[daip] gcc library 3.2

[Eric Greisen]

Sally Laurent-Muehleisen writes:

 > I'm running RedHat 8.0 on my system...unfortunately it's the latest
 > version of RedHat on which I'm assured of being able to run the Chandra
 > data analysis software I need.  I'm writing you to ask if, other than the
 > optimization issues you mention on your AIPS webpage for gcc/g77 version
 > 3.2.2 (and 3.2.3), if you know if the 3.2 libraries (included with RedHat
 > 8.0) will compile AIPS properly.  Also, how do I know for sure that the
 > routines mentioned in the OPTIMIZE.LIS file were compiled with the right
 > optimizations for my libraries?  I can't tell from the INSTEP2.LOG if the
 > optimization was used or not for the IOBSRC and PRJMAT routines (or for
 > the FILL, KWIK, and MASSGN routines for that matter).
 > 

We are using 3.2.3 for our main public release now and all seems okay.
The issue was the compiler not the libraries.  If IMAGR with DO3D and
DOCAL true runs correctly, then IOBSRC and PRJMAT were compiled
correctly.  Actually the INSTEP2.LOG should show that - compare the
fortran command line for these with adjacent lines.

FC        : as         LIST=FALSE PURGE=FALSE
FC        : plus       /usr/local/compilers/9/gcc_3.4.4/bin/g77 -c -fno-automatic -fno-globals -Wno-globals -malign-double -Wimplicit -Wall -O2 \
FC        :            /home/primate2/AIPS/31DEC05/LINUX/PREP/PRENAM.f

FC        : as         LIST=FALSE PURGE=FALSE
FC        : plus       /usr/local/compilers/9/gcc_3.4.4/bin/g77 -c -fno-automatic -fno-globals -Wno-globals -malign-double -Wimplicit -Wall  \
FC        :            /home/primate2/AIPS/31DEC05/LINUX/PREP/PRJMAT.f

Note the -O2 on PRENAM and not on PRJMAT.

You worry too much - if it works it is okay.

Eric Greisen