User Guide

Intel® VTune™ Profiler User Guide

ID 766319
Date 3/31/2023
Public

A newer version of this document is available.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

Document Table of Contents

gpu-offload Command Line Analysis

Explore code execution on various CPU and GPU cores on your platform, correlate CPU and GPU activity, and identify whether your application is GPU or CPU bound.

Syntax

vtune -collect gpu-offload [-knob <knobName=knobValue>] -- <target> [target_options]

Knobs: collect-cpu-gpu-bandwidth, collect-programming-api, enable-stack-collection, enable-characterization-insights, target-gpu.

NOTE:

For the most current information on available knobs (configuration options) for the GPU Offload analysis, enter:

vtune -help collect gpu-offload

Example

This example runs GPU Offload analysis with enabled tracing for GPU programming APIs on the specified Linux* application:

vtune -collect gpu-offload -knob collect-programming-api=true -- /home/test/myApplication

What's Next

When the data collection is complete, do one of the following to view the result: