Configuring your Network of Workstations

Either the directory that you are running your model is replicated on each machine, or it is a distributed directory (NFS or some other distributed file system). The second option is better.

Create in your home directory(ies) a .rhosts file with the workstations you will use. Start xpvm or pvm and add these workstations to the pvm network (PVM command add <workstation>; XPVM: Hosts$\rightarrow$Other Host or read Help$\rightarrow$Hosts to see how to create and use the .xpvm_hosts configuration file). Now click on the Tangram's Parallel check box and then the Simulate button. If you are using xpvm, you will be able to see the interaction between the workstations.

NOTE: if you cannot add a machine to your PVM network, try running rsh $<$machine$>$ and see if it requires the password to login. PVM is not able to deal with passwords. The next step is to check, on each machine, if the $PVM_ROOT environment variable has your PVM's root path (TANGRAM II needs it also). Tips:

• If you keep xpvm open, you should reset your Trace and your View before running your simulation again.
• If xpvm hangs, which is quite common, just kill it and restart again, your simulation runs with pvmd and will not be stopped.

If the trace generation is enabled, each workstation will write a trace file on it is own.