Pipelining Considerations - 2022.2 English

Versal ACAP Hardware, IP, and Platform Development Methodology Guide (UG1387)

Document ID
UG1387
Release Date
2022-11-16
Version
2022.2 English

Another way to increase performance is to restructure long datapaths with several levels of logic and distribute them over multiple clock cycles. This method allows for a faster clock cycle and increased data throughput at the expense of latency and pipeline overhead logic management.

Because devices contain many registers, the additional registers and overhead logic are usually not an issue. However, the datapath spans multiple cycles, and you must make special considerations for the rest of the design to account for the added path latency.