For all processors within a workstation cluster, the same release level of PE software is required. (This ensures that an individual PE application can run on any workstation in the cluster.)
When you use partitioning (available in PSSP Version 2 or later) on an IBM RS/6000 SP, you may have partitions at different levels of PE software. Within a partition, however, all the nodes must be at the same level of PE software. This ensures that an individual PE application can run on any node in the partition.
Table 3 lists the versions of PSSP and AIX required on a particular workstation cluster or partition. This list depends on the version of PE installed on that cluster or partition, and the possible migration paths.
Many of the compilers link to different libraries based on the AIX OSLEVEL value when they are installed. If you migrate just AIX, you will be using libraries for a back level. Be sure to change the compiler library links or reinstall compilers.
If you are operating in an environment where all the SP nodes are connected to the switch, when you install PE, the same level of the Message Passing Client Interface (MPCI) library will be installed on each SP node. MPCI, which is packaged with PSSP, is the primary interface to the point-to-point message passing protocols that support the SP Switch and SP Switch2.
All the nodes can participate in processing jobs. If, however, one node is later detached from the switch, that node can no longer participate in processing PE jobs and the MPCI libraries on that node are discarded. This is true for both the IP and User Space versions of MPCI. If you later reinstall or upgrade to a new version of PE, PE first checks to see if a version of the MPCI library already exists on the node. If one exists, it is not replaced. If one does not exist, a new version is placed on the node. In this case, the version of the (IP) MPCI library installed on the newly attached node may not match the version of the MPCI library previously installed on the other nodes. Thus the node will not be able to participate in running jobs because the levels of the MPCI library do not match. In this case, make the versions match by replacing or copying the necessary libraries.
How you plan your node resources will vary according to whether you are installing PE on an IBM RS/6000 SP or an IBM pSeries or RS/6000 network cluster.
The system administrator uses LoadLeveler to partition nodes into pools or features or both, to which he or she assigns names or numbers and other information. The workstation from which parallel jobs are started is called the home node and it can be any workstation on the LAN.
On an IBM pSeries or RS/6000 network cluster, you assign workstations to the following categories:
You need to identify these workstations running as execution nodes by name in a host list file.
An important aspect of planning your PE node resources is deciding which nodes will require which PE filesets or additional software. You do not need to install all of the PE filesets on every node. Refer to Software Requirements for more information on the filesets and their dependencies. This information will help you decide how to install PE and additional required software on your nodes.