amsterdam for some rib-eye steak and lamb, well-grilled on a skewer. In this hip-type restaurant, you just sit on long tables and eat standardized dishes off of standardized plates. That sounds like prison dining, but is actually really fun and, again, very lecker. This was a recommendation in the Boom magazine, whose advice to have one of their dessert cocktails twe also heeded. We had Brazilian Iced coffee out of a large conic cup, and were nearly perfectly happy when we left there to go explore the Jordaan part of Amsterdam by night. On the way back to the hostel, however, we got lost pretty badly. Not that it mattered much that we wandered around in circles, it was actually quite fun. It only took us four hours to get back to the hostel…

(Read on.)

The Lowdown on Amsterdam, Day 2 (Thursday)

2005-02-05T07:35:37-06:00

(Read Day 1 first, otherwise you might get confused.)

On the way out of the hostel yesterday, we had stumbled upon an outdoor gear store. I had been on the lookout for a new backpack for quite a while, and right there in that store I saw what I had been looking for all along, the Fjällräven Fox 60+10, a backpack with more extra features than you can shake a stick at. Having slept over the decision, we went out to buy it this morning. Speaking as of this posting, I thoroughly don’t regret buying it, especially since they had it marked down by 30%.

A while back, maybe in November, Josie and I had also stumbled upon these funny videos—we had enjoyed them without knowing whom exactly they were from. We just recognized the logo when we saw it, right there in our bedroom: Boom Chicago. Their magazine does not only contain a pretty good tourist guide, they also brought our attention to the fact that they have a improv (that’s improvised theater, like “Whose line is it anyway?”) show every night. Having liked their videos, we bought tickets at the theater at Leidseplein for their “Best of Boom” show.

The rest of the day was pretty rainy, so after getting the new shiny backpack back to the hostel, we went shopping, explored the Museumsplein and cooked until right before the show. The show itself was hilarious. Quoting any jokes here would most likely ruin them, especially since I didn’t take any notities during the Sjoord show. :P If you ever have the chance to go see Boom, run, don’t walk, to the ticket counter. (And use their magazine as a coupon to score a 3 Euro discount.) After the show, there was an improv jam for people from the audience, many of whom were apparently regulars. They put on a pretty funny show for the rest of us who were too scared to get on stage. (including me)

Read on for day 3.

madman news now syndicated from my journal site

2005-01-31T18:25:00-06:00

madman’s home page now semi-automatically draws its news from the madman category of my journal. I hope that this will encourage me to update madman’s news more frequently by making the process more convenient than writing PHP source code. (ugh) As an added benefit, you may comment on the news by clicking on the title links. Why not try that out right now? :P You may even syndicate the madman feed on your home page.

In other news, I’ve finished my degree and will hopefully have more time for madman for some time to come. Especially, the 0.94 release has been burning under my fingernails for quite some time. But there are still quite a number of bugs to catch. As always, it’ll be released when it’s done… :)

The Lowdown on Amsterdam, Day 1 (Wednesday)

2005-01-31T09:23:18-06:00

To say the least, the trip was great. Amsterdam is a very cool city, with all that water, all the bridges and the people, something is just right there. Sadly for you, there aren’t going to be any pictures, at least not for a while, since all we had was Josie’s point’n’shoot camera.

We arrived by train, which was surprisingly easy and convenient, even though at roughly 60 euros per way per person with BahnCard 50, rather expensive. We had to change twice, once in Mannheim and once in Köln. We got there around five o’clock, got our strippen-kaart, took the tram 5 to Leidseplein, which is the nightlife location in Amsterdam. We walked a few hundred meters, and found our hostel, the Flying Pig Palace Hostel in the Vossiusstraat, near Vondelpark. Out of the hostels in Amsterdam, it seems to one of the more mature ones. They have excellent facilities (free breakfast, free Internet, free kitchen for you to use) and the hostel is run by temp-worker backpackers who stay there free and do all the work (reception, room-cleaning) in return. We settled in an eight-person room in the lower “shelf” of a queen-size bunkbed. That sounds rickety, but actually works great.

For the night, we went back out to Leidseplein to find something to eat, only to discover that most of the tourist trap restaurants are way too expensive. If you keep walking, however, you find the places where the Dutch go, and in our case, that was an Italian restaurant that had yummy pizza for a mere 5€. After that, we kept walking downtown, explored a bit and found somewhere that served (stuntprijs! 8)) great apple cake and coffee for a mere 2,50€. We came back to the hostel a bit exhausted around 1am.

(Read on.)

Functional Analysis II Summary

2005-01-30T15:09:27-06:00

Just in case you are interested, Functional Analysis II Summary is a PDF link to my (German) summary of Dr. Christoph Schmoeger’s Functional Analysis class of 2001. While those that have taken that particular lecture and need to take an exam on it may be a dying breed, I’m posting it in the hope that it may be of use to some still. Source code is available upon request. The document was composed in TeXmacs.

Attachments

fa2.pdf

Groeten Uit Amsterdam

2005-01-27T05:25:29-06:00

Hi people, greetings from the Flying Pig in Amsterdam. I completed my degree! Did that last Monday at 10:47. (The last exam could have been better—I got a 1.7. But who cares now? }:)) Then, I completed my applications to grad schools in the US. Did I ever mention that most public schools’ application systems suck big rocks through thin straws? I got a rough load of that Monday night. Forms changing my input without being asked to, web servers crashing, “add another” buttons not working… But the applications are mostly done now. Phew!

In order to celebrate my being FINALLY DONE :), Josie and I removed ourselves from everybody else’s reality in Karlsruhe and decided to chill the rest of the week in Amsterdam. Hope y’all are having fun, too. We went out last night, and right now we’re about to seriously see the city by daylight for the first time. Still weird that everybody but us seems to have come for the drugs and the sex. :P

Tot ziens! :D

Dirk Werner rocks my life

2005-01-22T06:46:34-06:00

Dirk Werner is the author of an excellent functional analysis book, one of the two which I am presently using to prepare for my final exam, Rudin’s almost-classic being the other one. I enjoy Werner’s writing a lot—more than Rudin’s, and to be completely honest with you, he was the first to ever explain the Baire category theorem in a way that I found accessible.

But he also made my day today by having this collection of math photos on his home page. If you have a minute to spare, go there for hilarity.

The following picture was nicked from his homepage:

Attachments

san_francisco_100dpi.jpg

The Dinner Fairy

2005-01-21T07:47:31-06:00

While pretty seventies-looking, I actually like the brand name Penny came up with for their canned food. “Dinner-Fee”, which is German for the dinner fairy, took care of my lunch for today quite marvelously. Thanks, dinner fairy, you’re the best! :P

Note: Whilst writing this, I almost let my dinner fairy burn to the bottom of the pot. Thank heavens for kitchen timers… }:)

Sick before my Exam

2005-01-18T11:11:33-06:00

I was supposed to have my last exam to complete my Diplom degree on Friday, but unfortunately, I got sick with flu-like symptoms on Sunday. One of my examiners has already agreed to move the exam to Monday the 24th, and I hope the other will, too. Now if I didn’t have to complete the grad school applications at the same time, all would be well.

Update: The other examiner has agreed as well, so my exam is officially on Monday the 24th, at 10.

Wish me luck. :)

Sipgate and KPhone

2005-01-13T04:58:48-06:00

The future has just begun, our phone and DSL connections were switched over from the big pink monopoly to Arcor. Unfortunately, their phone rates border on highway robbery (by being at least double as expensive as cheaper providers). So, I decided to give Voice-over-IP a try, using kphone as my SIP client. These configuration instructions were sufficient, except for one minor (but important) detail: You may not use Umlaut characters as part of the Name field in KPhone, otherwise the sipgate server will report back something like (see the KPhone console output)

SIP/2.0 401 Unauthorized
Via: SIP/2.0/UDP 82.82.79.89;branch=RANDOM-CHARACTERS;rport=5060
CSeq: 1722 REGISTER
To: "Andreas Klöckner" <sip:5557868@sipgate.de>;tag=--------------HEX-DIGITS-------.HEX
From: "Andreas Klöckner" <sip:5557868@sipgate.de>
Call-ID: NNNNNNNN@192.168.113.4
WWW-Authenticate: Digest realm="sipgate.de", nonce="------HEX HEX HEX HEX------"
Server: sipgate ser
Content-Length: 0

Hope this helps someone out there.

Excuse me? We’re living in a… what?

2005-01-12T18:30:31-06:00

Well, until just now I was caught in the misbelief that we’re living in a western-style democracy. After watching the news today, this has suddenly changed, to my surprise. :O

Bremen’s senator of the interior is in trouble for authorizing the use of emetics by the police on someone suspected of dealing with drugs. The (foreign) suspect died in the aftermath. (Emetics are, by LEO’s definition, drugs that makes you throw up.) (Google News)

Honestly, I don’t care what the laws in this regard are. I believe that this is torture, and regardless of whether it is common practice or not, it needs to stop at once. I understand that it may be necessary for the police to restrict one’s freedom to move about temporarily and within reasonable limits, but strapping someone to a bed and making him swallow things against his will definitely goes way too far. (Especially if the bed looks like the ones they use for executions. But that is somewhat beside the point.)

But don’t feel relieved if you’re from the US, you got your share of bad news, too. A New York judge ruled today that police don’t need a judge’s assent to slap a GPS tracker on somebody’s vehicle. What is privacy, then, if they can track your every move? Oh, you’re British? Don’t worry, you’re rights are also being violated.

Are citizen’s rights being trampled on increasingly in recent times, just because new (and yet unregulated) technology makes this possible? Something needs to change. Fast. At least if we like our freedoms the way they are right now.

El sistema

2005-01-11T07:16:48-06:00

This image was nicked from here. I recommend this if your Spanish (like mine) is a tad rusty.

A strangely cleansing experience

2005-01-11T06:29:26-06:00

I recently got back to Karlsruhe from being home for the holidays. Besides the usual heap of problems waiting for me at my “normal” home, I was pretty surprised how easily you get spoiled again by the amenities of having a mom around (and a dad, to be fair). For the most part, no worries, no need to prepare breakfast, lunch or dinner. And especially, somewhere clean to live. As I got back to Karlsruhe, I was struck by how messy and dusty I had left the place before Christmas. (Or did it just get dirty and dusty while I was absent? Who knows.)

So, nearly the entirety of my first evening back in Karlsruhe was spent cleaning up and putting everything back in order, but even yesterday, the second day in, it didn’t quite feel like home. After running a few errands, I felt a strange need to rearrange my room—mostly, move the desk somewhere else and make use of some space that was previously unused. Oddly enough, that fixed my not being comfortable with my home.

I guess that was my reaction to feeling overwhelmed with stuff to do (cleaning)—so to speak, feeling powerless against my room. I had to show it who the man is, who’s wearing the pants, who’s in command. 8)

Sometimes life is so weird.

Just don’t use ACPI on old computers

2005-01-09T06:27:34-06:00

A brief hint for those using Linux: If you have an old computer, and Linux chooses to disable ACPI, leave it at that. That is, do not use acpi=force. I just had a case of a computer that appeared to work fine with it. Only that the ACPI option would make the serial port (apparently) lose interrupts, and thus put “breaks” or “pauses” into any longer download. The data transfer would stop, and then, with the next LCP Echo Request on the PPP link, resume. The data transfer was pretty “bursty”, not nearly exploiting the full (stable) transfer rate on ISDN of roughly 8 kilobytes/second.

Comment Spam

2005-01-08T21:56:45-06:00

I’ve had this site for barely two weeks, and even this short time was enough for comment spammers to catch on. The site emailed me with notifications of new comments, and even from the notification emails, I was pretty certain as to what had happened. Somebody was trying to direct your (and my) attention at the greatness of V!@GRA and to how cheap it was to buy from them. Oh, and they would provide a loan, too, should one be needed. With low, low rates. Personally, I find this pretty sick. What kind of person do you have to be to try and exploit a volunteer’s work for your own material benefit? I’m puzzled. Please enlighten me.

As the upshot of all this, anonymous comments, desirable as they may be, are now disabled. Sorry.

What were they thinking?

2005-01-06T04:56:18-06:00

The German tax administration has published software to file your taxes online, and they promise that you’ll get your money back faster if you do that. That sounds good, doesn’t it? But, as often, they missed the mark ever so slightly. Why, oh, why does the thing have to be called “Elster”? OK, it’s an acronym of Elektronische Steuererklärung, i.e. electronic tax return, but the word is also German for magpie. If I was the tax administration, I’d avoid any association with theft like the plague. :puzzled: Beats me.

In any case, the thing is Win32-only and looks like a good example of shoddy engineering. Unfortunately, my parents want it, and I had to install it, including all the Windows Updates it requires. You can hopefully feel my pain.

Review: “Der Palast von Alhambra” (board game)

2005-01-04T16:15:29-06:00

My parents gave me Der Palast von Alhambra by Dirk Henn of Queen Games for christmas, and I’m writing this review to warn you of buying this game. Despite a recommendation from my aunt Conny and the “Spiel des Jahres” (game of the year) prize of 2003, this game could stand improvement in various departments. But let me begin with a short description of what the game is all about.

Given money in four different “currencies” which you basically draw from a few stacks, you buy building blocks of your very own palace. There are a number of different block types, such as pavillon, tower, gardens, etc. The person to amass the majority of pieces of one type gets a given number of points in each of the three scoring rounds. If two players share a majority, they split the points. The person with most points wins.

Here’s my list of pet peeves:

The game is unnecessarily complex. For example, I don’t know why there have to be four different currencies. Additionally, The fact that each player needs to have a reserve stack is a foul hack to compensate for the fact you can build your way into a dead-end.
The game is non-interactive. Except for the shared card stacks and the shared data about who has how many buildings, one player’s actions really do not affect the others significantly. This game might be good if you hate talking while you play.
The game is very dependent on luck. Strategies, at least the ones that I’ve been able to devise, are usually foiled by missing means, such as a missing card of one color. Also, it doesn’t feel like the game rewards significant efforts fairly. For example, f you have invested heavily to build a majority in some building type and someone else matches it, you might end up with fewer points than the next-best player who might only have one building of that type.
The instructions omit important strategic advice. For example, the cheaper a given building block is, the more walled-off it is, typically. Walls severely restrict your building activity late in the game, while making little difference at the outset—but you typically don’t get that until it’s already too late, at least for the first few games.

PS: If you feel that my take of this game is entirely incorrect, please let me know in the comments.

PPS: This post is not meant to give me bad karma with my parents. I appreciated their gift very much.

More Drupal wizardry

2005-01-04T14:37:14-06:00

I’ve been molding this site according to my publishing needs, and in the process of doing so, I’ve ventured into the Drupal source code. I’ve made a few changes that I believe are generally useful. Besides submitting them to the respective maintainers, I’ll post them here for added accessibility.

I’ve modified the Drupal core to provide “Clean URLs” without needing proper mod_rewrite support. Instead, you only need to be allowed set the ErrorDocument to the provided 404.php page. There is a slight complication that the implementation handles properly: HTML forms submitted via HTTP POST cannot use these clean URLs since the form data does not propagate through the ErrorDocument redirection. This More Drupal wizardry against version 4.5.1 of Drupal implements this. (upstream)

Update 1: It turns out that my web hosting company, Alturo, does support this. In order to enable it, you need to say RewriteBase / in .htaccess. They do not document that, though. (In general, they document fairly little.) As a result of this, I won’t be maintaining the Clean-URLs-via-404 patch.

Update 2: I’ve put a modified version of the patch onto the Drupal bug tracker that applies to current Drupal CVS.

I’ve modified and fixed inline.module in several ways in this More Drupal wizardry: (upstream)
You can now specify [ inline:filename.ext] instead of only the number, for added link stability.
A bold, red NOT FOUND message is shown if the given file name or number does not exist.
The module previously generated spaces around the img and a tags. Those are gone.
Images are only inlined if linked to with [ inline:...]. If you say [ file:my.png], this will remain a file link.
The recent fix for multiple links in one node is part of this patch.
(The spaces were necessary to keep inline.module from picking up these example links.)
I’ve added two patches to the title module.

Attachments

Simple CD burning in Linux and the Real World(tm)

2005-01-01T19:28:34-06:00

There’s an issue surrounding recent (2.6.8 and newer) Linux kernels and CD burning. This Gentoo forum thread summarizes the technicalities. Basically, up until 2.6.7, everybody with read rights to a device was allowed to send control messages to it, and thus (for example) overwrite an IDE hard drive’s firmware, which is (obviously) undesirable.. In 2.6.8, a patch was introduced to fix this. This patch was a bit overzealous and disabled all control commands that aren’t obviously harmless, including those used for burning CDs.

The downside to this is that people who want to simply burn CDs and don’t have potentially malicious local users on their computers every day are out of luck: Until the related function in the kernel (called verify_command, in drivers/block/scsi_ioctl.c) is being more and more refined to allow what’s safe, CD burning simply will not work. For example with kernel 2.6.10, it still does not work as it used to, at least not with Debian sarge’s versions of k3b and cdrecord.

The attached Simple CD burning in Linux and the Real World(tm) against 2.6.10’s drivers/block/scsi_ioctl.c basically reverts everything to the pre-2.6.8 situation. It gives everyone with sufficient rights to open a file descriptor on a device the right to send any ioctl, which is potentially dangerous, but more convenient than the current mess.

Attachments

cd-allow-everything.patch

Drupal wizardry

2005-01-01T14:33:34-06:00

If you are trying to set up your personal site using Drupal, these insights might come in handy. At least, those were the points that cost me the most time to figure out.

As far as web writing markup goes, do not use Textile, do use Markdown. Not only does it work properly with Unicode (UTF-8, really), its markup is also more sensible, yet similar. Both Textile and Markdown have Drupal modules available for them. Now you know which one to use. I will gradually migrate my site over to it, but one thing at a time. Josie: Markdown has magically appeared in your modules list, you need to enable it and configure your input filters to use it.

Markdown is extensively documented.

Update: My site is migrated. Death to Textile! }:)

If you (like me) first set up your site on a home server and then want to migrate it onto some paid hosting site, and want to rename Drupal’s tables in the process (because the hosting company only gives you one database), do beware. When you post stories after this, Drupal will refuse to add nodes, complaining about duplicate node IDs. I was first confused into thinking that MySQL’s auto_increment columns were at fault for generating those duplicate IDs or otherwise had not survived the database dump’n’restore, and I spent an hour trying to fix that. Then I realized that Drupal has its own table called sequences for this purpose. To make a long story short: If you rename tables in drupal, don’t forget to also rename them in the sequences table.
“Clean URLs” can make your site unusable. Assume you have transferred a site from somewhere that does support Clean URLs to somewhere that doesn’t. At the new site, Drupal looks at the database , still thinks that they’re enabled, and outputs its URLs in “clean” form, so your admin interface (or, really, any link on your site) will not work. VoilÃƒÂ , your site is unusable. Even if you figure out that the “Clean URLs” setting in Drupal is mirrored in the variables table, changing that in there does not seem to help. A crude hack that saved my life twice is changing includes/common.inc.

Update: See the [link:More Drupal wizardry|post] about Clean URLs via the ErrorDocument directive.

Happy new year to you!

2005-01-01T07:46:24-06:00

I hope you had good holidays and that 2005 will turn out just the way you wanted it to. If there is one thing that I am hoping for the most, then it is that I do get into a good US graduate school for my PhD in the fall. Wish me luck! :D

Update: Josie told me to not wish a happy new year before it’s actually happened. So I’ve probably confused numerous people until now. I’ll only activate this entry once 2005 comes.

Update 2: Here goes.

Update 3: Happy birthday, Josie! :)

Parental Services, Inc.

2004-12-31T05:46:38-06:00

For the past few days, I’ve been doing what I suppose most kids who are at least a slight bit tech-savvy have to do when they get home for the holidays: Fix their parents’ high-tech stuff. The day before yesterday, I finished pushing my parents over to Linux, with a very tiring partition resize process. But it all worked out—Yay. And yesterday, I spent all day connecting my parents’ TV to their (pretty impressive, bass-wise) home stereo system. Obviously, making the actual connection was trivial and done within a few seconds. Hiding the cable was what took the most time, with lots of furniture-pushing, base-board-unscrewing and such things. Tiring, but I’m done for now. On to the pleasurable part of being home. :)

My least favorite day this week

2004-12-27T11:57:52-06:00

Today I took Josie to the airport to go home. As I’m writing this, I’m on my way to my parents’ house in Überlingen. For more than one reason, today wasn’t my favorite day this week.

Let’s start at the beginning: the last few days were extremely cool. I spent christmas with Josie in Ulm, and we hung out, ate and just generally had lots of fun, with very few obligations. Yesterday, Josie and I decided to make tiker.net our full-scale journal site. So, I ordered the upgrade from my web hosting company, and they filled the order almost immediately, even though it was a public holiday. When we got to Karlsruhe, I could of course not resist and just had to set up the site to see if it would work. This took until 3:45 am this morning, and except for a few glitches that I might write about in a separate entry, it worked out great. I had been a bit sad since we left Ulm because it seemed Josie was seriously leaving. Setting up the site helped me forget the sadness for a while (hacking generally does). Unfortunately, the sadness came back right after when Josie was rightfully upset with me for getting in the bed this late.

And then we had to get up at 6 am this morning, so I started the day with a royal 2:15 of sleep. We got to the airport while it was still dark, got her checked in, and hung out at Terminal 2 until it was time to go. Josie was in a hurry, because it was 9:50, and she was supposed to be at the gate at 10. She raced through security, was reminded by the ticket checker lady to say goodbye to me and didn’t even turn around to wave before she disappeared. That sucked. Fast forward to home, through two trains filled to the last spot with me sitting on the floor. As I walked in the door, I found the envelope from ETS that I had been hoping for and dreading for a long time: it contained my Subject GRE results. Turns out that I sucked. 660 out of 800, or better than 60% of the people. I sincerely hope that this doesn’t ruin my chances to get into a usable PhD program. It’s pretty easy to get frustrated and believe that all is lost. However, hope dies last, and I’ll continue to pursue admission in the US. To be safe, I’ll certainly look for opportunities for a doctorate in Germany in the new year.

Right now, I’m looking forward to a few days in the consoling abundance of food and company that is my parents’ house. (I’ll have to be studying, but still.) I just realized that I forgot my tooth brush.

Update: Much less frustrated now than when I wrote this. I still decided to post it, though.

A real journal on the real Internet!

2004-12-26T20:30:52-06:00

So, the site is finally out on the real Internet. Thanks to Josie for co-funding this server. This means that we’re now open 24 hours a day, fast and not behind a flaky DSL line. And the people rejoiced. :D

Math Archaeology

2004-12-08T06:11:07-06:00

While researching the specializations of the faculty at UMD, I stumbled upon a link to the Cornell Digital Math Library, which I find really neat. They have scanned old volumes in mathematics that have gone out of copyright, and made them freely accessible to anyone. No username/password mess. Along with other initiatives, such as ArXiv, this bodes well for the digital commons. Just imagine, all the world’s knowledge, freely accessible to anyone with comparably cheap technical means! Compare that to the effort it took just years ago to get a hold of, say, a copy of a historical mathematics book, such as Hilbert’s or Weierstrass’s. While their treatments of math are certainly not the most “up-to-date”, I usually find it most interesting to find out how someone came up with something in the first place, what ideas they had, which I firmly believe is of some value, even in today’s world of content overflow. If there only were more hours in the day…

Back to current priorities: grad school applications. :O

Another day of administrative joy

2004-12-06T14:02:52-06:00

Another day totally dominated by activities related to applying to grad school. Like trying to get my recommenders to send their letters the right way. Finding out that, in fact, UMD does not support online recommendation. Filling out stupid ETS score report forms using Scribus just to fax them off. (Btw. Scribus is really neat for filling out forms. Put a form PDF in the background, put some text layers and a signature image on it, bingo. Did that twice today. It does have its bugs though, at least the current Debian version does.) And :D Josie’s :D back from Spain. She says it was cool, if not totally stress-free. I’ll go see her tomorrow, or maybe Wednesday, depending on when Alessandro “da rock” F. 8) has time. :)

Here’s one to remind Josie of the long days in Ulm:

Sokoban

2004-12-05T10:47:41-06:00

So, I was wasting time before going to bed last night, like, by setting up this site, and by playing Sokoban. KSokoboan to be precise. Now, I made it up to level 12 of David W. Skinner’s Microban level set, which is rightfully said to be pretty easy. Well, level 12 wasn’t, at least not to me. (I admit that that’s embarassing, but who cares.)

So, I made it to the “I can prove this unsolvable” stage of puzzle solving. But since I reckoned that someone would have noticed that this level is unsolvable, my mind immediately went to cheating. Unfortunately, nobody had published move sequences to solve Microban. I was absolutely clueless as to how this level might be solvable, and I just had to know. So, I wrote a Python program to solve Sokoban levels. You’ll find sokosolve.py and a few demo levels in the sokoban-solver module of my 2004-public arch tree.

The program itself has two nested applications of A*, a usable (but half-baked) heuristic and a complete game model. In just 215 lines. Python never fails to impress me. Attached to this story you can see a dump of the program solving the level that had me so confused.

Attachments

m12-dump

Earthquake

2004-12-05T10:32:24-06:00

While I was setting up this site late last night, I felt like something was wrong. I heard my wardrobe make aching noises, and the ground started moving. First slightly, then more and more. At the end, it felt like our house was swaying from side to side with an amplitude of at least a few centimeters. Pretty scary stuff. Turns out the source of this was an earthquake whose epicenter was actually quite a ways from Karlsruhe.

Funny

2004-12-04T16:36:23-06:00

Someone just posted this to Planet Gnome. Appropriate. :)

Moved in.

2004-12-04T15:48:33-06:00

So, after spending an ungodly amount of time setting this up, I finally feel at home. Time to say something in here. Well, I should have been spending all this time working on my applications to the graduate schools which I plan to go to in the fall of 2005. My current best excuse is to see this as a breather from taking the GRE and the TOEFL this week. I’m feeling pretty good about the TOEFL, but my quantitative score on the GRE was 730—not all that great. I can’t tell you how much I loathe doing calculations in my head. I hope that this blooper won’t prevent me from getting into my chosen schools.

Josie’s in Barcelona with Kathrin at present. Hope they’re having fun. *wave*

Proprietary = bad.

2004-12-04T12:59:18-06:00

Ok, so I’m in the process of setting up this site. I thought, ha, even Josie has a blog. So, for the times when content on my own site is few and far between (like…err…now), I should just syndicate her blog, so I’ll be cool by association. Well, her blog provider, some site called getcrafty, decides that their users can’t let others syndicate their stuff. Like they own her view on life, and all that she wrote. As the title suggests, proprietary=bad.

This reminds of when Yonni’s (Josie’s sister’s) blog just went belly-up and they charged her $35 for just letting her download her own posts. (Josie: If you’re reading this, consider yourself invited to blog here. You’ll get your own site to mess with.)

Blog.^H^H^H^H^H Journal.

2004-12-04T12:39:31-06:00

Well, hi planet. Everybody has a ~~blog~~journal these days, so why not me? Let’s get started.

Update 1: Did anybody notice how cool I am? My blog can be accessed using SSL. Damn secure, man. Damn, damn secure. (Actually, not any more since the site moved to Alturo.)

Update 2: Drupal is pretty fun to mess with. Highly recommended.

Update 3: I really can’t get over this thing being called a “blog.” I’ll call mine a journal. “Blog” just sounds silly, sorry…

Madman bug/feature trackers closed down

2004-10-21T09:04:01-05:00

Trackers are closed now. The Sourceforge bug, feature and support trackers are closed now, effective immediately. Please turn to the mailing list for your support needs. Anonymous posting is allowed, but please request for yourself to be cc’ed. If you have data in the trackers, it has not been deleted and will influence future madman releases. I’m sorry that I had to take this step, but I unfortunately do not have the time to properly maintain them.

Thanks for your understanding.

CVS dropped as madman’s means of revision control

2004-10-19T09:06:46-05:00

Giving CVS the boot. Work on 0.94 is continuing nicely, with Guido as a notable addition to the development team. Shawn is considering a rewrite of the criteria language. And you’ll probably miss all this new development if you only check the SourceForge CVS—it is out of use as of today. The code is now maintained in my public arch repository. See the download page for details on how to get the tree.

Lots of new madman packages unleashed

2004-07-09T09:11:21-05:00

Package frenzy. Within minutes of each other, both Christian Hammers and Jaako H Kyro made packages available for their respective distributions. So now, if you have Fedora Core 1 or Debian, it’s never been easier to get madman.

Thanks to the packagers!

Friendly neighborhood spiderman completes work on madman during night

2004-05-25T09:20:32-05:00

Greetings from the night shift. Guess what. CVS HEAD madman is now based on Scott Wheeler’s TagLib. I’ve decided to delay many other features planned for 0.94 and release as soon as possible. Hang in there…

Rogue library programmer confounds madman author

2004-05-24T09:23:20-05:00

libid3tag. If only anybody had told me this: I was just investigating why some ID3 tags just aren’t written for some MP3s. I stumbled upon this piece of code, hidden deeply in the guts of libid3tag:

  /* hard general case: rewrite entire file */

  /* ... */
done:
  return 0;

To those not fluent in C++, this means that libid3tag’s implementer chose to just forego the hard case and leave that as an exercise to the interested reader. Grrr. :barf: So, I guess it’s going to be TagLib by Scott Wheeler.

A mixed bag of items

2004-05-24T05:26:35-05:00

More Slackware. SlackCare has also made Slackware packages, which they claim are more correct than Adam’s. Find the link to their packages on the download page.

Email trouble. Some people have complained that they were unable to get anything posted to the mailing list or sent personally to me. I’ll look into this, I’ll let you know right here if I find out what’s causing this.

Bugs in 0.93. As was to be expected, there were a couple of bugs in 0.93. Here’s the list so far:

The build system litters system directories with .sconsign files if you let it.
The main build script was lacking a comma.

So, nothing horribly broken. And plus, there are bugs in every release. That’s why there’s always the next one. :) [Apologies to Isaac Richards]

Way to go, Adam!

2004-05-21T09:29:37-05:00

Adam Ward made Slackware packages for madman. Get them from the download area.

New version of madman unleashed upon unsuspecting planet

2004-05-20T09:37:49-05:00

It has finally happened! The madman source has finally reached a state that I deem worthy of being called 0.93. Go grab it while it’s hot!

The most massive changes since the last stable release (which was gasp 0.91.1) include the addition of a very capable AutoDJ (Thanks, Shawn!) and a builtin web server, minimization to the system tray, a new build system as well as over 100 other fixes and improvements.

I hope to have downloadable binary packages for a few distributions soon.

Somebody else’s problem affects madman

2004-03-28T08:38:55-06:00

Is madman randomly crashing on you? The good news: It’s not madman’s fault. :) The bad news: XMMS 1.2.9 has a bug that causes programs controlling it to abort every once in a while. So, before complaining, please upgrade to XMMS 1.2.10, which is available from the usual place.

In other news, 0.93rc1 got quite a bunch of testing, Tim Dreessen deserves a big thank you for all the bugs he found and reported. 0.93rc2 is now officially available through the file release system.

A candidate asking for your support

2004-03-21T08:40:52-06:00

Madman 0.93 release candidate 1 has finally happened, while my thesis is keeping me busy… It’s a pretty massive update, with lots of potential breakage. Please give it a round of testing. Depending on the amount of actual problems, 0.93 will be out very soon.

You can grab this version from CVS as tag MADMAN_RELEASE_0_93RC1 or from the download page.

Major new features in this round: Rewritten AutoDJ (Shawn Willden’s work), Fixes for many string encoding bugs, an AutoTagger, passive popups and a new build system.

SHOCK! Mailing list appears out of thin air

2003-12-17T08:41:02-06:00

madman now has an actual mailing list, with people on it and everything. Wow. Little did I know. ;) Also, madman has an IRC channel to hang out on, that is #madman on irc.freenode.net. I can be found lingering there every once in a while, everybody else is invited to join. The next release, is, as always, late. Shawn and I are discussing some fairly major changes to madman’s core, but I still hope to release some a working drop of the current code as 0.93beta1. Hold tight, it’ll get done.I promise. ;)

Shawn = Mr. AutoDJ

2003-11-23T08:42:41-06:00

0.93 is still well underway. Shawn Willden agreed to rewrite the AutoDJ, and Christian Plagemann made a new logo for madman, which will be the official one from the next release. :) As for myself, I’m down to three simple items on the TODO list before 0.93 goes beta.

Weird. It keeps getting BETTER

2003-11-11T08:44:23-06:00

I just finished a pretty long coding session on madman, and it just keeps getting better and better. Tonight, passive song announcement popups and tray icon support went in. 0.93beta1 will be out real soon nowÃ¢Â„Â¢, there are only a couple of things left to do. Preheat those compilers.

SCANDAL! Release lost to change-o-mania!

2003-10-26T08:45:39-06:00

Oops. Release 0.92 will never happen. There are too many changes in the code base already after the release candidate. Instead, release 0.93beta1 will be the next one out. 0.92rc1 seems to be fairly high-quality, so I’d recommend using that for right now.

0.93 will have lots of new eye candy, will be lots faster and plus, it’ll be out soon.

Candidacy extended!

2003-09-29T09:48:31-05:00

Released version 0.92rc1. This is a source-only test release for the next big version of madman, 0.92. It contains many exciting new features like a web server, remote control and Auto DJ as well as the usual slew of fixes, this time, among others, for unicode cleanliness, UI and invalid characters.

The screenshots as well as all the rest of the site will be updated for the full release.

Please help get this release tested before we put it out there. Update: There’s a small bug in 0.92rc1 that concerns writing Ogg tags containing non-ASCII characters.

Quick fix released

2003-06-28T09:49:21-05:00

Released version 0.91.1 to fix a compilation issue regarding overloaded setAttribute calls. Reported by Frank Baumgart.

BEST madman release EVER!

2003-06-27T09:50:44-05:00

Released version 0.91 with many new features. Here’s the freshmeat announcement:

This release adds fuzzy search, plugins (for CD burning), a separate preferences dialog, and play history collection. Each song now has a ‘Performer’ field as well as play statistics and can be rated on a scale of 0 to 5 stars. Easy mass tagging was implemented. A configurable number of backups will be kept of a database. The build process now uses autotools. Searches were made asynchronous for much faster response times. Startup time was dramatically reduced.

Mr. Quicker-fixer-upper unleashed

2003-02-27T08:52:04-06:00

Released patch version 0.90.1 to fix dependency on Qt 3.1.This new release only requires Qt 3.0 or higher.

I humbly present…

2003-02-26T08:53:14-06:00

Released initial version 0.90.

Andreas Klöckner's web page

The State of OpenCL for Scientific Computing in 2018

OpenCL as a Standard

Implementations

Nvidia GPU

AMD GPU

POCL CPU

Intel Integrated GPUs

Apple GPU/CPU

Intel CPU (Updated August 2019)

Shipping Software

The Future

Conclusions

Nvidia Highlights PyCUDA/PyOpenCL Applications

Tutorial: Domain Specific Languages and Code Generation in Python

Dr. Dobb’s article series on PyOpenCL

The Devil’s dictionary of programming

New (Py)OpenCL tutorial

Quote of the day

Job? UIUC!

GPUs and Python at PyData NYC 2012

PyCUDA and PyOpenCL at Nvidia GTC 2010

Attachments

PyCUDA at PyCon Quattro (Italy)

Putting the “Hard” in “Hardware”

Why is GPU Computing significant?

You are being preyed on

Submitted: PyCUDA: GPU Run-Time Code Generation for High-Performance Computing

Abstract

PyCUDA talk at Nvidia’s GPU Technology Conference

Attachments

Nena vs Eminem

SciPy‘09: Advanced Tutorial on PyCUDA

Introducing UIH: Math + Unicode, Happy Together

Debug Python in Style

Git Commit Timelines

1 Billion seconds on this Earth

World Builder

Credit Crisis? What Credit Crisis?

boostmpi: Initial Release

CodePy: Initial public Release

Talk slides: PyCuda@MIT

Submitted: “Nodal Discontinuous Galerkin Methods on Graphics Processors”

Abstract

Boost Numeric Bindings 20081116

PyCuda 0.91

Brown Beamer Template

Download

Instructions

Attachments

Reproducibility in Scientific Computing

Joint release 0.92 of PyUblas and PyUblasExt

TagPy 0.94.1

Release 0.91.2 of Synoptic

First real public release of Synoptic

Boost Bindings 20080405

On Hybrid Scientific Codes, Part I: The Idea

Getting the Most out of the Home Row

Window management

Internet

Desktop environment

Barf Detergent and other Successes of Modern Marketing

Hey, Americans!

Forums!

Visualizing Greenhouse Gases

TagPy 0.93

Yuck!

Möbius Transforms

TagPy 0.92 released

Not enough Flash video yet?

PSI Smilies in Gaim/Pidgin

Attachments

Fermat’s Last Python Script

What they don’t tell you about VTK XML binary formats

Books, Books and more Books

Singing the praises of

New release of Boost.Bindings tracking Boost 1.35 CVS

D*mn sp*mmers

Trip to the Northwest

August 1

Dmn spmmers