This document describes the Xilinx® Power Design Manger (PDM) usage for power estimation. Accurate power estimation during the early design cycle is the key to the lowest possible power envelope for any design. Early estimation is crucial for choosing the right device, taking advantage of architectural benefits, changing the design topology, and using different IP blocks. This helps in making better trade-offs well ahead of the design phase, letting you meet specifications and get your product to market faster.
Xilinx offers two types of power estimation tools—PDM and XPE which are typically used for estimation before design implementation and Vivado® Report Power which is used during design implementation with better accuracy. Both tools have many features that helps you create low-power Versal® ACAP designs. Xilinx recommends the following power methodology to overcome any power challenges throughout the design cycle.
During the conceptualization and architectural exploration phases of a project, it is crucial to assess the power budget with limited availability of architectural details of the design. PDM addresses most of the early power estimation challenges. It is typically used in the pre-design and pre-implementation phases of a project and it assists with architecture evaluation, device selection, appropriate power supply components, and thermal management solutions specific to your application. PDM considers your design resource usage, toggle rates, I/O loading, and other factors. These factors, combined with the device models, help calculate the estimated power distribution. The device models are extracted from measurements, simulation, and extrapolation.
Accuracy of the PDM tool is dependent on two primary sets of inputs. They are:
- Device usage, thermal environment and solution, clock, enable, toggle rates, and other information you enter into the tool
- Device data models integrated into PDM (power characterization accuracy)
For accurate power estimates of your application, enter the information that is as complete as possible. Modeling a certain aspect of the design too conservatively or without sufficient knowledge of the design can result in unrealistic estimates. Some techniques to drive PDM to provide worst-case estimates or typical estimates are discussed in this document.