Teraminutes for the meeting of 29th October 2005 Present: Chiara, Laurent, Mireille, Yannick, Emmanuel, Henry

Status of T0003 : Laurent

 See notes at the end of minutes for Laurent's description of the current status of T0003
 stack generation scripts cannot be run when the images are being graded
 mix4: 150 QFITS-in each day: this will be slower once we start the sextractor re-extraction processes at the same time
 All processing (QFITS and SExtractror-reprocessing) for stage one should be finished by 14th of November, if there are no machine crashes.

WIRCam : Chiara

 Astrometric solution computed with the new version of SCAMP is much better than before
 WIRCam stacks re-done with this new astrometric solution
 Galaxy counts generated with chi2 images
 A tarfile will be produced containing all this work, and also the QF-out information
 Chiara will work on automatic generation of masks next week
 Chiara will produce (J-K)_2mass vs. (J-K)_WIRCam plots

SCAMP: Emmanuel

 Scamp works better now, but there is still a problem for images where there are many stars (PI data Mireille is processing). According to Emmanuel this is likely a relative weighting problem as function of lacal number dentiy of objects.
 Henry needs to re-do the cosmos-u data with the new version of SCAMP
 The new version of SCAMP also can produce an output XML file which can be used to generate a web-page summary
 Emmanuel and Chiara work on the 2MASS astrometric+photomtery calibration problem: when Chiara does a photometric calibration of WIRCam images with 2MASS catalogues, it works while when she does photometric+astrometric calibration is does not.

Florence Durret, Tache de Service

 Florence Durret has approached TERAPIX to participate in the TERAPIX project as part of her Tache du Service. Yannick needs to find a suitable task. Any suggestions?

Status Hardware: Emmanuel, Henry, Yannick

 mix7 has not arrived yet
 Emmanuel thinks that the repeated problems with mix5 comes from i/o crashes (after all, it always happens during a swarp coaddition phase)
 Some bios / kernel options can be changed: SMP or APIC bios option are under study. Keep swarping large images to see what works; for the moment the SMP bios option has been changed and the SWarp (cosmos data) restarted. We should now by the end of the week end whether SMP can fix the problem.
 yannick and the new kiravix arrived this week. They are both dual-core AMD-64 machines. Emmauel's old machine kiravix will move to Chiara's office with the clarix nickname. Chiara's old machine statix will move to Henry's office for Anis, with the anix nickname. The old epix that is located on the Anis's desk will move to Chiara's office, for Mathias.

Next tera-meeting: Nov.

Bilan de la semaine

 Nous avons franchi le cap des 10000 QF hier soir
 Il en reste 2000, c'est bientot la fin.
 Le taux de QF est tombe a 1500 (crash mix4 + correction/modif spica)
 Encore 2500 images a noter au moins ;) (courage !)
 Etat des nodes : mix3 (raid1/2) : 630/110 (8j/2j) mix4 (raid1/2) : 220/430 (3j/6j) mix6 (raid1/2) : 430/250 (6j/4j)
 Donnees PI Sanders tjs en attente du CFH

Nouveautes :
 Le patch des QF vient de partir => ralentissement des QF
 Des tests de Stacks vont etre lance => ralentissement des QF
 Par consequent, la fin des QF initialement prevu vers le 6 novembre sera reporte de plusieurs jours (4-6j a la louche)
 La repartition des donnees sur le cluster a pour consequence qu'actuellement seul mix4 et mix6 feront des stacks ! (mix4 : Deep, mix6 le reste).

Prevision sur spica :
 Terminer la possibilite de transferer un stack d'une machine a une autre.