Message 7/446 Huib Jan van Langevelde Oct 11, 2000 09:42:57 am +0200 Date: Wed, 11 Oct 2000 09:42:57 +0200 (MET DST) X-Sender: huib@juw21 To: Exploder for EVN Techies cc: "Walker; Craig" Subject: EVNtech: Sched report for the TOG Sender: owner-evntech@jb.man.ac.uk Precedence: bulk Below my view on Sched development relevant for the EVN in the last 6 months. Groeten, Huib ------------------------------------------------------------------------------ Huib Jan van Langevelde, Operations Manager, EVN MkIV Correlator at JIVE langevelde@jive.nl Web : http://www.jive.nl/ Joint Institute for VLBI in Europe Phone : +31-521-596515 Radiosterrenwacht Dwingeloo or : +31-521-596500 Postbus 2 Fax : +31-521-597332 7990 AA DWINGELOO Home : +31-528-221273 the Netherlands when traveling, Mobile: +31-621201419 STATUS OF SCHED WITH REGARD TO EVN/MKIV, TOG TORUN OCT 200 H.J.van Langevelde October 2000 My efforts in this area have 3 main aspects: Maintenance and development of the VEX related Sched code, maintenance of the EVN entries in the Sched catalogues, involvement in the VEX definition process (also relevant for the correlator). At the moment all Sched related efforts are under pressure, as they are not my main area of responsibility. Code maintenance can probably not easily be transfered, for the other areas we are considering different people at JIVE. Below I report on development in all 3 areas. VEX --- I am responsible for issuing an iteration of VEX definition 1.6. Amongst other things it has a much cleaner definition of tape/scan times. No progress has been made in the last 6 months. CATALOGUES ---------- Some changes are due in the catalogues. The last released version of Sched was issued 28 April 2000, which had all the VLBA4 stuff. Note that users currently produce VLBA4 schedules for Nt, Sh, Yb, but not for any of the other VLBA systems (Ef,Cm,Mh,Tr). Stations are encouraged to review entries regularly. A nasty problem was to fill the catalogs with reasonable S/X entries for geodetic stations (see below). SCHED/VEX CODE -------------- One experiment (GG41) was affected by the fact that Sched does not warn that automatic tape allocation can not be used for experiments that are to be correlated somewhere different than the VLBA processor. As a result GG041 had to be redirected to Socorro. Several users required a patch for the September session to allow them to schedule experiments with oversampling for the JIVE processor. The S/X session in May 2000 used several geodetic stations. Their standard setup uses a different IF to VC to formatter connection. Sched currently has no knowledge about this patching scheme, and therefore produces no reasonable defaults for these systems, nor can it check that the schedule prepared will work properly. Support for such systems will require quite some effort by both me and Craig Walker. There is currently no estimate when this capability could be available. In order to make 1Gb/s an operational feature, Sched needs to be upgraded for 2 head recording as well as 320 ips/16Mb/s/tr tape speeds. It seems the 2 head recording is the first step. Initial work on this capability has been carried out by me after a very intensive visit by Craig Walker. It is relatively straightforward, as the VLBA has an equivalent concept in the form of 2 drive recordings. Two head recording is practically finished and ready for testing with the FS. Allowing for quadruple speed recordings is probably slightly more involved, as there is no equivalent concept in Sched yet. Not only the VEX part, but really all areas of Sched would need changes to accommodate it. I estimate it will take several man weeks for us to make these enhancements. Recently it was discovered that the barrel-roll description Sched was using had been wrong. This is not affecting any experiments yet, but should be fixed soon.