Intel® Advisor User Guide

User Guide

Contents

Pane: GPU Roofline Chart

Use GPU Roofline chart to visualize actual performance of your GPU kernels against hardware-imposed performance ceilings. For more information about investigating GPU Roofline results, see Examine Bottlenecks on GPU Roofline Chart .
Families of Intel® X
e
 graphics products starting with Intel® Arc™ Alchemist (formerly DG2) and newer generations feature GPU architecture terminology that shifts from legacy terms. For more information on the terminology changes and to understand their mapping with legacy content, see GPU Architecture Terminology for Intel® X
e
 Graphics.

GPU Roofline Chart Controls

There are several controls to help you focus on the GPU Roofline chart data most important to you, including the following.
1
  • Select by Mouse Rect
    : Select one or more kernels by tracing a rectangle with your mouse.
  • Zoom by Mouse Rect
    : Zoom in and out by tracing a rectangle with your mouse. You can also zoom in and out using your mouse wheel.
  • Move View by Mouse
    : Move the chart left, right, up, and down.
  • Undo
     or
    Redo
    : Undo or redo the previous zoom action.
2
Use the filter drop-down to choose, which functions/loops to display on a Roofline chart. The following controls are available:
  • Use
    Operations
     pane to filter kernels by type of operations: INT or FLOAT.
  • Use the
    Memory Level
     pane to show results for each kernel in the chart: CARM, L3, SLM, GTI.
3
Use the
Compare
 drop-down to plot results from another Roofline chart on results of your current project.
Click the button to add results for comparison.
View and switch between the files that are currently compared in the
Compared Results
 pane.
After comparison, the recent results are saved. You can view the list of recent results in the
Ready for Comparison
 pane.
4
Add visual
Guidance
 to the GPU Roofline chart to make the interpretation of data easier, including performance limits and whether kernels are memory bound, compute bound, or both.
In the
Guidance
 drop-down toolbar, use the
Display roof rulers
 checkbox to enable showing a vertical line from a kernel to the nearest and topmost performance ceilings. To view the ruler, hover the cursor over a kernel dot. Where the line intersects with each roof, labels display hardware performance limits for the kernel.
The preview picture is updated as you select guidance options, allowing you to see how changes will affect the GPU Roofline chart's appearance. Click
Apply
 to apply your changes or
Default
 to return the GPU Roofline chart to its original appearance.
5
  • Roofline View Settings
    : Change the default scale setting to show:
    • The
      optimal
       scale (default), which adjusts to a chosen GPU Roofline chart view.
    • A
      constant
       scale, which adjusts to the tallest or widest view and does not change when a different GPU Roofline chart view is chosen.
  • Roof Settings
    : Change the visibility and appearance of roofline representations (lines):
    • Click a
      Visible
       checkbox to show/hide a roof line.
    • Click a
      Selected
       checkbox to change a roof line appearance: display the roof line as a solid or a dashed line.
    • Manually fine-tune roof values in the
      Value
       column to set hardware limits specific to your code.
  • Loop Weight Representation
    : Change the appearance of dots:
    • Point Weight Calculation
      : Change the
      Base Value
       for a point weight calculation.
      For a GPU Roofline chart, only
      Self Elapsed Time
       is available as a base value.
    • Point Weight Ranges
      : Change the
      Size
      ,
      Color
      , and weight
      Range
       of a dot. Click the
      +
       button to split a point weight range in two. Click the
      -
       button to merge a point weight range with the range below.
    • Point Colorization
      : Color dots by
      weight ranges
       or by
      type
       (vectorized or scalar). You can also change the color of loop with no self time.
6
  • Hover your mouse over a dot to display metrics and, if enabled, a roof ruler for it.
  • By default,
    Intel Advisor
     generates a roofline for
    GTI (Memory)
    , which reports memory traffic, in bytes, generated by all execution units.
    Double-click a dot or select a dot and press
    SPACE
     or
    ENTER
     to display labeled dots representing memory levels for the selected kernel. Lines connect the dots to indicate that they correspond to the selected kernel.
    If you have chosen to display only
    some
     memory levels in the chart using the
    Memory Level
     toolbar, unselected memory levels are displayed with
    X
     marks.
    To hide the labeled dots, do one of the following:
    • Select another kernel.
    • Double-click an empty space in the GPU Roofline chart.
    • Press
      SPACE
       or
      ENTER
      .
  • Click the
    +
     button next to a dot on a chart to break it into smaller dots representing groups of instances of the same source kernel. Instances differ by global and local size.
    • Hover over each instance to view its performance metrics.
    • Select a dot representing an instance to highlight it in the
      GPU
       pane and view detailed information about its performance and memory usage in the
      GPU Details
       tab.
    • Double-click a dot representing an instance to view how it utilizes each memory level.
  • Right-click a kernel dot or a blank area in the Roofline chart to perform more functions, such as:
    • Further simplify the GPU Roofline chart by filtering out (temporarily hiding a dot), filtering in (temporarily hiding all other dots), and clearing filters (showing all originally displayed dots).
    • Show/hide a side panel that displays metrics for a selected dot.
    • Add visual guidance to the GPU Roofline chart to make the interpretation of data easier. These options are the same as in the
      Guidance
       toolbar.

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.