DS891 Datasheet, PDF (34/41 Pages) Xilinx, Inc – Zynq UltraScale+ MPSoC Overview

Zynq UltraScale+ MPSoC Overview
Configurable Logic Block
Every Configurable Logic Block (CLB) in the UltraScale architecture contains 8 LUTs and 16 flip-flops. The
LUTs can be configured as either one 6-input LUT with one output, or as two 5-input LUTs with separate
outputs but common inputs. Each LUT can optionally be registered in a flip-flop. In addition to the LUTs
and flip-flops, the CLB contains arithmetic carry logic and multiplexers to create wider logic functions.
Each CLB contains one slice. There are two types of slices: SLICEL and SLICEM. LUTs in the SLICEM can be
configured as 64-bit RAM, as 32-bit shift registers (SRL32), or as two SRL16s. CLBs in the UltraScale
architecture have increased routing and connectivity compared to CLBs in previous-generation Xilinx
devices. They also have additional control signals to enable superior register packing, resulting in overall
higher device utilization.
Interconnect
Various length vertical and horizontal routing resources in the UltraScale architecture that span 1, 2, 4, 5,
12, or 16 CLBs ensure that all signals can be transported from source to destination with ease, providing
support for the next generation of wide data buses to be routed across even the highest capacity devices
while simultaneously improving quality of results and software run time.
Block RAM
Every UltraScale architecture-based device contains a number of 36Kb block RAMs, each with two
completely independent ports that share only the stored data. Each block RAM can be configured as one
36Kb RAM or two independent 18Kb RAMs. Each memory access, read or write, is controlled by the clock.
Connections in every block RAM column enable signals to be cascaded between vertically adjacent block
RAMs, providing an easy method to create large, fast memory arrays, and FIFOs with greatly reduced
power consumption.
All inputs, data, address, clock enables, and write enables are registered. The input address is always
clocked (unless address latching is turned off), retaining data until the next operation. An optional output
data pipeline register allows higher clock rates at the cost of an extra cycle of latency. During a write
operation, the data output can reflect either the previously stored data or the newly written data, or it can
remain unchanged. Block RAM sites that remain unused in the user design are automatically powered
down to reduce total power consumption. There is an additional pin on every block RAM to control the
dynamic power gating feature.
DS891 (v1.3) September 23, 2016
www.xilinx.com
Advance Product Specification
34