Global Arrays parallelization of ridft and rdgrad

ridft and rdgrad are parallelized with MPI using the Global Arrayshttp://www.emsl.pnl.gov/docs/global/ toolkit.

Those versions are automatically used when setting PARA_ARCH=SMP but can also turned on with PARA_ARCH=GA on clusters. Due to the explicit usage of shared memory on an SMP system, user has to be allowed to use sufficient shared memory:

• In addition to the usual stack size limit problem, make sure that your maximum shared memory you are allowed to use on your system is large enough (a few GB or 70-90 percent of the total memory should be sufficient):
  • cat /proc/sys/kernel/shmmax shows the amount of allowed shared memory
  • use sysctl to set new values, but you have to be root to do that!
• the default shared memory that is used is per process for the matrices is 300 MB for heap and 10 MB for stack. For large cases this can be too small and an error will be given in the output. In order to increase the default values, just set:

  $paroptions
ga_memperproc <stacksize> <heapsize>

  stacksize and heapsize have to be given in word, i.e. units of 8 Byte (1 MB is equivalent to 131072 word).