Altera recently revealed details about its highly anticipated Agilex 3 AI FPGA, as well as introducing 11 development kits for its Agilex 5 family. The Agilex 3 AI FPGA family is set to offer a wide range of logic elements, from 25K to 135K, along with a dual Cortex A55 ARM hard processor subsystem. This subsystem features a programmable fabric with AI Tensor Blocks, all built on the advanced Intel 7 process.
One of the key advantages of the Agilex 3 AI FPGA is its ability to accelerate FPGA development for AI inference using popular frameworks like TensorFlow, PyTorch, and the OpenVINO toolkit. This streamlined development process extends to other FPGA development flows, including the Quartus environment, making it easier for developers to harness the power of FPGA technology for AI applications.
For applications in the intelligent edge space, the Agilex 3 FPGA enables real-time computing for time-sensitive tasks such as those found in autonomous vehicles and industrial Internet of Things (IoT) environments. In smart factory automation scenarios involving machine vision and robotics, Agilex 3 FPGAs facilitate the seamless integration of sensors, drivers, actuators, and machine learning algorithms.
The Agilex 3 FPGA family boasts impressive performance metrics, offering up to 2.8 peak INT8 TOPS and FPGA AI Suite support. This support enables a push-button flow from industry-standard frameworks to FPGA bitstream, with a fabric speed reaching up to 345MHz. These capabilities make the Agilex 3 FPGAs a compelling choice for developers seeking high-performance solutions for AI and edge computing applications.
Altera's CEO, Sandra Rivera, emphasized the company's commitment to leveraging programmable logic to drive innovation across various sectors, including data centers, aerospace, defense, communications infrastructure, automotive, industrial, medical, and embedded markets. The company's focus on delivering full-stack hardware and software solutions positions Altera as a key player in the field of high-performance accelerated computing systems and intelligent edge applications.