4.3. Quantitative Timeline

The Quantitative Timeline (Views Menu->Charts->Quantitative Timeline) gives an overview of the parallel behavior of the application. It shows - over time - how many processes or threads are involved in which function. Along the time axis, the different functions are presented as vertically stacked color bars. The height of these bars is proportional to the number of processes that are currently within the respective function.

4.3.1. Mouse Hover

When the mouse hovers over the Quantitative Timeline, the status bar displays the present position of the cursor with respect to time. It also shows details pertaining to the function group under the cursor. Figure 4-14 shows that the mouse is currently at 2.01171 seconds and that it hovers over MPI.

4.3.2. Quantitative Timeline Settings

The Settings dialog box in the Quantitative Timeline has two tabs - the Preferences tab and the Functions tab.

Figure 4-14. Quantitative Timeline without a grid

4.3.3. The Context Menu

The context menu entry Ungroup in the Quantitative Timeline provides the option of ungrouping the given function group, like in the Function Profile (refer to Section Similarly, Regroup allows undoing a previous ungrouping.

The Hide option in the context menu conceals the chosen activity. To unhide all the hidden items, use the check boxes in the Functions tab of the Settings Dialog box.

Figure 4-17. Quantitative Timeline: context menu

The Move Group entry in the context menu changes the position of the different groups. The opted group can be moved to the top, to the bottom, upward by one position or downward by one position. Click on the legend to obtain a context menu with these options.

The other context menu entries are explained in Section 4.8.

4.3.4. Filtering and Tagging

Tagging in the Quantitative Timeline is indicated by a mesh pattern being placed over the tagged item. Figure 4-18 shows a Quantitative Timeline for the file poisson_icomm.single.stf with the tagged function MPI_Finalize on the right-hand side of the display.

Figure 4-18. Tagging the MPI_Finalize function in the Quantitative Timeline

Filtering in the Quantitative Timeline works the same way as in any other Chart. Figure 4-19 shows the result when only MPI_Finalize passes the filter.

Figure 4-19. Quantitative Timeline after filtering