Accelerators are becoming increasingly prevalent in distributed computation. FPGAs have been shown to be fast and power efficient for particular tasks, yet scheduling on FPGA-based multi-accelerator systems is challenging when workloads vary significantly in granularity in terms of task size and/or number of computational units required. We present a novel approach for dynamically scheduling tasks on networked multi-FPGA systems which maintains high performance, even in the presence of irregular tasks. Our topological ranking-based scheduling allows realistic irregular workloads to be processed while maintaining a significantly higher level of performance than existing schedulers.

This article summarizes our recent work and developments on SMPI, a flexible simulator of MPI applications. In this tool, we took a particular care to ensure our simulator could be used to produce fast and accurate predictions in a wide variety of situations. Although we did build SMPI on SimGrid whose speed and accuracy had already been assessed in other contexts, moving such techniques to a HPC workload required significant additional effort. Obviously, an accurate modeling of communications and network topology was one of the key to such achievements. Another less obvious key was the choice to combine in a single tool the possibility to do both offline and online simulation.

This paper presents a novel approach for allocating resources in systems with multiple accelerators. It has three main contributions. First, a new model based on Birkhoff's representation theory in capturing the ordering properties of resource allocation requests (RArs). Second, an effective technique for resource allocation based on this model, targeting systems with multiple accelerators. Third, the evaluation of the proposed approach for Maxeler MPC-X multi-accelerator systems, demonstrating time-efficiency and 30%-50% failure-rate decrease (FRD) on random input dataset.

We propose algorithms for allocating multiple resources to competing services running in virtual machines on heterogeneous distributed platforms. We develop a theoretical problem formulation and compare these algorithms via simulation experiments based in part on workload data supplied by Google. Our main finding is that vector packing approaches proposed in the homogeneous case can be extended to provide high-quality solutions in the heterogeneous case, and combined to provide a single efficient algorithm. We also consider the case when there may be bounded errors in estimates of performance-related resource needs. We provide a heuristic for compensating for such errors that performs well in simulation, as well as a proof of the worst-case competitive ratio for the single-resource, single-node case when there is no bound on the error.

We propose a novel job scheduling approach for homogeneous cluster computing platforms. Its key feature is the use of virtual machine technology to share fractional node resources in a precise and controlled manner. Other VM-based scheduling approaches have focused primarily on technical issues or extensions to existing batch scheduling systems, while we take a more aggressive approach and seek to find heuristics that maximize an objective metric correlated with job performance. We derive absolute performance bounds and develop algorithms for the online nonclairvoyant version of our scheduling problem. We further evaluate these algorithms in simulation against both synthetic and real-world HPC workloads and compare our algorithms to standard batch scheduling approaches. We find that our approach improves over batch scheduling by orders of magnitude in terms of job stretch, while leading to comparable or better resource utilization. Our results demonstrate that virtualization technology coupled with lightweight online scheduling strategies can afford dramatic improvements in performance for executing HPC workloads.

Simulation is a popular approach for predicting the performance of MPI applications for platforms that are not at one's disposal. It is also a way to teach the principles of parallel programming and high-performance computing to students without access to a parallel computer. In this work we present SMPI, a simulator for MPI applications that uses on-line simulation, i.e., the application is executed but part of the execution takes place within a simulation component. SMPI simulations account for network contention in a fast and scalable manner. SMPI also implements an original and validated piece-wise linear model for data transfer times between cluster nodes. Finally SMPI simulations of large-scale applications on large-scale platforms can be executed on a single node thanks to techniques to reduce the simulation's compute time and memory footprint. These contributions are validated via a large set of experiments in which SMPI is compared to popular MPI implementations with a view to assess its accuracy, scalability, and speed.

We propose a novel job scheduling approach for homogeneous cluster computing platforms. Its key feature is the use of virtual machine technology for sharing resources in a precise and controlled manner. We justify our approach and propose several job scheduling algorithms. We present results obtained in simulations for synthetic and real-world High Performance Computing (HPC) workloads, in which we compare our proposed algorithms with standard batch scheduling algorithms. We find that our approach widely outperforms batch scheduling. We also identify a few promising algorithms that perform well across most experimental scenarios. Our results demonstrate that virtualization technology coupled with lightweight scheduling strategies affords dramatic improvements in performance for HPC workloads.

We propose a novel approach for sharing cluster resources among competing jobs. The key advantage of our approach over current solutions is that it increases cluster utilization while optimizing a user-centric metric that captures both notions of performance and fairness. We motivate and formalize the corresponding resource allocation problem, determine its complexity, and propose several algorithms to solve it in the case of a static workload that consists of sequential jobs. Via extensive simulation experiments we identify an algorithm that runs quickly, that is always on par with or better than its competitors, and that produces resource allocations that are close to optimal. We find that the extension of our approach to parallel jobs leads to similarly good results. Finally, we explain how to extend our work to dynamic workloads.