Advogato blog for gilbou

11 Mar 2008

Tue, 11 Mar 2008 14:55:03 GMT

been watching the sarah lacy disaster. she should get fired. how can someone so incompetent and selfish claim herself to be a journalist... i have never ever seen an interview with someone behave so unprofesionally, talking of herself and always making closed questions. the facebook guy even told her : "perhaps it would get more interesting if you would ask.. some questions" (!!) she clearly wanted it to be all about her and she later even told the public to "screw" itself. this my friends is having yourself revealed to the world for what you really are.she kept interrupting the guy all the time (the crowd started to talk about it and they could not believe it). she asked no interesting question, none. she kept making some bold statements, talking to the crowd in what i have to call a vulgar way. worst interview ever. just watch the mess and judge by yourself.

lucasr : check joshua redman's albums. i started jazz with those. love it.

5 Mar 2008

Wed, 5 Mar 2008 19:16:00 GMT

got a new video card, a nvidia 8600 gt. the overclocked model from club3d (it's in sweden i think). it's not really to get a better world of warcraft from linux (works like a charm by the way, on debian etch 4.0 32 bits + cedega 6.0.5). i had a 7900 gt previously and could not try to play with Cuda. at work (cnrs) we got a cluster. and its admin bought two tesla cards from nvidia which are dedicated to mathematics (128 threads par cpu, two per card). the good thing about it and compared to the ati solution (this is gonna change very soon i think under's amd guidance) is we use nvcc which are wrappers (not very clean if you dive your nose in) around gcc. it's moving fast. first, each card offers us 500 Gflops so we have a theorical 1 Tflop with just two cards loaded on two machines. it's not very portable and you really need to know the architecture behind it to have it work properly. the first week we got the cards was funny. once the card was plugged the machine was not even booting. it only worked on a very very few specific dell machines. after a nvidia developer gave us some daily code, it finally worked. well the machine was booting but once the nvidia's cuda driver was installed it did not see the card. it took over two weeks to have the cards available from debian on the cluster ;) and of course all pre-packaged stuff is only for redhat, suse.. well. i am a bsd monk so i really dont like linux. but god i do love debian (i never thought i would ever say that). the nice thing about the whole thing is you buy a cheap video card, and you can from linux write cuda code. then, we just move that code to the cluster where the tesla cards run it with their raw computing power. that's cool to be able to compile and run it on small scales and then just load it on the cluster for full range calculations. the only thing that really annoys me if that most Cuda functions are UPPERCASE like this and i got a happy hacking keyboard without capslock key... :/ well tonight i'm going to install and try at home for the first time to use cuda on my 8600 and i'd like to implement something easy like a md5 hash or so. and then upload it to the cluster but it's not really something you can get scalar execution from (well, it's a stream you work upon so it's gonna be disk-bound in performance).

another problem are mathematicians. most of them if not all know how to use matlab or libraries to their fields of work (petsc, and so on) and for cuda to really work it has to be as transparent as possible. most know how to write C code, but we're not really seeing them wanna learn the inners of those tesla cards to get real fast, optimized scalar code to run on those gpus. we need either fortran libraries that use as much cuda as possible or to have as much as possible matlab functions to be redirected to the gpu through some addon/lib for matlab. most of our mathematicians do code, but their work is mathematics. not spending days the nose deep in C with the tesla architeture in mind for best performance. hope nvidia will improve this on those parts :)

5 Feb 2008

Tue, 5 Feb 2008 23:53:29 GMT

im pissed off. really. i got an ibm thinkpad x30. works fine under netbsd. i can put a YEARS old netbsd or openbsd or whatever on this machine, and everything works. since last week i have tried on this machine : debian 4.0 etch, lenny and i just finished install of unstable. no network. nothing. the ethernet plug is there, lighted, and dhclient fails and never gets any answer. wifi doesnt work either. how can a netbsd years old have everything to work from this piece of crap, and the latest 2.6.22-3 kernel and debian unstable cant even get a fscking dhclient to work ? it's the third time i try to install debian on this machine and it either gets stuck on the login screen of gnome because the orinoco drivers gets nuts (debian lenny, debian etch) or the eth1 gets no answer from the dhcp, while another box on etch 4.0 much more recent (asus n2m32 sli...) works fine. i really hate, hate this. i cant believe this crap recent kernel is unable to have this machine work on default install. been searching for over an hour with google, there's nothing about the thinkpad x30 having problems with debian anywhere. ethernet is intel 82801CAM (ICH3) PRO/100 VE (LOM) (rev 42). this is why i hate linux, and love debian at the same time. debian is wonderful at work, we just put it everywhere. but this piece of crap keeps burning me when any several years old or the latest snapshot of any bsd just works fine with no config, no kernel compilation out of the box. and for the third time tomorrow i guess i will have no choice but to remove debian and get back to bsd. i feel like killing penguins at close range with great violence and blood everywhere.

searching again and again i found people with the same problem http://ubuntuforums.org/archive/index.php/t-182618.html http://www.webservertalk.com/archive99-2007-6-1910758.html

5 Feb 2008

Tue, 5 Feb 2008 14:29:39 GMT

installed under debian unstable a xps from dell, model m1730. freaky machine. it runs a dual core from intel, and is loaded with two nvidia 8700M GT video cards with SLI link and.. a PhysX card. my god. after a 2.6.22 kernel install, I added the kernel headers and installed the nvidia 169.09 drivers. what is strange is on the day before, I did the same driver install on a very same debian unstable snapshot, and two scripts located in /etc/init.d, namely nvidia-kernel and nvidia-glx, were screwing the driver on each reboot. I had to rm those two files to keep the 169.09 working on the dell laptop (a d630). on the xps, after the kernel module compilation and install, opengl 32 bits install, those two scripts were not present in /etc/init.d. if you are installing the nvidia drivers and they break on reboot, rm the nvidia-* scripts from /etc/init.d ;)having the SLI to work was very easy. edited the xorg.conf and forced the MultiGPU option to yes and selected AFR (alternate frame rendering) as method rather than Auto because in this mode the SLI/Dual GPU rendering will only work if the application being run is recognized.

been installing OCS and Munin-client on every box on the network at the CNRS here and it's a pretty good stuff. Problem is the load on the processor on the server side goes up for each machine we had to Munin... and it goes up a little too fast 8)

26 Jan 2008

Sat, 26 Jan 2008 00:43:30 GMT

up to now i used cedega from debian. worked ok but i wanted to check another distro. so i downloaded a fedora core 8 because it has a dvd so i could avoid downloading most packages to install (you have such considerations when you only have a slow 1 mbit link). what a piece of crap. i install it, then the update says there's updates to do. and of course, it fails. it first crashes and asks me to report a bug, then does not crash but is unable to resolve dependencies, then it seems to work but asks me a cdrom that does not exist. i installed from dvd and it asks me for cd number 1. i insert the install dvd, and it refuses to use it. what the fuck ? so they currently distribute an operating system install media which if you install in the default setting without touching anything, is unable to update, does not even recognize its install media and asks for a disk you dont have. yeah, sure. well after 30 min install and 10 minutes of crashes, error and stupid way of working, i got rid of that crap. im gonna put the debian back where it was, and aptitude update/upgrade and basta.

28 Dec 2007

Fri, 28 Dec 2007 16:09:07 GMT

gnupg ftp seems down. i found some mirrors, downloaded two versions of gnupg-1.4.8.tar.gz but none do check the md5 sum as reported on gnupg.org. been forced to go back to a verified version of gnupg-1.4.6.tar.gz i keep at home. be cautious if you download or compile gnupg : check the md5 sum and sha1 sum and be prudent. could be just some problem but having the main ftp down and downloading files that pass gzip -t but fail on hashes.. fishy ?

18 Dec 2007

Tue, 18 Dec 2007 23:19:56 GMT

spent most of the day reworking my console and screen settings. wandered inside netbsd 4,0_rc5 and found with joy sendmail seems really gone for postfix by default (at last that sendmail crap is gone). checked the zsh but still keeping my pdksh (i dont even recall why i love so much the korn shell, but there was some good reason i seem to have forgot). wondering how to add to sftp or scp a stdin > file thru ssh connection (aim is to dd something, have it piped to scp/sftp in batch mode without interaction and have the file being dropped on another box). probably going to end with a nfs exported thru ssh tunnel and dd dump from disk/part to some file inside this nfs :/ mm and if you have netbsd and a pc with suspend to ram you can help: testers are required to test the new acpi code for suspend to ram on netbsd.. check mailing-lists about it. christmas is coming and i already know i will have (1) no vacations and (2) be doing a full audit at work before i got to write doc, create a db to dump audit data and dmesg (already spent one full day working on a bootable cd that boots and dumps a file with first ethernet's mac as filename with dmesg inside with a combo of boot-cd that dumps dmesg to floppy which stores while i'll go from box to box). at least i should be alone at work with none of those pesky humans around me, precious server cpu cycle leechers. and it's cold, i get to work by bike and need over one hour just to be able to type anything properly. sucks.

10 Dec 2007

Mon, 10 Dec 2007 20:58:27 GMT

spent part of the day trying to get freebsd to install on a thinkpad x31. we did use pxe to boot the x31 from a netbsd loaded x30. pxe boot is ok thru tftp and dhcpd, but the x31 did freeze all the time. on freebsd 6.2 we had to remove all /tftpd/kernel/acpi* and it would grab the kernel and mfsroot and freeze at random after keyboard selection or alike. searching online showed a freebsd trouble with x31 related to either acpi and, very probably, the ata1 and the dock (docked x31 dont freeze). so we downloaded freebsd-7 to try it from the ftp, dumped its cd's boot/ to /tftpd but it would not even boot.. we tried to remove the acpi stuff the hard way (by removing the freebsd's kernel acpi*.ko files) but also by tweaking the load_acpi=NO and the hint.acpi.0.disabled=1 (since we pxeboot the laptop and have it load mfsroot+kernel at /tftpd/kernel/kernel we dont get the usual cdrom install menu that offers a non-acpi kernel). neither freebsd 6.2 nor the 7 we got from the ftp did work.. no fun.

17 Nov 2007

Sat, 17 Nov 2007 11:26:19 GMT

it's been a long time since i posted anything. well. left club internet. discovered the gsxr 750. still happy with netbsd and checking openbsd from time to time ;)

10 Oct 2006

Tue, 10 Oct 2006 17:26:15 GMT

today i tried a new firefox extension that lets you do whois from any web page. you just click it and it appears. so while i was on google.com home page i clicked the link and obviously something wrong appears : GOOGLE.COM.ZOMBIED.AND.HACKED.BY.WWW.WEB-HACK.COM GOOGLE.COM.VN GOOGLE.COM.UA GOOGLE.COM.SUCKS.FIND.CRACKZ.WITH.SEARCH.GULLI.COM GOOGLE.COM.PLZ.GIVE.A.PR8.TO.AUDIOTRACKER.NET GOOGLE.COM.MX GOOGLE.COM.IS.POWERED.BY.MIKLEFEDOROV.COM GOOGLE.COM.IS.NOT.HOSTED.BY.ACTIVEDOMAINDNS.NET GOOGLE.COM.IS.APPROVED.BY.NUMEA.COM GOOGLE.COM.HAS.LESS.FREE.PORN.IN.ITS.SEARCH.ENGINE.THAN.SECZY.COM GOOGLE.COM.BR GOOGLE.COM.AU GOOGLE.COM i sent an email to google (france) but they dont seem to react about it. if anyone knows anybody at google that could fix that whois record... :) its last update seems to be : Last update of whois database: Tue, 10 Oct 2006 13:23:10 EDT