Intel® Trace Analyzer and Collector

Summary Page

Summary page provides you with summary information on the time your application spent in MPI.

To access the Summary page, go to View > Summary Page, or click the toolbar button.

Advanced Aggregation

The advanced aggregation function is based on process aggregation and allows lightening the Event Timeline chart for big traces above 1K processes. By default, the function is automatically enabled for traces with above 768 processes, but you can change this threshold using the Edit Configuration Dialog Box.

Advanced Aggregation Settings in the Edit Configuration Dialog:

MEM-FLUSHBLOCKS

Syntax

MEM-FLUSHBLOCKS <number of blocks>

Variable

VT_MEM_FLUSHBLOCKS

Default

1024

Description

This option controls when a background thread flushes trace data into the flush file without blocking the application. It has no effect if AUTOFLUSH is disabled. Setting this option to a negative value also disables the background flushing.

Zoom Stack

The zoom stack supports navigation in time by storing previously displayed time intervals. You can restore these time intervals when it is necessary. Keyboard or mouse navigation is quite intuitive without detailed knowledge of the zoom stack. Nevertheless, an explanation is given below for the sake of complete reference.

Consider a trace file spanning the time from 0 to 100 seconds. When the first View is opened the zoom stack looks like this:

Defining and Recording Scopes

Scope is a user-defined region in the source code. In contrast to regions and functions, which are entered and left with VT_begin/VT_end() or VT_enter/VT_leave(), scope does not follow the stack based approach. It is possible to start scope a, then start scope b and stop a before b, that is they can overlap one another:

Subscribe to Intel® Trace Analyzer and Collector