www.arareko.net

"Only the Spirits of the Air know what awaits me behind the mountains, but still I travel onwards."

- Old Eskimo proverb

Archive for June, 2005

Sunday, June 12th, 2005

Aunt Yola’s requiem (a.k.a. Judith’s wedding)

I don’t plan to get into the full details about it, but…

The soundtrack at the church’s ceremony was performed by the Classical Orchestra of México. The highlights were: Exultate Jubilate K-165 and Messe K-626 “Domine Jesu Christu”, both from W.A. Mozart. No need to say who selected the music score, right?

As always, Beto and I managed to be the center of disorder at the party, with the exception that this time we didn’t got very drunk, so we behaved a little bit better: we only brought our Chuck Taylor Converse shoes for the dancing ritual…

Aunt Yola was very happy and to say it in some way… she looked quite decent. Her hair was arranged for the occasion (with all the dead nature a wedding could require…) and with an excellent pair of tennis shoes which allowed her to jump all that stupid “music for the masses” that is played live at weddings…

Of course, she was caught by the camera when she tried to escape from the party… riding the bottle of tequila! Anyway… Congratulations Judith!!

Saturday, June 11th, 2005

hdf5′s Segmentation faults

Some time ago I wrote about my porting process of Zori (Alan‘s software for Monte Carlo simulations using parallel computing) to FreeBSD. At the time, some ports (mpich & hdf5) were broken and needed an update…

Time passed and those ports were updated by their maintainers. Now they compile OK with the “normal” configure options. I could install lam7 and mpich for testing hdf5 parallel support. Here’s where things got stuck again…

I tweaked hdf5 port’s Makefile to enable parallel support and to use any of the MPI compilers:

# ex:ts=8
# New ports collection makefile for: HDF5
# Date created: Jul 7, 2002
# Whom: ijliao
#
# $FreeBSD: ports/science/hdf5/Makefile,v 1.33 2005/05/11 03:44:40 ijliao Exp $
#

PORTNAME= hdf5
PORTVERSION= 1.6.4
CATEGORIES= science archivers graphics
MASTER_SITES= ftp://hdf.ncsa.uiuc.edu/pub/outgoing/hdf5/${PORTVERSION}/

MAINTAINER= koziol@ncsa.uiuc.edu
COMMENT= Hierarchical Data Format library (from NCSA)

USE_REINPLACE= yes
USE_LIBTOOL_VER=13
LIBTOOLFILES= configure c++/configure
CONFIGURE_TARGET= --build=${ARCH}-portbld-freebsd${OSREL}
CONFIGURE_ARGS= --enable-cxx --enable-production --enable-threadsafe --disable-debug --with-pthread
ALL_TARGET= lib progs
INSTALLS_SHLIB= yes

OPTIONS= PARALLEL "Build with parallel support" Off

.include

.if defined(WITH_PARALLEL)
# Build with lam7
BUILD_DEPENDS+= mpicc:${PORTSDIR}/net/lam7
CC= /usr/local/bin/mpicc
# Build with mpich
#BUILD_DEPENDS+= ${LOCALBASE}/mpich/bin/mpicc:${PORTSDIR}/net/mpich
#CC= /usr/local/mpich/bin/mpicc
.endif

post-patch:
@${REINPLACE_CMD} -e 's|-lpthread|${PTHREAD_LIBS}|g' ${WRKSRC}/configure
@${REINPLACE_CMD} -e 's|-D_THREAD_SAFE|${PTHREAD_CFLAGS}|g ; s|[\$$]LDFLAGS -pthread|\$$LDFLAGS ${PTHREAD_LIBS}|g' ${WRKSRC}/config/freebsd
@${REINPLACE_CMD} -e 's|-D_THREAD_SAFE|${PTHREAD_CFLAGS}|g ; s|[\$$]LDFLAGS -pthread|\$$LDFLAGS ${PTHREAD_LIBS}|g' ${WRKSRC}/c++/config/freebsd
@${REINPLACE_CMD} -e 's|-O3||g' ${WRKSRC}/config/gnu-flags

post-install:
.for i in gif2h5 h52gif h5debug h5diff h5dump h5import h5ls h5repart h5jam h5unjam
@${STRIP_CMD} ${PREFIX}/bin/${i}
.endfor

.include

For both MPI compilers the configure stage went perfect (parallel support enabled & MPI compiler detected). The problems arised at the make process.

hdf5 compilation brokes with any of the MPI compilers, but at different stages. Here are the last lines of the output using lam7 compiler:

./H5detect.c:1385: warning: cast discards qualifiers from pointer target type
./H5detect.c:1385: warning: cast discards qualifiers from pointer target type
./H5detect.c:1385: warning: comparing floating point with == or != is unsafe
./H5detect.c:1385: warning: padding struct to align 'x'
./H5detect.c: In function `detect_alignments':
./H5detect.c:1408: warning: padding struct to align 'x'
./H5detect.c:1409: warning: padding struct to align 'x'
./H5detect.c:1410: warning: padding struct to align 'x'
LD_LIBRARY_PATH="$LD_LIBRARY_PATH`echo -pthread | sed -e 's/-L/:/g' -e 's/ //g'`" ./H5detect > H5Tinit.c || (test $HDF5_Make_Ignore && echo "*** Error ignored") || (rm -f H5Tinit.c ; exit 1)
Segmentation fault (core dumped)
*** Error code 1

Stop in /usr/ports/science/hdf5/work/hdf5-1.6.4/src.
+ exit 1
*** Error code 1

Stop in /usr/ports/science/hdf5/work/hdf5-1.6.4.
*** Error code 1

Stop in /usr/ports/science/hdf5.

And with mpich compiler:

../../src/H5FDmpi.h:74: error: initializer expression list treated as compound expression
../../src/H5FDmpi.h:79: error: `MPI_Datatype' has not been declared
../../src/H5FDmpi.h:80: error: `MPI_Datatype' has not been declared
../../src/H5FDmpi.h:80: error: ISO C++ forbids declaration of `btype' with no type
../../src/H5FDmpi.h:80: error: ISO C++ forbids declaration of `ftype' with no type
../../src/H5FDmpi.h:86: error: `MPI_Comm' does not name a type
*** Error code 1

Stop in /usr/ports/science/hdf5/work/hdf5-1.6.4/c++/src.
+ exit 1
*** Error code 1

Stop in /usr/ports/science/hdf5/work/hdf5-1.6.4/c++.
+ exit 1
*** Error code 1

Stop in /usr/ports/science/hdf5/work/hdf5-1.6.4.
*** Error code 1

Stop in /usr/ports/science/hdf5.

Again, I had to e-mail hdf5‘s port maintainer and ask him for guidance. Zori porting will have to wait a bit more…

Friday, June 10th, 2005

Bioinformaticians in Molecular Biology Labs

I found this essay about bioinformatics via Technorati.

Even though it’s small, it addresses some of the common questions about the area. Interesting…

Thursday, June 9th, 2005

Audioscrobbler

A few days ago, Jordi suggested me to try Audioscrobbler, a service that according to its website:

…builds a profile of your musical taste using a plugin for your media player (Winamp, iTunes, XMMS, etc.). Plugins send the name of every song you play to the Audioscrobbler server, which updates your musical profile with the new song. Every person with a plugin has their own page on this site which shows their listening statistics. The system automatically matches you to people with a similar music taste, and generates personalised recommendations.

So I signed up for an account and filled my profile with the usual stuff. After a few minutes of browsing the site, I arrived to the developers section and saw the possibility of using my Audioscrobbler RSS feed to create a nice add-on to the blog (see the Recent Music section in the right menu). There you’ll be able to know what music I’m listening at the moment and “maybe” get an idea of my current mood…

Some friends are also trying the service. I suggested Ion that my simple add-on could be a nice plugin for his great Jaws software. He considered the idea

Tacvbo had some scripting fun retagging all his music with ID3v2 tags. He always thought that ID3v2 tags would be useless to him, until he discovered that Audioscrobbler plugins needed them!

Now we’re waiting to see what kind of statistics this service generates and most important, what music recommendations it gives to us. In the meantime, we’ll keep on playing…

Tuesday, June 7th, 2005

Mr. Incredible?

Somebody sent me this test. I thought I’d be more like Jack-Jack (unpredictable & with temper), but he wasn’t included in the test… I don’t really feel like the result it gave me. What do you think?

Which Incredibles Character Are You? brought to you by Quizilla.

Monday, June 6th, 2005

The Loneliness of the Long Distance Runner

Beto and I started today our running training.

We must get strong enough for the México-Cuernavaca trail-running marathon (46 km). We plan to run it by September, when (we suppose & plea) the weather would be nicer, with rainy and colder days…

The next song (a nice suggestion by Octavio) has to be our mantra for the next months:

iron maiden – the loneliness of the long distance runner

tough of the track,
with the wind,
and the rain that’s beating down on your back.
your heart’s beating loud
and goes on getting louder
and goes on even more till the sound
is ringing in your head,
with every step you tread,
and every breath you take,
determination makes
you run,
never stop,
gotta win, gotta run till you drop,
keep the pace,
hold the race,
your mind is getting clearer,
you’re over halfway there but the miles
just never seem to end
as if you’re in a dream,
not getting anywhere.
it seems so futile.

run, on and on,
run, on and on,
the loneliness of the long distance runner.

i’ve got to keep running the course,
i’ve got to keep running and win at all costs,
i’ve got to keep going, be strong,
must be so determined and push myself on.

run over stiles, across fields,
turn to look at who’s on your heels,
way ahead of the field,
the line is getting nearer but do
you want the glory that goes,
you reach the final stretch,
ideals are just a trace,
you feel like throwing the race,
it’s all so futile.

run, on and on,
run, on and on,
the loneliness of the long distance runner…

Sunday, June 5th, 2005

Some Vespa adventure…

Today, Beto suggested to go training to Ciudad Universitaria riding the well-known Mambo-Tango and her sister (my uncle’s red Vespa).

The big tour began from Santa Mónica (at Beto’s place). Periférico wasn’t so empty, so we had to take care while being between all the cars. I had to make an emergency stop at Av. Reforma (I ran out of gas!). Once the tank’s reservoir was set up, we continued to Av. Revolución to the gas station.

We arrived to Espacio Escultórico by mid-day and got our usual sunburning session, then we went to Filosofía boulders. We were very dehydrated and pumped from the 2 laps we made at E.E. and somehow we managed to climb some new long hanging traverse at the parking lot’s boulder

The day was so hot that we drank 2 liters of fruit water at La Michoacana in Av. Universidad, then we headed to Coyoacán to continue the tour. There we quickly decided (after seeing that El Jarocho was crowded with beautiful women) to sip a cup of mocha and have some rest. Yeah, nice excuse…

After the nice rest, we continued by Av. Insugentes to La Condechi to get some more exciting landscape (yes, we can’t get enough… never!) and then returned to Santa Mónica. We finished the tour by 9:30 p.m. and called it a very fun & productive day. Now it’s time to sleep!

Saturday, June 4th, 2005

¡Feliz cumpleaños, tarado!

A ver animal, teníamos un compromiso pendiente para antes del día de hoy…

Por causas del destino (comúnmente llamadas: pretextos) no se pudo concretar. Espero que antes de que transcurra otro año, lo cumplamos. Septiembre parece una buena época ¿no?, al parecer con más lluvia y menos calor, nos da tiempo de entrenar… Ponte a chingarle que no quiero verte chilloteando a medio cerro, jajaja

¡Un abrazo socio! ¡¡Felicidades!!

Wednesday, June 1st, 2005

500 Workunits!

Back to my SETI@home statistics. I’ve got a new certificate for completing 500 workunits for the project:

Some of the statistics:

Name: arareko
Results Received: 507
Total CPU Time: 2427 hr 50 min
Average CPU Time per work unit: 4 hr 47 min 19.1 sec
Average results received per day: 3.64
Last result returned: Wed Jun 1 13:07:44 2005 UTC

Currently processing unit No. 508