header

Categories::

People::

stesie
r4m (de|en)
backhaus
bunnylabs

Projects::

SAGE
Kopete SILC
Junge Linke (de)

Bubble::


Mon, 09. Apr 2007

SuperKarambaR

Again, you have to write some code during EasterHegg. As I was way too lazy to do actual work I geeked out and played with Superkaramba - the widget engine for the KDE desktop. So I hacked together two widgets to display flickr images on your desktop. Karambrrr displays one image at a time and PhotoStackr displays - you guess it - a stack of images.

screenshot

Btw. when I say “hacked together” I mean it, so things may go wrong for you … and, yes, I am aware that these names are ridicilous.

posted at: 19:26 :: permanent link

Tue, 21. Nov 2006

Memory Profiling

One thing SAGE is lacking is the ability to profile memory usage of a calculation. There are some tools available for Python allowing to inspect the used memory at a given point in time, however nothing seems to exist to draw pretty pictures of mem usage during the calculation. A workaround is to use a shell script sampling the SAGE process’ RAM usage every couple of seconds and write that information to some file. Such a script was posted to this forum and after some slight patching + wrapping. the memory usage of my F4 implementation computing a lexicographical Gröbner basis for a CTC ideal with B=4 and Nr=3 looks like this.

F4 memory consumption

The graph doesn’t say much - because the Python garbage collector might interfear - besides that the memory consumption is 3.5 GB max for this Gröbner basis calculation. Still, I guess 2.x GB after the calculation finished points to a memory leak somewhere.

Elsewhere: Looks like Uni-Bremen is benchmarker’s heaven.

posted at: 15:21 :: permanent link

Mon, 23. Oct 2006

MAGMA Online Calculator

In case you haven’t noticed there is a free MAGMA Online Calculator (which is based on code by SAGE’s William Stein btw.). Also in case you haven’t noticed Python has pretty neat http support built-in. So some lines like 

import urllib
cmd = urllib.quote(cmd)
url = "http://magma.maths.usyd.edu.au/calc/?input=%s"%cmd
fp = urllib.urlopen(url)
answer = fp.read()

allow you to use that online calculator in your Python program if you are willing to accept the horrible lag and the 20s timeout. If you put a couple of pre-/post-processing lines around it you end up with something somehow usable. Of course, you could also use it as a fall back to interface (a stripped down) MAGMA from SAGE if there is no local MAGMA installation available. But that would probably only cause bad vibes with the MAGMA people.

posted at: 12:52 :: permanent link

Tue, 17. Oct 2006

The Need for Speed

First, I am aware that the current version of MAGMA is v2.13-5 and that MAGMA’s F4 implementation was heavily optimized in version 2.11-8 as e.g. shown here. That said, I had the chance to play with MAGMA v2.11-2 a bit and was actually a bit surprised how well the now truly open-source Singular 3-0-2 performed compared to it. # construct a random CTC ideal with blocksize=10 and number of round = 1
sage: F,s = ctc_random_MQ(Nr=1,B=10,order=”degrevlex”)

sage: time _ = F.ideal().groebner_basis(“magma:GroebnerBasis”)
CPU times: user 2.60 s, sys: 0.17 s, total: 2.77 s
Wall time: 124.8 # MAGMA’s GroebnerBasis

sage: time _ = F.ideal().groebner_basis()
CPU times: user 2.36 s, sys: 0.12 s, total: 2.49 s
Wall time: 42.01 # Singular’s groebner
 Again, I know that the most recent version of MAGMA is much faster than the most recent version of Singular when it comes to Gröbner bases and I know that Allan Steel did an amazing job with his F4 implementation. However, the benchmarks shown above indicate that the open-source world is not as far off as I used to believe. Feel free to criticize my benchmarks, if they are wrong, I’ll correct them as soon as possible.

Elsewhere: You might be interested in this document which lists tips and tricks how to make your Pyrex code faster and should be of interest beyond SAGE. Also my quick comparison of malloc replacements could be of interest outside of SAGE - now that omalloc is GPL’d.

posted at: 18:01 :: permanent link

Thu, 14. Sep 2006

SAGE Days 2

are coming up and I’m going to be there. I’ll give a talk on Gröbner bases in SAGE and two smaller, informal talks: One on Pyrex and one on Algebraic Attacks on Block Ciphers. Spoiler-warning: These links point to my beta or alpha quality slides.

Speaking of slides: I adapted the nice Oxygen Beamer theme to sport some SAGE references:

screenshot

You may find the respective sources here, here, and here.

posted at: 01:13 :: permanent link

Mon, 29. May 2006

sage.math.washington.edu

“This is computer for very open collaboration among SAGE developers and testing of intense SAGE calculations. It is a special-purpose 64-bit computer built by Western Scientific that has 64GB of RAM and 16 AMD Opteron cores.” (sage.math.washington.edu) … and I have shell access. You may browse everybody’s home directory which of course includes mine.

posted at: 14:15 :: permanent link

Wed, 26. Apr 2006

SAGE AJAX Calculator

William Stein released a new AJAX based SAGE calculator for those desiring to try out SAGE without installing any extra software. It is currently a bit slow as every character does a complete round trip between the server and the client before being displayed.

Elsewhere: Check this neat comparison of computer algebra systems.

posted at: 00:51 :: permanent link

Fri, 14. Apr 2006

Native Compiler Toolchain on OpenZaurus

These are some tips and tricks on how to setup a native compiler toolchain right on your Zaurus (SL-5500) running the OpenZaurus distribution. I found it to be a painful process and will try to ease the pain for you a bit. All this applies to OpenZaurus 3.5.3., the situation might be better with 3.5.4.

Some packages are available via the OpenZaurus feed for your platform and the quickest way to add the missing ones is to use the Debian ARM stable repository. You can install any debs like you would do with ipkgs, however some postinstall scripts might fail: Ignore or remove them.

posted at: 15:27 :: permanent link

Sun, 09. Apr 2006

GnuPG 1.4.2.2 and 1.4.3 for OSX

As the MacGPG project does not offer binaries for the critical GnuPG security update (1.4.2.2) but only instructions how to build the new package I offer binaries of GnuPG version 1.4.2.2. (critical bugfix) and the new version 1.4.3. I couldn’t believe it myself but there seem to exist Mac users without a working C compiler installed. So here is what you get:

Please get the signature (.asc) as well and verify (as the 1.4.2.2 fixes a critical bug which makes this check useless you actually cannot verify the signature) the signature by typing: 

gpg --verify 'TAR.GZ or PKG file'

Please note that these signatures only state that I had possession of these files. So you see the same files as I did when uploading them. Neither do I know what Till did to his binaries nor do I fully control BigMac to rule out any attack.

Finally there is no reason to trust me so build it yourself.

Final words: Backup everything before using these binaries as I do not know if they work at all and please report any bugs you stumble across back to me. Don’t blame me if your computer explodes.

posted at: 16:26 :: permanent link

Fri, 07. Apr 2006

Howto use your favorite Python debugger with SAGE

Even though SAGE or iPython has a built in command line debugger here is how to use DDD or Eric3 with pieces of SAGE written in Python or scripts written in SAGE:

DDD

Use a shell script like this or type into a shell: 

#!/bin/bash

# change this to your SAGE_ROOT
export SAGE_ROOT="Your SAGE ROOT here"

# setup environment
source $SAGE_ROOT/local/bin/sage-env
# make sure to add pydb the enhanced Pyton Debugger
export PYTHONPATH=$PYTHONPATH:/usr/share/python-support/pydb

# and run DDD
ddd --debugger "python /usr/bin/pydb"

Now you should be able to open a SAGE source file (use “Open Source” for that), press “Run” and “Continue” and your SAGE script should be running. I checked with this a tiny example like this: 

from sage.all import *
r=MPolynomialRing(GF(127),2,'x')
r.gens()

Breakpoints, displays (only string representation) and backtraces worked.

Eric3

You will need working sage-pyton and sage scripts somewhere in your $PATH for this one. Write a shell script like this: 

#!/bin/bash
export SAGE_ROOT=/opt/sage/
source $SAGE_ROOT/local/bin/sage-env
sage-python ${*}

Configure Eric3 to use this script as a special Python interpreter. Make sure to choose “without Qt” as the “Debug Client” and it should work. Additionally you may want to generate an .api file for Eric3 by calling the gen_python_api.py script found at the Eric3 website with the appropriate environment variables set up (use sage-env for this). Or use this one for SAGE v1.1.2.

posted at: 17:55 :: permanent link

Sat, 01. Apr 2006

Current state of the kopete-silc-plugin

There have been some nice bugfixes lately as stesie and I picked up the kopete-silc-plugin again.

At least my planned features are:

If something is missing, let me know. Use it and report bugs.

posted at: 17:19 :: permanent link

Tue, 24. Jan 2006

Software for Algebra and Geometry Experimentation

Speaking of SAGE: It’s a pretty cool open-source computer algebra system. Instead of being written from scratch it uses many existing special purpose systems (Singular, GAP, PARI, NTL) to get to the juice. But I guess the best thing is that it is written in Python, i.e. it is using a real object-oriented programming language with all bells and whistles. No for loops starting with 1 but an object-oriented interface to mathematics. Download it, read the reference manual, or try it online.

posted at: 23:09 :: permanent link

Python never, ever frees memory

Just found out that Python doesn’t release any memory it has allocated, but keeps it for later use. It is designed to do so. So if you - let’s say - use SAGE to construct a MPolynomialRing with 20000 variables for some reason and you’re happy enough to have 2GB of RAM to support that, you’ll have a sage.bin process hanging around never releasing those 2GB until you kill it. See Evan Jones’ website for some more details.

posted at: 23:00 :: permanent link

Fri, 06. Jan 2006

kopete_silc Development

As reported elsewhere I’ve joined stesie in the kopete_silc team. We are developing a Kopete plugin for the beautiful SILC protocol. SILC is basically a chat protocol with heaps of good crypto poured in so encrypted group chats, file transfers and so on are possible. The kopete_silc plugin so far supports (group) chat, basic file transfer, some channel management, and some public key handling: It is usable even though there are SEGFAULTs. You might want to report some unknown on the mailinglist.

Stesie provides public WebCVS access and some Debian packages for Sid. I’m going to provide packages for Debian Sarge.

posted at: 23:16 :: permanent link

Wed, 04. Jan 2006

ValueKonverter 0.1.3 released

Now the Widget actually scales when the size of the application is changed. It took me a while to figure out how to do it. It can be found at the usual places.

posted at: 23:11 :: permanent link

Wed, 14. Dec 2005

ValueKonverter 0.1.2 releases

Again two bugs have be squashed:

Grab sources here and Debian packages here (sarge) and here (sarge + KDE 3.4).

If you’re running any other Debian distribution a apt-get build-dep valuekonverter followed by a apt-get -b source valuekonverter should do the trick.

posted at: 11:09 :: permanent link

Wed, 19. Oct 2005

ValueKonverter 0.1.1 released

Two bugs have been removed:

The libdisasm bug has not yet been solved but the debian package features some accurate build dependencies and conflicts now. Grab the source here and debian packages either in the Sarge repository or the custom one.

posted at: 12:17 :: permanent link

Mon, 10. Oct 2005

ValueKonverter released

Even though there are binary and source packages for the ValueKonverter available, there are some issues left to be solved

posted at: 01:32 :: permanent link

Valid XHTML 1.0 Strict Valid CSS! blosxom