Efinix Doubles Titanium FPGA Family To Address Growing AI Demands
The newly available devices, manufactured by TSMC, feature high transceiver speeds and low power consumption for use in AI-powered machine vision and industrial systems.
Efinix, a programmable logic solutions provider, has announced a significant expansion to its Titanium FPGA family, doubling the number of available products. The new devices are designed to address the growing demands of AI-driven systems and other high-performance, edge-based applications by offering increased capabilities, including high-speed transceivers and enhanced processing cores.
Efinix’s Titanium FPGA family is now made up of 20 available products. Image used courtesy of Efinix
Titanium FPGA Family Expansion
The Titanium FPGA family expansion is focused on delivering a mix of energy efficiency, high performance, and a small form factor, all important characteristics for modern edge computing deployments. The Taiwan Semiconductor Manufacturing Company (TSMC) manufactures the FPGAs on its 16/12 nm process.
The FPGAs are engineered for applications that require on-the-fly data analysis and real-time decision-making. Efinix notes that these FPGAs are in production for various markets, with a particular focus on AI-powered machine vision in industrial systems.
New capabilities in the expanded Titanium family include high-speed transceivers operating up to 25.8 Gbps, a key feature for high-bandwidth data transfer. The latest devices also integrate a 64-bit RISC-V processor SoC, providing more robust processing power on-chip. Additionally, the MIPI interfaces have been enhanced to support higher speeds, which is beneficial for systems that rely on camera and sensor data, such as those used in robotics and smart devices.
Efinix’s Quantum compute fabric features a 16 nm process node to enable low power consumption in a small form factor. Image used courtesy of Efinix
Balancing Power, Performance, and Size
The architecture of the Titanium FPGAs, which Efinix refers to as the Quantum compute fabric, is designed to balance power, performance, and area efficiency. This balance is particularly important in applications where space and energy consumption are limited. The company’s integrated development environment, Efinity, is a complete design suite that supports the full design flow from RTL to bitstream, enabling engineers to prototype and deploy solutions quickly.
With the new additions, the Titanium family now includes a wider range of logic elements, with devices ranging from 35K to 2M logic elements. The availability of these new products in various packages, including down to a minuscule 3.5 mm x 3.4 mm WLCSP, allows for integration into highly compact systems. Efinix also highlights the inclusion of features like bitstream authentication and encryption for enhanced security, as well as single-event upset detection to ensure system integrity in harsh environments.
Efinix’s Efinity software enables users to build designs for the Titanium FPGAs. Image used courtesy of Efinix
FPGAs Engineered for AI-Powered Systems
Efinix’s expansion of the Titanium product line provides engineers with more options for designing systems that require flexible, high-performance, and low-power processing at the edge. The new devices are well-suited for a wide range of applications, including industrial automation, robotics, medical devices, and automotive systems. The integration of high-speed transceivers and a RISC-V processor on-chip streamlines system design by reducing the need for external components, simplifying PCBs, and accelerating time-to-market for complex AI-driven applications.
