AMD Kintex UltraScale+ FPGAs: Power-Efficient And Fast

The FinFET node with the best price/performance/watt ratio.

AMD UltraScale+ vs. 7 Series FPGAs: Doubling Performance/Watt

This movie demonstrates the AMD 16 nm Kintex UltraScale+ devices’ dual-operating voltage’s performance, power, and versatility. The presentation shows the power and performance gains over AMD’s wildly popular 7 Series family.

AMD Kintex UltraScale+ Product Advantages

AMD Kintex UltraScale+ devices offer the best price/performance/watt balance in a FinFET node, making them an affordable option for applications requiring high-end features like 100G connectivity cores and 33 Gb/s transceivers. Applications like wireless MIMO technology, Nx100G wired networking, data centre networking, and storage acceleration are all well suited for our mid-range family, which is perfect for both packet processing and DSP-intensive tasks.

Best Price/Performance/Watt Balance on FinFET

Integration of Programmable Systems

• System logic cells of up to 1.2 million
• Integrated 100G Ethernet MAC with KR4 RS-FEC, PCIe Gen4, and 150G Interlaken cores for on-chip memory integration UltraRAM

Improved System Efficiency

• 6.3 DSP computing performance in TeraMACs
• More than twice as much system-level performance per watt as Kintex 7 FPGAs, mid-speed grade transceivers with 16G and 28G backplane capability, and 2666 Mb/s DDR4

BOM Cost Cuts

Fractional PLL integration and 12.5 Gb/s transceivers in the slowest speed grade lower the cost of clocking components.

Overall Power Savings

• 60% less power compared to 7 series FPGAs
• Performance and power options for voltage scaling
• Logic cells packed more tightly for dynamic power reduction

Enhanced Design Efficiency

• Vivado Design Suite was co-optimized for quick design closure.
• SmartConnect technology for applications involving intelligent IP integration.

Applications

Explore New Possibilities with the Kintex UltraScale+ FPGAs.

PON Access

One of the primary broadband access technologies used by numerous carriers is the Passive Optical Network (PON). In order to process the aggregated end-user traffic and provisioning in accordance with QoS regulations, packet processing and traffic management operations are essential. Kintex UltraScale+ FPGAs are well-suited to carry out Layer 2 to Layer 4 packet processing tasks as categorization, filtering, lookup, and packet forwarding because of their programmable logic and hardware acceleration blocks.

In addition to saving millions of LUTs for hardware differentiation, the integrated 100G Ethernet MAC facilitates effective uplink control. The Kintex UltraScale+ FPGA’s DSP resources, serial connectivity, memory, and logic performance, along with its power efficiency and optimal price point, make it appropriate for applications like high-volume 100G PON OLT line cards, which are always looking for a balance between performance, cost, and power.

Mobile Backhaul

In recent years, wireless traffic has increased dramatically. Applications for mobile backhaul and radio access networks (RAN) are becoming more innovative as a result of the growing demand for more data capacity. Multiple FPGAs are needed to accommodate the 112 MHz signal bandwidth supported by traditional microwave frequencies.

For the development of point-to-point microwave modems with increased packet processing in a single chip, the Kintex UltraScale+ FPGA is the perfect platform. In the dynamic wireless market, faster feature deployment and end-product differentiation are made possible by increased logic resources in comparison to the preceding generation. Additionally, designers can benefit from easy timing closure and power savings of up to 20% made possible by integrated IP.

Data Center Network Acceleration

New network functions are developing too quickly for fixed-function devices, and network port speeds are increasing quicker than conventional server nodes can keep up. With high-performance packet processing and datapath offloading, sophisticated SerDes technology, and 100G Ethernet IP to enable unparalleled scalability and connection for quick data transfer, Kintex UltraScale+ FPGAs provide the perfect answer to these problems.

Network designers may expedite a wide range of network tasks while lowering total cost of ownership thanks to Kintex UltraScale+ FPGAs’ robust performance and power balancing. Kintex UltraScale+ FPGAs offer deterministic, ultra low-latency, and flexible networking solutions with economical power consumption, making them an attractive option for the data centre whether speeding complex algorithms or introducing new unique functionality.

Get Started

Utilise the kit’s proven hardware, software support, tools, design examples, and documentation to accelerate your design cycle and achieve a quick time-to-market.

Kintex UltraScale+ KCU116 Evaluation Kit

For assessing important Kintex UltraScale+ features, particularly the 28 Gbps transceiver performance, the Kintex UltraScale+ FPGA KCU116 Evaluation Kit is perfect.

Overview of the Kintex UltraScale+ Product

Kintex UltraScale+ devices are co-optimized with the Vivado Design Suite and utilise the UltraFAST design technique to expedite time to market, based on the UltraScale architecture’s ASIC-class advantage.

Product Table