[Wrf-users] Re: About compilation of wrf v2.2

[Brent Shaw]

I have been able to successfully compile WRF v2.2 using PGI on a Linux
cluster using the default settings (with -fastsse).  Likewise, the
solve_em.f90 compilation took a long time, but did eventually complete.
During the time it was compiling, the memory usage on my compile node
reached 3.5GB, but the machine has 4GB, so maybe that is why our machine
was able to finish.  I have compiled both the serial and the DM-Parallel
(RSL-Lite with nesting but no periodic LBCs) version with success.  For
what it's worth, our cluster is based on AMD Opteron servers (x86_64),
running PGI version 6.2-5 (I think this is the latest).  The MPI part is
Myrinet, using MPICH-GM wrapped around PGI. 

Best regards,
Brent
------------------------------

Message: 2
Date: Thu, 4 Jan 2007 00:11:47 +0800 (CST)
From: eddycarl at 126.com
Subject: [Wrf-users] About compilation of wrf v2.2
To: " wrf-users at ucar.edu " <wrf-users at ucar.edu>
Message-ID: <459BD5C3.00002F.01930 at bj126app65.126.com>
Content-Type: text/plain; charset="gb2312"

Dear all,

Anyone can compile wrf model v2.2 which is released in Dec. 2006 ? I am
doing the compilation on a pc-cluster with PGI complier. However,the
compilation always does not complete! I checked the compile log and It
seems that when compiling a program "solve_em.f90",it took a very long
time, but simply could not finish it. PGI compiler version does not
count,as I have changedthe version to the latest one, but things go on
like before ... Does anyone can help with this problem? Thank you!

Regards,
Carl