AI Engine DevelopmentSee Vitis™ Development Environment on xilinx.com See Vitis-AI™ Development Environment on xilinx.com |
AI Engine Performance and Deadlock Analysis Tutorial¶
Version: Vitis 2021.2
Introduction¶
This tutorial introduces you to performance analysis and optimization methods, and shows you how synchronization works in graph execution. It also demonstrates the analysis of a hang issue using an example.
Before You Begin¶
IMPORTANT: First, set up the Vitis™ unified software platform and install the VCK190 base platform. Follow the instructions in the Vitis Software Platform Release Notes (v2021.2).
Before you begin this tutorial, run the following steps:
Set up your platform by running the
xilinx-versal-common-v2021.2/environment-setup-cortexa72-cortexa53-xilinx-linux
script as provided in the platform download. This script sets up theSYSROOT
andCXX
variables. If the script is not present, you must runxilinx-versal-common-v2021.2/sdk.sh
.Set up your ROOTFS to point to
xilinx-versal-common-v2021.2/rootfs.ext4
.Set up your IMAGE to point to
xilinx-versal-common-v2021.2/Image
.Set up your
PLATFORM_REPO_PATHS
environment variable based on where you downloaded the platform.
Objectives¶
After completing this tutorial, you will understand how to:
Read lock signals in Vitis Analyzer.
Profile design performance.
Identify a performance bottleneck.
Analyze a deadlock (hang issue) in the design.
Steps¶
Step 1: Learn about the execution model of an AI Engine graph, including how memory with lock and FIFO works. Learn performance measurement methods. Identify performance bottlenecks and practise optimization. Go to AI Engine Graph Execution and Measurement.
Step 2: Learn about how a deadlock might occur with an example design. Learn how to analyze a deadlock in the AI Engine simulator, hardware emulation, and hardware flows. Go to AI Engine Deadlock Analysis.
Tip: In this tutorial, the designs are self-contained in each step. You can choose to start at any step depending on your experience and requirements. Be aware that the concepts and options introduced in the previous step might not be repeated later. It is recommended to start from the beginning and progress through to completion.
Revision History¶
July 2021: Initial release.
Support¶
GitHub issues will be used for tracking requests and bugs. For questions go to forums.xilinx.com.
© Copyright 2021 Xilinx, Inc.
Licensed under the Apache License, Version 2.0 (the “License”); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.