Rate This Document
Findability
Accuracy
Completeness
Readability

Comparison Analysis

For the same type of analysis tasks, you can select the same node or different nodes to compare the analysis results. In this way, you can quickly learn the differences between different analysis results, locate performance metric changes, and identify the effect of optimization methods.

  1. Click next to System Profiler.

    Choose Comparison Analysis from the drop-down list. The page for creating a task is displayed.

  2. Set the parameters as required. See Figure 1. Table 1 describes the parameters.
    Select an analysis scenario from the drop-down list, click Node-to-Node Comparison or Task-to-Task Comparison, and select the objects to be compared and analyzed.
    • Tasks in different scenarios cannot be compared and analyzed.
    • Tasks that do not comply with the current comparison mode cannot be selected.
    • If the components in specified tasks are different (for example, openGauss vs. MySQL in the database scenario), a dialog box will be prompted.
    • During the comparison of hotspot function analysis tasks, if the sampling ranges selected during task creation are different, the generated comparison report is useless.
    Figure 1 Creating a comparison analysis task
    Table 1 Parameters for creating a comparison analysis task

    Parameter

    Description

    Task Name

    Task name. The name must meet the following requirements:

    1. Contain only letters, digits, and underscores (_).
    2. Contain 1 to 64 characters.

    Select comparison object

    Select the object to be compared and click Hotspot Function Analysis to view the differential flame graph.

    When creating a comparison analysis task, you can click next to the analysis report to create it.

  3. Click OK.

    You can click next to the task name to delete all data related to the task. Exercise caution when performing this operation.

  4. Wait until the analysis completes.

    On the onCPU differential tab page (default), you can view the running details of the application.

    • The upper part aggregates the macro metrics of the current flame graph application, including the clock cycle, number of instructions, instructions per cycle (IPC), OS, and host name. The lower part lists the differential flame graph you select. There is a search module at the top of the flame graph. The searched functions are displayed in purple.
    • The differential flame graph can only indicate the differences between two flame graphs. Therefore, when you need to compare more than two flame graphs, select the base node and comparison node manually. Only one flame graph of the two selected nodes can be displayed at a time. You can click to switch between the benchmark object and the comparison object.
    • You can hover the mouse pointer over a function on the top of the flame graph and click the function to view its details. Only the last function can be displayed.
    • You can search for functions in the search box. You can confirm Case sensitive to initiate case-sensitive query. This option is not selected by default. After the query, the flame graph automatically adds a background color to the query result and a border to the first function. The search box supports regular expression match and fuzzy query.
    • During search, you can click or or press the up or down arrow key to switch the search result. You can press Enter to expand the selected function block, click + or - or scroll the mouse wheel to zoom in or zoom out the function block, and click to view how to use the flame graph.
    Figure 2 onCPU differential flame graph