[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20011207: new Unidata-Wisconsin machine shipped (cont.)

>From: Dee Wade <address@hidden>
>Organization: SSEC
>Keywords: 200111071940.fA7JeI100958 Unidata-Wisconsin DataCenter ADDE

Dee,

>Thanks Tom.
>Warning.....our TC people are swamped and your boxes may sit here for awhile
>before they get around to them.  I have 2 brand new computers that have been
>waiting for their attention.
>I will let you know when they arrive though

OK.  In case it helps, this machine has been "locked down" just about
as hard as it can be (security, that is) and has been configured with
the IP addresses (two for the RAID units) you provided earlier.  Even
though we realize that your TC people will be looking hard at the
machine before installation there, we are confident that the list of
chores that they will need to do is very short.  If it helps to get
things moving, we can provide a list of what was done here and why it
was done.

On a totally different note, I loaded the LDM and my latest McIDAS
release on this box and was running XCD for a number of days before it
was shipped.  The new thing about this is that my distribution of
-X/-XCD was entirely built using gcc/g77.  This combination of
compilers helped me find some previously undiscovered bugs in XCD code
which I have fixed (I will be updating the CVS repository at SSEC in
the next couple of days with the latest fixes).  I also modified
'makefile' and 'mccomp.sh' for -X to allow one to simply specify use of
the gcc/g77 combination (specify '-g77' as the VENDOR macro value
instead of '-gcc' (which causes the gcc/mcfc combo to be used) or
'-vendor' (which causes the vendor compilers to be used).  I am pretty
happy that I have caught all of the bugs in McIDAS code (the final
tally was 65 modules modified) that has prevented using g77 for so
long, but, like anything else, time will tell if I got _all_ of them.

Tom