[BLAST_ANAWARE] jobs on spud7

From: Chi Zhang (zhangchi@MIT.EDU)
Date: Sun Apr 25 2004 - 23:01:47 EDT


Hi,

I have put 4 lrn's on spud7 as blast for debugging Blastlib2. anounce here
lest they being killed.

during the last two months' running, there have been 41 runs where
crunch could not finish and the root file was left open. 20 of them are
marked both good runs in RUNLIST and VALID in elog.

I debugged the crash a little bit.

runs 6024 and 6049 crunched through without any change in code.
runs 6022, 6031, 6199 and 6235 indicate a small bug in TBLDst.cc. changes
have been made and these 4 runs are being crunched right now on spud7. As
a result, spud7 would probably be buzy the next 8 hours.

somehow I failed to build BlastLib2 from my account on spuds(build is OK
from bud10 though). complained about some syntax error in gcc lib and
include files. I was forced to debug from blast account.

Changes on code are made to a copy in cvs_v3/BlastLib2_dev and
cvs_v3/Blast_Params_dev to avoid directly changing codes used for online
crunching.

runs 5903 and 5994 are zipped on spud1 disc so are unreadable. another 12
runs before 45 degree spin angle was established are zipped too.

Chi



This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:30 EST