Getting Started With Smart Placement

Table of Contents

• Introduction
  • When is smart placement required:
  • What not to expect from Smart Placement?
• Specific implementation of smart placement
  • Manual Implementation
  • SW Instrumentation implementation

Introduction

Usually used when application foot print is more than internal memory size (3.5MB for case of AM263Px). More on this can be read at Smart Placement

When is smart placement required:

Smart Placement would:

1. Place critical code section needed for deterministic latencies to appropriate memories
2. Identify and place code section which are bottleneck to system performance and place them in faster memory.
3. Improves XIP performance.

What not to expect from Smart Placement?

Smart placement is dependent on profiling data and the input given by the user on the priority of the function. The result achieved by smart placement tool is dependent on these factors. In absence of this factor, the full effect of smart placement would not be achieved.

Specific implementation of smart placement

Manual Implementation

Use this when

1. it is not suitable to run the application on the target and collect runtime profiling data.
2. Developer of the application have knowledge about the criticality of various functions/Tasks.

More on this can be read at Manual Smart Placement

SW Instrumentation implementation

This method gives an automated way of finding critical functions. A more frequent function is more critical function. Use this when:

1. Application runtime profiling and profiling data extraction is possible.
2. Higher profiling coverage is required.

More on this can be read at Software Instrument Smart Placement