In the automotive industry, there is currently great interest in supporting driver-assist and autonomouscontrol features that utilize vision-based sensing through cameras. The usage of graphics processing units (GPUs) can potentially enable such features to be supported in a cost-effective way, within an acceptable size, weight, and power envelope. OpenVX is an emerging standard for supporting computer vision workloads. OpenVX uses a graph-based software architecture designed to enable efficient computation on heterogeneous platforms, including those that use accelerators like GPUs. Unfortunately, in settings where real-time constraints exist, the usage of OpenVX poses certain challenges. For example, pipelining is difficult to support and processing graphs may have cycles. In this paper, graph transformation techniques are presented that enable these issues to be circumvented. Additionally, a case-study evaluation is presented involving an OpenVX implementation in which these techniques are applied. This OpenVX implementation runs atop a previously developed GPU-management framework called GPUSync. In this case study, the usage of GPUSync's GPU management techniques along with the proposed graph transformations enabled computer vision workloads specified using OpenVX to be supported in a predictable way.

Dataflow software architectures are prevalent in prototypes of advanced automotive systems, for both driver-assisted and autonomous driving. Safety constraints of these systems necessitate real-time performance guarantees. Automotive prototypes often ensure such constraints through over-provisioning and dedicated hardware; however, a commercially viable system must utilize as few low-cost multicore processors as possible to meet size, weight, and power constraints. In short, these platforms must do more with less. To this end, we develop cache-aware and overhead-cognizant scheduling techniques that lessen guaranteed response times without unnecessarily constraining platform utilization. We implement these techniques in PGMRT, a portable middleware framework for managing real-time dataflow applications on multicore platforms. The efficacy of our techniques is demonstrated through overhead-aware schedulability experiments and runtime observations. Results for our test platform show that cache-aware clustered scheduling outperforms naïve partitioned and global approaches in terms of schedulability and end-to-end response times of dataflows.

Motivated by computational capacity and power efficiency, techniques for integrating graphics processing units (GPUs) into real-time systems have become an active area of research. While much of this work has focused on single-GPU systems, multiple GPUs may be used for further benefits. Similar to CPUs in multiprocessor systems, GPUs in multi-GPU systems may be managed using partitioned, clustered, or global methods, independent of CPU organization. This gives rise to many combinations of CPU/GPU organizational methods that, when combined with additional GPU management options, results in thousands of "reasonable" configuration choices. In this paper, we explore real-time schedulability of several categories of configurations for multiprocessor, multi-GPU systems that are possible under GPUSync, a recently proposed highly configurable real-time GPU management framework. Our analysis includes the careful consideration of GPU-related overheads. We show system configuration strongly affects real time schedulability. We also identify which configurations offer the best schedulability in order to guide the implementation of GPU-based real-time systems and future research.

This paper describes GPUSync, which is a framework for managing graphics processing units (GPUs) in multi-GPU multicore real-time systems. GPUSync was designed with flexibility, predictability, and parallelism in mind. Specifically, it can be applied under either static-or dynamic priority CPU scheduling, can allocate CPUs/GPUs on a partitioned, clustered, or global basis, provides flexible mechanisms for allocating GPUs to tasks, enables task state to be migrated among different GPUs, with the potential of breaking such state into smaller "chunks", provides migration cost predictors that determine when migrations can be effective, enables a single GPU's different engines to be accessed in parallel, properly supports GPU-related interrupt and worker threads according to the sporadic task model, even when GPU drivers are closed-source, and provides budget policing to the extent possible, given that GPU access is non-preemptive. No prior real-time GPU management framework provides a comparable range of features.

Architectures in which multicore chips are augmented with graphics processing units (GPUs) have great potential in many domains in which computationally intensive real-time workloads must be supported. However, unlike standard CPUs, GPUs are treated as I/O devices and require the use of interrupts to facilitate communication with CPUs. Given their disruptive nature, interrupts must be dealt with carefully in real-time systems. With GPU-driven interrupts, such disruptiveness is further compounded by the closed-source nature of GPU drivers. In this paper, such problems are considered and a solution is presented in the form of an extension to LITMUS^RT called klmirqd. The design of klmirqd targets systems with multiple CPUs and GPUs. In such settings, interrupt-related issues arise that have not been previously addressed.

Real-time locking protocols employ progress mechanism(s) to ensure that resource-holding jobs are scheduled. These mechanisms are required to bound the duration of priority-inversion blocking (pi-blocking) for jobs sharing resources. Examples of such progress mechanisms include priority inheritance and priority donation. Unfortunately, some progress mechanisms can cause any job, including those that never request shared resources, to be blocked upon job release. This paper presents a variant of priority donation for globally-scheduled systems that only causes blocking for jobs waiting for shared resources. Additionally, this variant of priority donation is employed to construct a new suspension-based locking protocol called the replica-request donation global locking protocol (R2DGLP), which is asymptotically optimal for both mutex and k-exclusion (i.e., multi-unit) resources. This work is motivated by multicore systems where tasks may share I/O devices (e.g., GPUs) where critical sections can be long. In such applications, progress mechanisms that cause jobs that do not access I/O devices to be blocked to ensure progress can be detrimental from a schedulability perspective.

Graphics processing units (GPUs) are becoming increasingly important in today's platforms as their increased generality allows for them to be used as powerful coprocessors. In this paper, we explore possible applications for GPUs in real-time systems, discuss the limitations and constraints imposed by current GPU technology, and present a summary of our research addressing many such constraints.