DRIVING THE FUTURE: GPU VIRTUALISATION IN AUTOMOTIVE
The automotive industry is undergoing a profound transformation. As vehicles evolve into sophisticated, software-defined platforms, the demand for smarter, safer, and more efficient systems is accelerating. At the heart of this evolution lies a powerful enabler: GPU virtualisation.
An example of GPU multitasking in automotive using Imagination’s HyperLane virtualisation technology.
THE SHIFT TOWARD CENTRALISED VEHICLE ARCHITECTURES
Traditionally, vehicles relied on a distributed Electronic/Electrical (E/E) architecture, with each function such as infotainment, ADAS, or climate control, which is handled by a dedicated Electronic Control Unit (ECU). In modern vehicles, this has resulted in over 100 ECUs per car, increasing system complexity, weight, and cost.
To overcome these challenges, manufacturers are transitioning toward centralised E/E architectures, where fewer, higher-performance processors handle a broader range of tasks. This shift is enabled by virtualisation technologies, which allow multiple applications and operating systems to run concurrently (securely and efficiently) on shared hardware. In fact, by 2032, it’s estimated that nearly 30% of vehicles will adopt zonal E/E architectures, further accelerating the move toward centralised compute and GPU virtualisation.
WHY GPU VIRTUALISATION IN AUTOMOTIVE MATTERS
As vehicles adopt increasingly graphical and compute-intensive systems, ranging from high-resolution digital dashboards to AI-powered driver assistance, the GPU becomes a critical part of the vehicle’s computing infrastructure.
But not all workloads are equal. Safety-critical functions, such as the instrument cluster or rear-view camera, must be prioritised over non-critical applications like media playback. In legacy architectures, separating these by ECU made this easy. In centralised platforms, GPU virtualisation becomes essential.
Imagination’s HyperLane technology enables multiple virtual machines (VMs) to share a single GPU securely. Each VM is assigned a dedicated hardware lane, ensuring isolation, prioritisation, and near-native performance, with no need for software-based hypervisors.
REAL-WORLD BENEFITS OF GPU VIRTUALISATION IN AUTOMOTIVE
1. CONSOLIDATION WITHOUT COMPROMISE
OEMs and Tier-1 suppliers can now consolidate previously separate domains such as navigation, driver monitoring, ADAS, and infotainment onto a single SoC or ECU, without sacrificing performance or safety.
By embedding virtualisation at the hardware level, there’s no need to rely on third-party hypervisors. This reduces cost, integration complexity, and certification effort.
2. GUARANTEED SAFETY AND ISOLATION
Each VM has dedicated memory space via HyperLane. If one system fails (e.g. the entertainment UI), others continue functioning unaffected. This deterministic separation is critical for ISO 26262 and ASIL-B/D compliance.
3. HIGH-PERFORMANCE MULTITASKING
The demand for real-time compute in autonomous and driver-assist systems continues to grow. Hardware-based virtualisation eliminates the performance bottlenecks of software hypervisors, delivering low-latency, high-bandwidth access across all workloads.
MARKET MOMENTUM: A GROWING NEED FOR VIRTUALISATION
The automotive software market alone is expected to grow from $31.7 billion in 2024 to over $91.9 billion by 2032, driven by demands for smarter, connected vehicles. From advanced infotainment to autonomous functions, GPUs are no longer optional, they’re essential.
While traditional hypervisor solutions exist, many chipmakers now embed virtualisation directly into the processor IP.
Imagination is leading this shift with GPU IP designed from the ground up for automotive virtualisation.
IMAGINATION: BUILT FOR THE ROAD AHEAD
Our PowerVR GPUs are proven across automotive IVI and cockpit systems. Designed with safety, performance, and flexibility in mind, they offer:
- Up to 16 hardware HyperLanes per GPU
- Full isolation between VMs
- Support for safety-critical and consumer applications in parallel
- Real-time performance guarantees
- Advanced QoS and prioritisation mechanisms
Whether you’re building the dashboard of tomorrow or deploying edge AI for self-driving assistance, our GPU IP delivers the virtualisation, scalability, and safety modern automotive systems demand.
Explore the architecture, benefits, and implementation of GPU virtualisation in automotive systems in our in-depth white paper. Download the white paper on GPU Virtualisation.
Interested in learning more about our efforts in Automotive? Download our in-depth white paper on Solving the Mixed-Criticality Challenge in Automotive Controllers to explore how GPU virtualisation plays a vital role in enabling safe, efficient system consolidation.
