With growing Field Programmable Gate Array (FPGA) device sizes and their integration in environments enabling sharing of computing resources such as cloud and edge computing, there is a requirement to share the FPGA area between multiple tasks. The resource sharing typically involves partitioning the FPGA space into fix-sized slots. This results in suboptimal resource utilisation and relatively poor performance, particularly as the number of tasks increase. Using OpenCL’s exploration capabilities, we employ clever clustering and custom, task-specific partitioning and mapping to create a novel, area sharing methodology where task resource requirements are more effectively managed. Using models with varying resource/throughput profiles, we select the most appropriate distribution based on the runtime, workload needs to enhance temporal compute density. The approach is enabled in the system stack by a corresponding task-based virtualisation model. Using 11 high performance tasks from graph analysis, linear algebra and media streaming, we demonstrate an average 2.8×2.8\times higher system throughput at 2.3×2.3\times better energy efficiency over existing approaches.

中文翻译：

---

基于 FPGA 的计算系统上高效、动态的多任务执行


随着现场可编程门阵列 (FPGA) 设备尺寸的不断增长及其在支持共享计算资源（例如云和边缘计算）的环境中的集成，需要在多个任务之间共享 FPGA 区域。资源共享通常涉及将 FPGA 空间划分为固定大小的插槽。这会导致资源利用率不佳和性能相对较差，尤其是当任务数量增加时。利用 OpenCL 的探索功能，我们采用巧妙的集群和自定义、特定于任务的分区和映射来创建一种新颖的区域共享方法，在该方法中可以更有效地管理任务资源需求。使用具有不同资源/吞吐量配置文件的模型，我们根据运行时、工作负载需求选择最合适的分布，以增强时间计算密度。该方法通过相应的基于任务的虚拟化模型在系统堆栈中启用。使用图形分析、线性代数和媒体流中的 11 项高性能任务，我们证明了与现有方法相比，系统吞吐量平均提高了 2.8×2.8 倍，能源效率提高了 2.3×2.3 倍。