[CPMD-list] Intel compiler 10 seg fault resolution on quadcore Xeon.

[Joseph Hargitai]

It appears that using version 10 of Intel, on quadcore 2.33Ghz machines the -xT flag causes the seg-fault. The interconnect is Infinband. 

Using Atlas from the CPMD site and the following conf file works. (it also works with 
  CC='/usr/mpi/intel/mvapich-0.9.9/bin/mpicc'. 

Optimization level -O2 also works, but there is no increase at all in speed. This is stripped to minimal optimization:


IRAT=2
     CFLAGS=''
     CPP='/lib/cpp -P -C -traditional'
     CPPFLAGS='-D__Linux -D__PGI -DFFT_DEFAULT -DPOINTER8 -DLINUX_IFC \
        -DPARALLEL'
     FFLAGS='-pc64 -unroll'
     LFLAGS=' -L. -latlas_x86_64'
     FFLAGS_GROMOS='-Dgood_luck $(FFLAGS)'
       FC='/usr/mpi/intel/mvapich-0.9.9/bin/mpif77 -c'
       CC='/usr/mpi/gcc/mvapich-0.9.9/bin/mpicc'
       LD='/usr/mpi/intel/mvapich-0.9.9/bin/mpif77'


as far as the results for various node/cpu combinations: (8 cores per node) 


5 nodes, 8 cpus, for 40 cpu total 

  ELAPSED TIME :    0 HOURS 20 MINUTES 51.31 SECONDS     
 

reverse:

8 nodes 5 cpus, for 40 cpu total

 CPU TIME :    0 HOURS 14 MINUTES 26.52 SECONDS     
   


8 nodes 6 cpus  for 48 cpu total

     CPU TIME :    0 HOURS 10 MINUTES  5.86 SECONDS     
  


6 nodes 8 cpus - 48 cpu total

   CPU TIME :    0 HOURS 13 MINUTES 40.81 SECONDS     
   ELAPSED TIME :    0 HOURS 14 MINUTES 18.46 SECONDS     



6 nodes  7 cpus -  42 cpu total  
 CPU TIME :    0 HOURS 14 MINUTES 31.68 SECONDS     
 

1 node 8 cpus - 8 cpu total -  

 CPU TIME :    1 HOURS 31 MINUTES 55.62 SECONDS     
  

2 nodes 8 cpus - 16 cpu total

       CPU TIME :    0 HOURS 44 MINUTES 42.87 SECONDS    
  
j