Dear all,
I want to compile LAMMPS on Blue Gene/Q machine. I was wondering if there is any available Makefile that can help me to compile it. I know that there is one Blue Gene/L in MAKE directory of LAMMPS but it is not working for Blue Gene/Q.
Thank you,
search the web and ye will find enlightenment!
jeff hammond from argonne has posted detailed instructions for
compiling and using lammps on BG/Q here:
https://www.alcf.anl.gov/user-guides/lammps
Thanks for your reply.
I have tried this one, but it gives me the following error:
Thanks for your reply.
I have tried this one, but it gives me the following error:
----------------------------------------------------------------------------------------------------------------
"/bgsys/drivers/toolchain/V1R2M0-efix23/gnu-linux/lib/gcc/powerpc64-bgq-linux/4.4.6/../../../../powerpc64-bgq-linux/sys-include/sys/param.h",
line 74.9: 1540-0848 (S) The macro name "MIN" is already defined with a
different definition.
"pointers.h", line 35.9: 1540-0425 (I) "MIN" is defined on line 35 of
"pointers.h".------------------------------------------------------------------------------------------------------------------
The macro name MIN and MAX is defined in my system.
there need to be some define or setting to turn this off. IBM is known
to do many very stupid things in their system headers, but most of the
time there is a workaround. you either need to read through the docs
and figure this out or work with a person that knows the system well,
to sort this out. having unprotected commonly used macros like MIN/MAX
in a system head would be a blunder of epic proportions that is only
rivaled by defining "hz" as a preprocessor constant (which IBM
engineers managed to do on AIX).
in short, this is not a LAMMPS problem.
axel.
I did not see this error when I built LAMMPS-24Apr13 on BGQ a week or
two ago. I use "COMPILER = XLC" (this is defined in my Makefile.bgq,
which is posted on the website Axel mentioned) when building with
MEAM, REAX, REAXC and OMP packages. What changes did you make to my
makefile or additional options/packages did you use that caused this
error?
While I understand Axel's frustration with IBM and could share my own
stories, I believe that the idiotic definition of MIN/MAX can be
blamed on the GNU/Linux community, as indicated by the copyright noted
below.