Accelerate Your Design with Cutting-Edge Software Tools

Giles Peckham, a regional marketing director at Xilinx, and Adam Taylor, an embedded systems consultant, discuss how to speed up vehicle electronics design using a system-optimizing compiler

Vehicle electronics have come a long way over the years. They’ve evolved from just basic engine systems and an AM radio to contemporary cars equipped with advanced technology like engine control, ADAS (Advanced Driver-Assistance Systems), traction and stability control, infotainment systems, and even autonomous driving features.

This rapid technological advancement brings several challenges that designers need to tackle:

1. Performance: Many vehicle features require real-time, low-latency, and highly predictable performance. This includes ADAS, ECU (Engine Control Unit), and traction and stability control systems.
2. Security: Car electronics perform critical functions where failures can lead to injury or death. Therefore, it’s vital to implement robust security measures to prevent unauthorized modifications.
3. Safety: Systems must comply with automotive safety standards as defined by ISO 26262.
4. Interfacing: The systems need to connect with various sensors, drives, and other actuators.
5. Power Efficiency: These systems must work efficiently within fixed power constraints.
6. Software Flexibility: They should adapt to different standards and conditions across various markets.

To address these issues, developers are utilizing heterogeneous System on Chip (SoC) devices. These SoCs combine a multi-core processing unit with other specialized co-processors like GPUs, DSPs, or programmable logic.

By combining a processing unit with programmable logic, a tightly integrated system is created. This allows the programmable logic’s parallel processing abilities to handle high-performance tasks and interfacing, while the processing unit manages high-level decision-making, communications, and system management. This collaboration results in a system that’s more responsive, predictable, and power-efficient, as it can offload tasks to programmable logic.

For interfacing, heterogeneous SoCs support a variety of industry-standard interfaces via either the processing system or programmable logic. They’re adaptable enough to handle legacy or custom interfaces using programmable logic, though this might require an external PHY for the physical layer of the protocol to ensure wide-ranging connectivity.

Security-wise, some heterogeneous SoCs offer easily implementable device-level and system-level security features, such as encrypting and authenticating the boot and configuration process. If the SoC uses Arm processors, TrustZone technology can secure the software environment by creating separate, isolated worlds with restricted hardware access, facilitated by a hypervisor. Additional design options, like functional isolation, can further enhance security based on specific requirements.

Traditionally, developing a heterogeneous SoC meant dividing the design between the processor system and the programmable logic, requiring two separate development teams. This method increased costs, extended development times, and raised technical risks. It also locked certain design functions to either the processor cores or the programmable logic, making subsequent optimizations challenging.

By using a system-optimizing compiler, these complexities and inefficiencies can be reduced, helping streamline the design process.