Intel® Hyperflex™ Architecture High-Performance Design Handbook

ID 683353
Date 10/04/2021

A newer version of this document is available.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

Document Table of Contents Loop Pipelining and Synthesis Optimization

The loop pipelining technique initially appears to create more logic to optimize this loop, resulting in less devices resources. While this technique may increase logic use in some cases, design synthesis further reduces logic through during optimization.

Synthesis optimizes the various clouds of logic. In the preceding example, synthesis ensure that the cloud of logic containing g*g*g*g is smaller than implementing four instances of block g. This reduction in size is because the LUT actually has six inputs, and logic collapses, sharing some LUTs. In addition, the Hyper-Retimer retimes registers in and around this smaller cloud of logic, thus making the logic less timing-critical.