FPGA-Based Prototyping

Enabling early software integration and firmware development prior to ASIC or SoC availability

Overview

FPGA-based prototyping is the process of implementing/synthesizing an ASIC RTL on a platform comprising one or more FPGAs. Prototyping is done before tape out as part of the pre-silicon system validation flow but can be used post tape out for software development. The prototyping platform also includes interfaces to peripherals and memories that are used with the targeted ASIC.

Hardware verification and SW/FW development are the dominant factors in SoC design cost. Before tape out, hardware and software co-validation during prototyping allows developers to bring up the software and implement custom features before physical parts are available. Moreover, the design flow can be co-optimized by using the AMD Vivado™ Design Suite, which reduces cost and tape out risk, and improves efficiency and time-to-market.

For FPGA-based prototyping, AMD solutions:

  • Provide high-performance FPGAs for faster verification of targeted designs
  • Reduce board space requirements and complexity
  • Deliver flexible I/O to create a contiguous device
  • Deliver advanced debug, simulation acceleration, and interactive design tuning

With the Virtex™ 7 2000T FPGA and the Virtex UltraScale™ VU440 FPGA, AMD has been the market leader for the highest capacity FPGAs. The 16nm Virtex UltraScale+™ family now includes the world’s largest FPGA, the Virtex UltraScale+ VU19P FPGA, achieving three consecutive generations of high-end leadership.

Virtex 7 2000T

Built with SoC prototyping in mind

  • 2M logic cells, 6.8B transistors
  • 36x 12.5Gb/s serial transceivers
  • 46Mb of block RAM
  • 1,200 I/Os
  • 1st generation SSI technology

Virtex UltraScale VU440

Extending device density lead to 4X at 20nm

  • 5.5M system logic cells, 20B transistors
  • 48x 16.3Gb/s serial transceivers
  • 89Mb of block RAM
  • 1,456 I/Os
  • 2nd generation SSI technology

Virtex UltraScale+ VU19P

Industry’s highest capacity FPGA

  • 9M system logic cells, 35B transistors
  • 80x 28Gb/s serial transceivers
  • 94.5Mb of block RAM
  • 2,072 I/Os
  • 3rd generation SSI technology

 

Documentation
Design Example

Breakthrough performance and integration for ASIC prototyping and emulation can be realized with AMD UltraScale™ architecture. Virtex™ UltraScale devices simplify design partitioning through high logic capacity, over 90% device utilization, ASIC-like clocking, enhanced routing, and high-speed transceivers for pin multiplexing. This breakthrough architecture coupled with Vivado™ Design Suite provides the ideal solution for tackling the demands of leading-edge ASIC and SoC platforms.

Solution Summary and Benefits

  • Breakthrough device capacity reduces the number of partitions and simplifies board layout
  • Enhanced routing and co-optimization with the Vivado Design Suite ensures over 90% device utilization
  • ASIC-like clocking efficiently maps complex ASIC and SoC clock trees
  • High-speed transceivers enable efficient pin multiplexing between FPGAs and support the I/O interfacing requirements for next generation systems

UltraScale Architecture Benefits

  • Massive I/O Bandwidth
    • > 1 Tbps chip to chip bandwidth available
    • Low latency transceiver enables chip/chip interconnect
  • Massive Data Flow & Routing
    • Supports native wide ASIC busses with high performance
  • ASIC-like Clocking
    • Maximum flexibility for complex SoC prototyping
  • System Peformance
    • 15-30% performance improvement per device
    • 3X improvement due to reduced partitioning
  • Power Management
    • Up to 35% System Power reduction
Block Diagram