A Software Stack for Next-Generation Automotive Systems on Many-Core Heterogeneous Platforms

Abstract

The advent of commercial-of-the-shelf (COTS) heterogeneous\ud many-core platforms is opening up a series of opportunities\ud in the embedded computing market. Integrating multiple\ud computing elements running at smaller frequencies allows obtaining\ud impressive performance capabilities at a reduced power\ud consumption. These platforms can be successfully adopted to\ud build the next-generation of self-driving vehicles, where Advanced\ud Driver Assistance Systems (ADAS) need to process unprecedently\ud higher computing workloads at low power budgets. Unfortunately,\ud the current methodologies for providing real-time guarantees\ud are uneffective when applied to the complex architectures\ud of modern many-cores. Having impressive average performances\ud with no guaranteed bounds on the response times of the critical\ud computing activities is of little if no use to these applications.\ud Project HERCULES will provide the required technological\ud infrastructure to obtain an order-of-magnitude improvement in\ud the cost and power consumption of next generation automotive\ud systems. This paper presents the integrated software framework\ud of the project, which allows achieving predictable performance\ud on top of cutting-edge heterogeneous COTS platforms. The\ud proposed software stack will let both real-time and non real-time\ud application coexist on next-generation, power-efficient embedded\ud platform, with preserved timing guarantees

A Perspective on Safety and Real-Time Issues for GPU Accelerated ADAS

“…We argue that methodologies for GPU scheduling and memory interference arbitration in SoCs will play a key role for the automotive industry players: these are not only important from a real-time perspective, but it is also a necessity dictated by the requirements of both temporal and spatial isolation as recommended by ISO-26262. We also highlighted how methodologies for error detection and redundancy for GPUs are still inspired by what has been presented in CPU-related literature: moreover, we argue that even if such mechanisms are well established for HPC oriented devices, these very same approaches can also be extremely useful in embedded automotive scenarios [36].…”

Contention-Aware GPU Partitioning and Task-to-Partition Allocation for Real-Time Workloads

“…Waves tend to be executed simultaneously within all the SMs. A sub-optimal configuration of number of blocks, therefore can cause the last wave to be served by a small number of SMs, hence this last wave causes to increase the execution time of the whole kernel while under-utilizing the GPU 3 .…”

“…Such architectures are usually based on multi-core CPUs coupled with massively parallel accelerators. While the adoption of such heterogeneous platforms satisfies the computational needs for ADAS-related workloads, engineering difficulties arise when the system is subject to real-time and safety requirements [3,10,18].…”

Graphic Interfaces in ADAS

“…Pre-processing of data involves aggregating nearby road users positions and reprojecting them through a homography to the respective locations, hence translating from GPS coordinates to the local coordinate system used in our 3D rendering. As far as scheduling is concerned, we are able to bound both period, CPU and GPU compute time for our 3D graphic application: for testing and measurements we used the NVIDIA Jetson TX2 4 which is a well-known development board featuring a six-core CPU and a high performance integrated GPU fully compliant the latest OpenGL and OpenGLES profiles. Such development platform allowed us to measure framerates (e.g.…”

“…ISO 26262) imposed by certification authorities, since the co-existence of critical components from the ADAS (Advanced Driving Assisted System) stack, and non-critical, HMI and infotainment systems, is only possible through a carefully design of both computing hardware and software ecosystem. Researchers acknowledge that software artifacts such as Real-Time Operating Systems and hypervisors are key technologies to solve this problem [4], as they can provide the amenable property of isolation among components. In this work, we analyze two real usecases under development in the EU project Prystine [10] and we highlight how our implementation efforts balanced regulations and requirements for both an ergonomics and performance perspective.…”