Defining Input Delays - 2020.2 English

UltraFast Design Methodology Guide for Xilinx FPGAs and SoCs

Document ID
UG949
Release Date
2021-02-18
Version
2020.2 English

The input delay is defined relative to a clock at the interface of the device. Unless set_clock_latency has been specified on the source pin of the reference clock, the input delay corresponds to the absolute time from the launch edge, through the clock trace, the external device and the data trace. If clock latency has already been specified separately, you can ignore the clock trace delay.

Figure 1. Input Delay Computation

The input delay values for the both types of analysis are:

Input Delay(max) = Tco(max) + Ddata(max) + Dclock_to_ExtDev(max) - Dclock_to_FPGA(min)
Input Delay(min) = Tco(min) + Ddata(min) + Dclock_to_ExtDev(min) - Dclock_to_FPGA(max)

The following figure shows a simple example of input delay constraints for both setup (max) and hold (min) analysis, assuming the sysClk clock has already been defined on the CLK port:

set_input_delay -max -clock sysClk 5.4 [get_ports DIN]
set_input_delay -min -clock sysClk 2.1 [get_ports DIN]
Figure 2. Interpreting Min and Max Input Delays

A negative input delay means that the data arrives at the interface of the device before the launch clock edge.