Subject: New .TRK/.CAL files for runs 7295-8333 (28-MAR-1994 till 20-SEP-1994)
Ioanni.Katsavounidis@lngs.infn.it
Date: Mon May 15 2000 - 09:34:27 EDT
Ciao a tutti,
Today I've checked the production of TRK/CAL files for DDM17-DDM13.
Here is the summary:
(1) 147 runs crashed the code completely
(2) 5 runs crashed in DLWICO
(3) 4 runs crashed in ONLINE_TRACKER
(4) 3 runs crashed due to high optimization (OK with standard .EXE)
Note that the first 2 classes of problem runs should have no effect in
the re-calibration procedure, since all 152 runs involved are dated prior
to RUN007473 (considered as the "beginning of full Attico").
That leaves only the 3rd class (4 runs) that each had just a single problem
event which got skipped when launched a more "conservative" version of the
MAKETRK code.
In more detail:
- Runs 7295-7429 and 80859-80893 for DDM17 (till 22-APR-1994) had serious
problems that crashed execution immediately and thus no .TRK/.CAL files
were produced. Given that these runs correspond prior to the "official"
beginning of full Attico running, did nothing about it.
- Runs 7431,7439,7440,7441 and 7442 for DDM17 (22-25 Apr. 1994) had problems
on line 15893 in the DLWICO (streamer-tube tracking) routine. Relaunched
MAKETRK with strict error-handling that eventually produced .TRK/.CAL
files for the full run, even though it still skipped problem events
(in DLWICO)
- Runs 7579 (DDM16), 7849 (DDM15), 7894 (DDM15) and 8159 (DDM13) have had
problems with the ONLINE_TRACKER subroutine that crashed due to "illegal
floating point number". This is variable "SLOPED" in line 29268 of
subroutine ONLINE_TRACKER in DREAMLIB.FOR (DREAM V1.32). Apparently,
the contents of IBUF(19:20) that are equivalenced to SLOPED are wrong.
Again, relaunched MAKETRK with strict error-handling that eventually
produced .TRK/.CAL files for the full run, even though it still skipped
problem events (in ONLINE_TRACKER)
- Runs 7705 (DDM16), 7814 (DDM15) and 8048 (DDM14) have had problems with
the highly-optimized version of the MAKETRK code; problem went away when
re-ran it with the "standard" executable.
A presto,
Ioannis
This archive was generated by hypermail 2a24 : Mon May 15 2000 - 09:34:30 EDT