mpi-rank

Specify which MPI process data you want to import.

Syntax

-mpi-rank <number>

Arguments

<number> is the rank of the process whose data will be imported.

Default

If an MPI rank is not specified, and there is more than one result directory in the project because the result partition is shared, then a rank is chosen at random. Specifying a rank is recommended.

Actions Modified

import-dir

Description

Use the mpi-rank action-option to specify which process's data you want to import for viewing. When collecting analysis data on a cluster, the data is stored in unique subdirectories under the project directory, named rank.#. When importing the data, mpi-rank is used to specify which processes' data should be imported. Only one processes' data can be imported and viewed at a time.

Example

This example imports MPI analysis data from the rank.3 cluster, reading source files from the specified search directory, and writing the result to the new-advi project directory.

$ advixe-cl -import-dir ./advi -mpi-rank 3 -project-dir ./new-advi -search-dir src:=./src