[daip] Problem in 31DEC04 version of AIPS

Eric Greisen egreisen at nrao.edu
Fri Sep 24 10:33:54 EDT 2004


Poonam Chandra writes:
 > > The daip address reaches all of us and I have been following your
 > > discussion with Amy.  You keep referring to problems with adverbs but
 > > have never said what they are.  That is what I was trying to ask
 > > about.
 > 
 > My dear Eric,
 > 	Even I don't know what are the problems with adverbs!!!

   I did not ask you to diagnose a correction!  You keep saying there
is something wrong.  Why do you say that?  Surely, your remarks are
based on something - some message, error events or something else.
With some description of that I might have some idea.

 > 
 > >
 > > BUT - I think I know what your trouble with 31DEC04 is.  You used the
 > > gcc 3.3 for it and I will bet you used some other version of gcc for
 > > 31DEC03.  gcc 3.3 seems to have lots of problems with AIPS.  A failure
 > > in optimization can often cause the calibration routines to discard
 > > all the data - VPLOT and UVPLT use those routines, PRTUV and LISTR
 > > (SCAN anyway) do not.  Try LISTR in another OPCODE or UVPRT and I
 > > suspect you will see no data.
 > >
 > 
 > I did use gcc 3.3 when I installed 31DEC03 version. Infact, I had
 > installed gcc 3.3 at that time only and used the same version this time
 > while upgrading AIPS. But that time I had redhat 9 on my machine and now I
 > have Fedora core 2. Can that be a problem? Which gcc version shall I use
 > for installing AIPS? You are right. LISTR doesn't work with opcode 'MATX'.
 > I seem to be in soup!!

You are still using gcc 3.3.????? which was build on and for a RedHat
9 machine and you are now using the same binaries of the compiler on a
fedora core 2 system?  That alone might cause trouble.  You could
consider re-installing the compiler under your current OS.  But we
have had numerous reports of trouble under 3.3.3 and 3.3.4 - I use
3.2.3 with no problem - but under RedHat 9 (build for RH 9).

Eric Greisen




More information about the Daip mailing list