English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Domain-specific accelerators are used in various computing systems rangingfrom edge devices to data centers. Coarse-grained reconfigurable arrays (CGRAs)represent an architectural midpoint between the flexibility of an FPGA and theefficiency of an ASIC and are a promising candidate for servicing multi-taskedworkloads within an application domain. Unfortunately, scheduling multipletasks onto a CGRA is challenging. CGRAs lack abstractions that capture hardwareresources, leaving workload schedulers unable to reason about performance,energy, and utilization for different schedules. This work first proposes aCGRA architecture that can flexibly partition key resources, including theglobal buffer memory capacity, the global buffer memory bandwidth, and thecompute resources. Partitioned resources serve as hardware abstractions thatdecouple compilation and resource allocation. The compiler uses theseabstractions for coarse-grained resource mapping, and the scheduler uses themfor flexible resource allocation at run time. We then propose two hardwaremechanisms to support multi-task execution. A flexible-shape execution regionincreases the overall resource utilization by mapping multiple tasks withdifferent resource requirements. Dynamic partial reconfiguration (DPR) enablesa CGRA to update the hardware configuration as the scheduler makes decisionsrapidly. We show that our abstraction can help automatic and efficientscheduling of multi-tasked workloads onto our target CGRA with highutilization, resulting in 1.05x-1.24x higher throughput and a 23-28% lowerlatency in a multi-tasked cloud workload and 60.8% reduced latency in anautonomous system workload when compared to a baseline CGRA running singletasks at a time.

Hardware Abstractions and Hardware Mechanisms to Support Multi-Task Execution on Coarse-Grained Reconfigurable Arrays