Part I: Vectorization – Introduction
February 23, 2022
Looking for ways to accelerate your compute-intensive applications, like image processing or sensor fusion in safety-critical systems?
For applications relying heavily on linear algebra, vector processing promises exciting performance gains.
In their next-generation of embedded, safety-certified MCUs, Infineon, amongst other vendors, will incorporate a so-called “vector unit”, an accelerator that can perform multiple identical operations on a stream of data in parallel. This concept, called single instruction, multiple data (SIMD), has been in use for decades in the desktop and server area and is now finding its way into safety-critical embedded systems. With this kind of hardware, applications that rely heavily on linear algebra, like image processing, sensor fusion or inference in AI systems, can be sped up by a factor larger than ten.
This kind of acceleration is not achievable by general-purpose multi-core processors because of the hardware overhead involved in managing multiple threads and their synchronization. In contrast, vector processors execute vector calculations as part of their normal program flow, with hardware managing synchronization. Therefore, there is no danger of deadlocks or race conditions, which alleviates the burden of qualifying your applications for safety critical systems.
Overall, vector processing is a welcome addition to software acceleration in embedded, safety-critical systems, that will enable powerful, real-time capable, yet energy efficient applications.
This post is the first part of a series on vectorization. In the upcoming weeks we will cover the peculiarities of the hardware and how to program these kinds of accelerators. We will also show you how emmtrix Parallel Studio can help you make use of the hardware efficiently and how your model-based workflow with Simulink can benefit from our vectorization solution.
Visit our website “vectorization” for more information and/or register for our webinar “Vectorization for Infineon AURIX™ TC4x” which we are scheduling around embedded world 2022.