>From: Jeff Squyres <jsquyres_at_[hidden]>
>......
>So this should not have happened. :-) I can think of 4 reasons offhand
>that you might be seeing this behavior:
>
>1. When you re-configured LAM and re-built it, you did not "make clean"
>first. This, therefore, may not have re-compiled the wrapper compilers
>proprly. As a quick test, you can do the following to see if this is what
>happened:
>
> cd /directory/where/you/have/lam/source/code/lam-6.5.9
> cd tools/hf77
> ./hf77 -showme
> # should show "f77..."
> make clean all
> ./hf77 -showme
> # show now show "g77..."
>
>If this is what happened, then do a top-level "make clean all install",
>and you should be ok.
Yes, that was what happened. I did "make clean all install", and now I am
able to compile and execute those little tester like xtc_CsameF77 and
xfmpi_sane. So, now I move on to the "real" BLACS tester - that in the
BLACS/TESTING directory. I have difficulty there but I will open a new
thread to discuss that: See Problems compiling BLACS tester - argument
types not matched.
>4. You have the environment variable LAMH77 set to "f77", which overrides
>the underlying compiler that mpif77 uses at run time.
Where can I find all these environ variables?
Finally, to summarize the resolution of the initial problem that I
encountered, for the purpose of archiving in case the interim messages were
removed during the archiving, the problem compiling/executing a couple of
Fortran77-related testers were resolved by installing LAM 6.5.9, using
compilers from the same vedor, reinstalling using "make clean all install",
changing the compiler flags to be consistent with the underlying C and F77
compilers, and all that.
And Jeff Squyres is the nicest guy in the world!!
_________________________________________________________________
Tired of spam? Get advanced junk mail protection with MSN 8.
http://join.msn.com/?page=features/junkmail
|
