Attendance:
Warren, Joe, John, Peter, Stuart, Jolien, Duncan, Ben, Greg, Rauha,
Alan.
Issues with LDAS:
John's problem is to do with the v4/v5 frame library problem. One
uses degrees for azimuth angle, other uses radians. asci2frame is
still writing v4 frames and ldas works with v5 frames. The problem
is due to a bug in ascii2frame. It looks like there is confusion
between versions of the frame spec. Stuart will call Albert and try and
get this sorted out ASAP. Jolien has only seen a spec dated Jan.
Peter has one dated Feb. New version has degrees, old version has
radians.
Tania reported lat/long swapped. This is a problem with the DAQ
system. All the frames written so far are wrong. Conversion from deg
min sec to decimal degrees is incorrect as well. Peter will contact
framecpp authours and the sites to get this fixed.
Stuart will propose that both azimuth and altitude will be radians
and mail lal-discuss with result. And slap Albert for messing with
the frame spec.
LDAS at Lab:
Stuart still hopes for release tomorrow, maybe morning. Big issiue
is writing and reading proc frames. Think that the issues with
diskcache are sorted. Last diskcache bug in 0.1.105 has been sorted.
Will be a new version on ldas-test tomorrow to test proc frame
reading and last bugs before release.
LDAS at UWM:
is working well. no jobs seen stuck in the diskcache, no jobs lost
by manager, no missing nodes or search users. 5000 jobs run since
0.1.97 (bank sim, e7 analysis, eff dist plot)
Release of LAL/LALWrapper:
Jolien will get lal and lalwrapper released after the ldas release.
Other business:
Duncan asked if there may be a problem with multiple users
generating simulated frames with may have same ifo codes/timestamps.
Peter said best way around this is to force users to use the new
naming convention, where the type can be anything (not just R or F),
e.g. H-duncan-6000000000-16.gwf and construct -framequery { { H } {
duncan } { ... We should do this at uwm now we have multiple users.
|