From owner-nwchem-users@emsl.pnl.gov Tue Jun 5 05:04:54 2007 Received: from odyssey.emsl.pnl.gov (localhost [127.0.0.1]) by odyssey.emsl.pnl.gov (8.13.8/8.13.8) with ESMTP id l55C4rs5012705 for ; Tue, 5 Jun 2007 05:04:54 -0700 (PDT) Received: (from majordom@localhost) by odyssey.emsl.pnl.gov (8.13.8/8.13.8/Submit) id l55C4rNh012704 for nwchem-users-outgoing-0915; Tue, 5 Jun 2007 05:04:53 -0700 (PDT) X-Authentication-Warning: odyssey.emsl.pnl.gov: majordom set sender to owner-nwchem-users@emsl.pnl.gov using -f X-Ironport-SG: No_SBRS X-Ironport-SBRS: None X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: Aq1HAEDuZEbBkCEm/2dsb2JhbAAVghs3iXmOSQ X-IronPort-AV: E=Sophos;i="4.16,385,1175497200"; d="scan'208,217";a="34015071" Message-ID: <4665523E.6090106@cesga.es> Date: Tue, 05 Jun 2007 14:08:30 +0200 From: =?ISO-8859-1?Q?Jose_Carlos_Mouri=F1o_Gallego?= User-Agent: Thunderbird 1.5.0.7 (X11/20060909) MIME-Version: 1.0 To: Dunyou Wang CC: nwchem-users@emsl.pnl.gov Subject: Re: [NWCHEM] nwchem 5.0 execution problems References: <4663FB62.8030803@cesga.es> <466434FB.3010006@pnl.gov> In-Reply-To: <466434FB.3010006@pnl.gov> Content-Type: multipart/alternative; boundary="------------070608000908070502060504" X-Virus-Scanned: by amavisd-new Sender: owner-nwchem-users@emsl.pnl.gov Precedence: bulk This is a multi-part message in MIME format. --------------070608000908070502060504 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Hi, Thank you for your response. I'll put the options I have used, and the versions of compilers, librarys, and OS. OPTIONS and environment variables: ------------------------------------------- export LARGE_FILES=TRUE export NWCHEM_TOP=/home/cesga/jmourino/Aplicaciones/nwchem-5.0 export NWCHEM_TARGET=LINUX64 export BLASOPT="-L/home/cesga/jmourino/intel/mkl/9.0/lib/em64t -lmkl_solver -lmkl_lapack -lmkl_em64t -lmkl_lapack64 -lmkl -lvml -lguide -lpthread" gmake nwchem_config NWCHEM_MODULES=all cd $NWCHEM_TOP/src gmake clean gmake 64_to_32 # I've tried both, using the icc compiler and using the gcc. And the error is the same gmake USE_64TO32=y HAS_BLAS=yes FC=ifort gmake FC=ifort CC=icc USE_64TO32=y HAS_BLAS=yes COMPILER versions ------------------------ jmourino@svgd ~/Aplicaciones/nwchem-5.0> ifort --version ifort (IFORT) 9.1 20061101 Copyright (C) 1985-2006 Intel Corporation. All rights reserved. jmourino@svgd ~/Aplicaciones/nwchem-5.0> icc --version icc (ICC) 9.1 20061101 Copyright (C) 1985-2006 Intel Corporation. All rights reserved. jmourino@svgd ~/Aplicaciones/nwchem-5.0> gcc --version gcc (GCC) 3.4.4 20050721 (Red Hat 3.4.4-2) Copyright (C) 2004 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO MKL version --------------- Intel(R) Math Kernel Library 9.0 for Linux* OS --- jmourino@svgd ~/intel/mkl/9.0/doc> cat /etc/redhat-release Red Hat Enterprise Linux ES release 4 (Nahant) jmourino@svgd ~/> uname -a Linux ************* 2.6.9-22.0.2.ELsmp #1 SMP Tue Jan 17 06:12:06 CST 2006 x86_64 x86_64 x86_64 GNU/Linux I hope all this infomation cauld help. Carlos Dunyou Wang escribió: > Hi Carlos, > > On your PC cluster (RH/LINUX x86_64), I wonder what MKL libs you linked > with. On a 64 bit platform, the em64t libs need to be linked with. > Here are the libs options you need to link with on your x86_64 LINUX64 > platform: > > setenv BLASOPT '-L$(MLK_LIB_PATH)/em64t -lmkl_lapack -lmkl_em64t > -lguide -lpthread' > > then: > > make FC=ifort USE_64TO32=y HAS_BLAS=y link > > Hope this helps > Dunyou > > > Jose Carlos Mouriño Gallego wrote: > >> Hi all, I'm new in the list. >> >> I'm a applications technician in a Supercomputer Center in Galicia, >> Spain. >> We have several HPC systems and we wonder to install nwchem 5.0 on >> several of them. >> >> My first try was in a HP TRU 64 with DECOSF1, linking with native math >> library cxml (with the 64to32 correction). >> After the first compilation, some tests fails, but, downgrading some >> compilation options, the tests are ok now. >> >> Then, I tried to compile it in a Superdome, HPUX system with IA64 >> processors and MLIB as native math library. >> (HP-UX sd1 B.11.23 U ia64 4294967295 unlimited-user license) >> The compilation was OK, but almost all the tests fails. The error in the >> tests is not comparing the output, the reason is a BUS error in the >> executable. (one test is OK, tests/autosym/autosym) >> >> 0:Bus error, status=: 10 >> 0:Bus error, status=: 10 >> 0: ARMCI aborting 10 (0xa). >> 0: ARMCI aborting 10 (0xa). >> >> After this, I tried to compile it in a PC cluster. The nodes have RH >> Operating system and Pentium IV EM64T processors (Linux svgd >> 2.6.9-22.0.2.ELsmp #1 SMP Tue Jan 17 06:12:06 CST 2006 x86_64 x86_64 >> x86_64 GNU/Linux) >> I've used the Intel Fortran compiler, the gcc C compiler and the Intel >> MKL math library (making the 64to32 correction) >> >> The reslut is that only 4 or 5 tests ends properly, the others fails >> executing the program. >> >> 0:Segmentation Violation error, status=: 11 >> 0:Segmentation Violation error, status=: 11 >> Last System Error Message from Task 0:: Inappropriate ioctl for device >> 0: ARMCI aborting 11 (0xb). >> 0: ARMCI aborting 11 (0xb). >> system error message: Inappropriate ioctl for device >> >> >> Any ideas? >> >> Carlos >> >> >> > > > -- ______________________________________________________________________ J. Carlos Mouriño Gallego Tel.: +34 981 56 98 10 Tecnico de Aplicaciones. Fax: +34 981 59 46 16 CESGA e-mail: jmourino@cesga.es Avda. de Vigo s/n. Campus Sur. 15705 - Santiago de Compostela. SPAIN ______________________________________________________________________ --------------070608000908070502060504 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Hi, Thank you for your response.

I'll put the options I have used, and the versions of compilers, librarys, and OS.

OPTIONS and environment variables:
-------------------------------------------

export LARGE_FILES=TRUE
export NWCHEM_TOP=/home/cesga/jmourino/Aplicaciones/nwchem-5.0
export NWCHEM_TARGET=LINUX64

export BLASOPT="-L/home/cesga/jmourino/intel/mkl/9.0/lib/em64t -lmkl_solver -lmkl_lapack -lmkl_em64t -lmkl_lapack64 -lmkl -lvml -lguide  -lpthread"

gmake nwchem_config NWCHEM_MODULES=all

cd $NWCHEM_TOP/src
gmake clean
gmake 64_to_32
# I've tried both, using the icc compiler and using the gcc. And the error is the same
gmake USE_64TO32=y HAS_BLAS=yes FC=ifort
gmake FC=ifort CC=icc USE_64TO32=y HAS_BLAS=yes


COMPILER versions
------------------------

jmourino@svgd ~/Aplicaciones/nwchem-5.0> ifort --version
ifort (IFORT) 9.1 20061101
Copyright (C) 1985-2006 Intel Corporation.  All rights reserved.

jmourino@svgd ~/Aplicaciones/nwchem-5.0> icc --version
icc (ICC) 9.1 20061101
Copyright (C) 1985-2006 Intel Corporation.  All rights reserved.

jmourino@svgd ~/Aplicaciones/nwchem-5.0> gcc --version
gcc (GCC) 3.4.4 20050721 (Red Hat 3.4.4-2)
Copyright (C) 2004 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO


MKL version
---------------

Intel(R) Math Kernel Library 9.0 for Linux*

OS
---

jmourino@svgd ~/intel/mkl/9.0/doc> cat /etc/redhat-release
Red Hat Enterprise Linux ES release 4 (Nahant)

jmourino@svgd ~/> uname -a
Linux ************* 2.6.9-22.0.2.ELsmp #1 SMP Tue Jan 17 06:12:06 CST 2006 x86_64 x86_64 x86_64 GNU/Linux

I hope all this infomation cauld help.

Carlos






Dunyou Wang escribió: 
Hi Carlos,

On your PC cluster (RH/LINUX x86_64), I wonder what MKL libs you  linked
with.  On a 64 bit platform, the em64t libs need to be linked with. 
Here are the libs options you need to link with on your x86_64  LINUX64
platform:

setenv BLASOPT  '-L$(MLK_LIB_PATH)/em64t -lmkl_lapack -lmkl_em64t
-lguide -lpthread'

then:

make FC=ifort  USE_64TO32=y HAS_BLAS=y link

Hope this helps
Dunyou


Jose Carlos Mouriño Gallego wrote:
  
Hi all, I'm new in the list.

I'm a applications technician in a Supercomputer Center in Galicia,
Spain.
We have several HPC systems and we wonder to install nwchem 5.0 on
several of them.

My first try was in a HP TRU 64 with DECOSF1, linking with native math
library cxml (with the 64to32 correction).
After the first compilation, some tests fails, but, downgrading some
compilation options, the tests are ok now.

Then, I tried to compile it in a Superdome, HPUX system with IA64
processors and MLIB as native math library.
(HP-UX sd1 B.11.23 U ia64 4294967295 unlimited-user license)
The compilation was OK, but almost all the tests fails. The error in the
tests is not comparing the output, the reason is a BUS error in the
executable. (one test is OK, tests/autosym/autosym)

0:Bus error, status=: 10
0:Bus error, status=: 10
 0: ARMCI aborting 10 (0xa).
 0: ARMCI aborting 10 (0xa).

After this, I tried to compile it in a PC cluster. The nodes have RH
Operating system and Pentium IV EM64T processors (Linux svgd
2.6.9-22.0.2.ELsmp #1 SMP Tue Jan 17 06:12:06 CST 2006 x86_64 x86_64
x86_64 GNU/Linux)
I've used the Intel Fortran compiler, the gcc C compiler and the Intel
MKL math library (making the 64to32 correction)

The reslut is that only 4 or 5 tests ends properly, the others fails
executing the program.

0:Segmentation Violation error, status=: 11
0:Segmentation Violation error, status=: 11
Last System Error Message from Task 0:: Inappropriate ioctl for device
 0: ARMCI aborting 11 (0xb).
 0: ARMCI aborting 11 (0xb).
system error message: Inappropriate ioctl for device


Any ideas?

Carlos


    


  


-- 
______________________________________________________________________
J. Carlos Mouriño Gallego               Tel.: +34 981 56 98 10
Tecnico de Aplicaciones.                Fax:  +34 981 59 46 16
CESGA                                   e-mail: jmourino@cesga.es
Avda. de Vigo s/n. Campus Sur.
15705 - Santiago de Compostela. SPAIN
______________________________________________________________________
--------------070608000908070502060504--