-- SandraCastro - 2015-04-13

---

Use-Case 1: Run on a cluster, on interactive node

---

Description

Run CASA from a self-contained binary package. Run CASA in parallel using MPI.

Performance

Pre-Conditions

1. CASA binary package is pre-installed in the user account.

Basic Course

1. Login to interactive node.
2. Move to desired working directory.
3. Start CASA with mpi and chose the number of cores to use in the node.
   • Example: <path-to-binaries/>mpicasa -n <num_cores> <path-to-binaries/>casapy

Alternate Course

1. Login to interactive node. Use more than one interactive node.
2. Move to desired working directory.
3. Create hostfile with names of nodes and number of cores per node.
   • Example of hostfile:

     nmpost01 slots=16
     nmpost02 slots=15

4. Start CASA with mpi and give the name of the hostfile, as created above.
   • Example: <path-to-binaries/>mpicasa -hostfile myhostfile <path-to-binaries/>casapy

Post-Conditions

---

Use-Case 2: Run on a cluster, in different nodes, submit jobs with Torque

---

Description

Run CASA from a self-contained binary package. Run CASA in parallel using MPI and submit the jobs to remote nodes using Torque.

Performance

Pre-Conditions

Basic Course

Alternate Course

Post-Conditions

---

Use-Case 3: Run on a cluster, build CASA from source

---

Description

Run CASA on a cluster. CASA is built from source. This is the most typical use-case for CASA developers who would like to make changes in their local CASA and test-run the changes on the cluster.

Performance

Pre-Conditions

Basic Course

Alternate Course

Post-Conditions

---

Use-Case 4: Run on a user desktop

---

Description

Performance

Pre-Conditions

Basic Course

Alternate Course

Post-Conditions