Premium SoC with next generation C7™ NPU for advanced AI in ADAS and software defined vehicles

The TDA5 high-performance compute SoC family is

designed to deliver safe and efficient edge AI

performance, with capabilities of up to 1200 TOPS,

with integrated C7™ NPU and chiplet-ready

architecture. This enables seamless progression to

L3 autonomous driving in software-defined

vehicles, where conditional automation is possible

in both urban and highway environments.

Additionally, the TDA5 SoCs are well-suited for

similar applications in industrial transportation,

humanoid and industrial robots, and aerospace and

defense with their advanced sensor processing,

cybersecurity, and functional safety features.

The TDA5 SoCs feature multiple specialized

subsystems, each tailored to address the growing

demand for high-performance compute and

cross-domain applications. These subsystems

include dedicated processing cores and hardware

acceleration for security, vision processing, edge

AI, display rendering, and networking. By

offloading these tasks, the SoCs’ MPU and MCU

cores are freed up to focus on user application

software. Furthermore, support for PCIe, Ethernet,

and other automotive standard peripherals enables

safe, secure, and high-speed data transfer between

different components and systems.

Building on two decades of leadership in the

automotive processor market, the TDA5 architecture

is designed to provide a scalable and

high-performance solution for a wide range of

applications. As an evolution of the TDA4 family,

software developed for TDA4 can be easily scaled

to the TDA5 family, with minimal rework required.

This allows for the reuse of software assets and

enables the development of more complex and

sophisticated applications. The TDA5 family’s

unique combination of high-performance real-time

and analytics cores, along with the latest C7™ NPU

and next-generation accelerators for image signal

processing, makes it an optimized solution for

various camera, radar, sensor fusion, and AI

applications.

The TDA5 family provides high-performance compute

capabilities for both traditional computer vision

and deep learning algorithms, with

industry-leading power/performance ratios. The

high level of system integration enables lower

costs for advanced automotive platforms,

supporting multiple sensor modalities in

centralized ECUs, multiple sensor domains, or a

centralized automotive computer. Key processing

and acceleration cores include the

latest-generation C7™ DSP with scalar and vector

cores, dedicated deep learning acceleration,

latest application and GPU cores for general

compute, a vision and imaging subsystem, video

codec, network packet processing accelerator, and

Ethernet switch, as well as a dedicated security

subsystem. These cores are carefully integrated

into an SoC architecture designed from the ground

up to support functional safety at a system level,

ensuring the highest level of reliability and

performance in safety-critical applications.