AMBER Archive (2005)

Subject: AMBER: just one more try:) xleap, opteron, pgi and segmentation fault. (fwd)

From: Lars Packschies (packschies_at_rrz.uni-koeln.de)
Date: Fri Nov 04 2005 - 06:09:40 CST

Dear List,

  some days ago I posted one question. Does anybody know someone who can
possibly answer that one? Or maybe yome Opteron user knows how to compile
Amber correctly so that xleap works?

  Thanks in advance,

      Lars

---------- Begin Forwarded Message ----------
(original Post)

Dear Members,

  I have problems getting xleap to run. Amber8 (fully patched) is installed
on a dual opteron cluster in 64 bit using the PGI compilers 5.2-4 (serial
version). I applied the hints and tips on the PGI amber page
(<http://www.pgroup.com/resources/amber/amber8_pgi52.htm>). Everything
seems to run perfectly but xleap returns a segmentation fault.

  For any reason I cannot compile Amber (or leap...) in 32 bit mode. PGI
suggests to pactch the leap src Makefile (inserting X11LIBDIR after
"XHOME". I configured Amber with

  ./configure -nopar -opteron -acml -lapack pgf90

  and compiled with

  make -e YACC="/usr/bin/bison -y" X11LIBDIR="lib" serial

  In the end, it returns

/usr/bin/ld: skipping incompatible /usr/X11R6/lib/libXt.so when searching
for -lXt
/usr/bin/ld: skipping incompatible /usr/X11R6/lib/libXt.a when searching
for -lXt
/usr/bin/ld: cannot find -lXt
make[2]: *** [xaLeap] Error 2
make[2]: Leaving directory `/home/a0580/amber/amber8/src/leap/src/leap'
make[1]: *** [install] Error 2
make[1]: Leaving directory `/home/a0580/amber/amber8/src/leap'
make: *** [serial] Error 2

  libXt.so is there (pointing to libXt.so.6. The cluster runs the Rocks
distribution (version 3 something).

  Long story short: Is there a way to compile Amber8 fully in 32 bit mode?
  As far as I understand, only leap is tried to be linked to the 32 bit
(Xt-)library, while all the other code is compiled natively in 64 bit. Or
do I really have a problem with an incompatible 32 bit X library?

  However, I'b be happy just with xleap runnung, all other binaries seem to
be fine.

  Do you have any suggestions?

  Thanks in advance,

   Lars

---------- End Forwarded Message ---------- 

-- 
             ,H  + Dr. Lars Packschies  ZAIK/RRZK, Robert-Koch-Str.10 +
H      H    O    | Chemistry Dpt. Support            D-50931 Cologne  |
 `O-H. |  ,' `H  | Comp.Dpt., Univ. of Cologne  Phn (+49)221-478-7022 |
      `O-H       + Packschies_at_rrz.uni-koeln.de  Fax (+49)221-478-5568 +

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu