Intel® Advisor User Guide

ID 766448
Date 7/13/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents
Give Feedback

Offload Modeling Report Overview

Review the controls available in the main report of the Offload Modeling perspective. You can view the interactive HTML report or a graphical user interface.

In the Accelerated Regions and Summary reports, you can drag and drop, close/open, collapse/expand panes to change the report layout.

Switch between perspectives or different Offload Modeling report tabs.

Review the summary of estimated offload characteristics for your application. The pane highlights total speedup, number of loops and functions offloaded, and a fraction of code accelerated.

Select code regions to show in the report based on offload type:

  • Show all code regions in your code.
  • Show only code regions recommended for offloading.
  • Show only code regions not recommended for offloading.

  • Click the button to show per-program recommendations, which you can collapse/expand.
  • Click the button to see the collection log including featured events separated by analyses, full collection log, and application output.
  • Create a snapshot for the current project results. For details, see Create a Read-only Result Snapshot.
  • Click a + button to open previously closed panes.

Review application performance measured on a host platform and its performance modeled on a target platform. For details about metrics reported, see Accelerator Metrics.

Depending on a perspective configuration and accuracy level, you might see different metrics reported and some metrics might be not accurate. Refer to the following topics for interpretation details:

Switch between Data Transfer Estimations and Details tabs for more information about loops selected in Code Regions:

  • Examine details about estimated data transfers in the loop and memory objects tracked and review data transfer recommendations. For details about how to read this pane, see Examine Data Transfers for Modeled Regions.

    NOTE:
    You need to enable light or full data transfer analysis before running the perspective to see metrics in this pane,

  • Examine offload details about the loop, such as loop type, measured and estimated execution time, estimated offload speed-up (for offloaded loops), and reason for not offloading, time for several bounded-by factors. If the loop is not recommended for offloading, this tab also reports the reason for it. For details, see Investigate Non-Offloaded Code Regions.

Switch between Source and Top-Down tabs for more information about loops selected in Code Regions:

  • Examine the source code and offload details for each source line. Select a loop in the Code Regions table to focus on the corresponding part of the source code.

  • View the loop/function hierarchy in a stack and its metrics.