First off, I'm aware of the changes to libtool, beta versions of LAM, etc.
It's not so much that I can't compile LAM. (Well, maybe it is; but I'm
chipping at each new occurance as it comes... and the fix is the same in
every case so far...)
I've read most of the posts about compiling LAM with the PGI compilers. I
also have found an interesting solution that I haven't found in the
mailing list archives, and while it works for me (for now), I'd like to
get some comments on... well I think what I've done is too easy, and I
suspect there's more to it than what I've done.
After looking through the ./configure script(s), I found repetitions of
the following (or something similar to it):
***
linux*)
case $CC in
icc* | ecc*)
lt_prog_compiler_wl_F77='-Wl,'
lt_prog_compiler_pic_F77='-KPIC'
lt_prog_compiler_static_F77='-static'
;;
ccc*)
lt_prog_compiler_wl_F77='-Wl,'
# All Alpha code is PIC.
lt_prog_compiler_static_F77='-non_shared'
;;
esac
;;
***
Hmm...
So, for each occurance, I've added a pgcc case...
**
pgcc*)
lt_prog_compiler_wl_F77='-Wl,'
lt_prog_compiler_pic_F77='-fPIC'
lt_prog_compiler_static_F77='-static'
;;
**
Obviously, there are minor differences from one instance to another, but
by and large, they are identical.
After adding a pgcc case, everything seems to just work (from a
compilation standpoint).
As I said previously -- I'm not sure if this is the best way to go about
this; espescially knowing very little about the internals of
autoconf/automake/whatever produces the 'configure' script. But editing
the ./configure script (and then re-configuring and re-compiling) seems to
do the trick.
--
Troy Telford
Linux Networx
ttelford_at_[hidden]
(801) 649-1356
|
