Visualizer Window

The Tools > Visualize and Tools > Visualize Distribution commands, and the $visualize intrinsic send an array’s values to the Visualizer so that it can display these values. A “view across” display can also be visualized, in which case, the process or thread index forms one axis of the display.

The Launch Strings Page within the File > Preferences Window dialog box allows you to specify the command TotalView will use to start the Visualizer.

The Maximum array rank field sets the maximum rank of the array that TotalView will export to the Visualizer. The Visualizer cannot visualize arrays of rank greater than two; if you are using another visualizer, however, or if you are dumping binary data, you can set the limit here.

The Visualizer can be used in two ways: it can be launched by TotalView to visualize data as you debug your programs, and it can be run from the command line to visualize data dumped to a file in a previous debugger session.

Visualizing your program’s data uses two interactions:

You interact with TotalView to choose what you want to visualize and when it should make snapshots of your data.

You interact with the visualizer to choose how you would like your data to be ­displayed.

The TotalView handles the first of these interactions, extracting data and marshalling it into a standard format that it sends down a pipe. The Visualizer then reads the data from this pipe and displays it for analysis.

The Visualizer has two types of windows:  

Dataset Window

A single main window lists the datasets that you can visualize. You can use this window to set global options and to create views of your datasets.

View Window

A data window contain images of the datasets. By interacting with a View Window, you can change its appearance and set dataset viewing options. Using the Dataset Window, you can open several View Windows on a single dataset to get different views of the same data.

For more information on the Visualizer, see the TotalView Users Guide.