charm installation problem

From: Govardhan Reddy (greddy1_at_umd.edu)
Date: Sun May 27 2007 - 16:43:43 CDT

Hello,
I am trying to compile namd2.6 source code to make it work in
parallel on a cluster equipped with Xeon EM64T processors. I am
using intel compilers to compile the source code. The charm-5.9
installs without any error but when I run the test programs I get the
following error.
The test program "simplearrayhello" works fine but the other program
"pingpong" does not work properly. I get the following error when I
run the program:
************************************************************************
***************
Running on 2 processors.
Starting run at: Sun May 27 15:41:20 EDT 2007
-------------------
../../../bin/charmc pingpong.ci
Intra-processor Pingpong..
./charmrun ./pgm +p1

Running on 1 processors: ./pgm
Roundtrip time for 1D Arrays is 70.312500 us
Roundtrip time for 2D Arrays is 63.671875 us
Roundtrip time for 3D Arrays is 62.890625 us
Roundtrip time for Fancy Arrays is 61.718750 us
Roundtrip time for Chares (reuse msgs) is 60.546875 us
Roundtrip time for Chares (new/del msgs) is 62.890625 us
p1_18893: p4_error: interrupt SIGSEGV: 11
rm_l_1_18894: (3.828125) net_send: could not write to fd=5, errno = 32
p1_18893: (3.828125) net_send: could not write to fd=5, errno = 32
Inter-processor Pingpong..
./charmrun ./pgm +p2

Running on 2 processors: ./pgm
Roundtrip time for 1D Arrays is 70.703125 us
Roundtrip time for 2D Arrays is 63.671875 us
Roundtrip time for 3D Arrays is 63.281250 us
Roundtrip time for Fancy Arrays is 62.109375 us
Roundtrip time for Chares (reuse msgs) is 60.156250 us
Roundtrip time for Chares (new/del msgs) is 62.890625 us
p1_18908: p4_error: interrupt SIGSEGV: 11
rm_l_1_18909: (3.832031) net_send: could not write to fd=5, errno = 32
p1_18908: (3.832031) net_send: could not write to fd=5, errno = 32
Epilogue Initiated
Removing /scratch/470180.b08l02.oic.ornl.gov on node(s):
b01n052.oic.ornl.gov
Floaters flushed on node(s):
Epilogue Complete
   
************************************************************************
**********************
I tried to google the error but the results are not of much help. I
used the following commands to intall charm-5.9.

./build charm++ mpi-linux icc --no-shared -DCMK_OPTIMIZE=1 -I/opt/
mpich-ch_p4-icc-1.2.7/include -L/opt/mpich-ch_p4-icc-1.2.7/lib -
lmpich -pmpich -lpthread

I have edited the charm-5.9/src/arch/mpi-linux/conv-mach.sh file. The
contents of this are as follows:

CMK_CPP_CHARM='/lib/cpp -P'
CMK_CPP_C='/opt/mpich-ch_p4-icc-1.2.7/bin/mpicc -E'
CMK_CC='/opt/mpich-ch_p4-icc-1.2.7/bin/mpicc '
CMK_CXX='/opt/mpich-ch_p4-icc-1.2.7/bin/mpicxx '
CMK_CXXPP='/opt/mpich-ch_p4-icc-1.2.7/bin/mpicxx -E '
CMK_CF77='/opt/mpich-ch_p4-icc-1.2.7/bin/mpif77'
CMK_CF90='/opt/mpich-ch_p4-icc-1.2.7/bin/mpif90'
CMK_RANLIB='ranlib'
CMK_LIBS='-lckqt -lmpich '
CMK_LD_LIBRARY_PATH="-Wl,-rpath,$CHARMLIBSO/,/opt/mpich-ch_p4-
icc-1.2.7/lib"
CMK_NATIVE_LIBS=''
CMK_NATIVE_CC='/opt/intel/cc/9.1.045/bin/icc '
CMK_NATIVE_LD='/opt/intel/cc/9.1.045/bin/icc'
CMK_NATIVE_CXX='/opt/intel/cc/9.1.045/bin/icpc '
CMK_NATIVE_LDXX='/opt/intel/cc/9.1.045/bin/icpc'
CMK_NATIVE_CC='/opt/intel/cc/9.1.045/bin/icc '
CMK_NATIVE_CXX='/opt/intel/cc/9.1.045/bin/icpc '
CMK_F90LIBS='-L/usr/absoft/lib -L/opt/absoft/lib -lf90math -lfio -
lU77 -lf77math '
CMK_MOD_NAME_ALLCAPS=1
CMK_MOD_EXT="mod"
CMK_F90_USE_MODDIR=1
CMK_F90_MODINC="-p"

************************************************************************
*************
Can somebody please help me to resolve this problem
Thank You
Greddy

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:44:45 CST