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

Emerging real-time multi-model ML (RTMM) workloads such as AR/VR and dronecontrol involve dynamic behaviors in various granularity; task, model, andlayers within a model. Such dynamic behaviors introduce new challenges to thesystem software in an ML system since the overall system load is not completelypredictable, unlike traditional ML workloads. In addition, RTMM workloadsrequire real-time processing, involve highly heterogeneous models, and targetresource-constrained devices. Under such circumstances, developing an effectivescheduler gains more importance to better utilize underlying hardwareconsidering the unique characteristics of RTMM workloads. Therefore, we proposea new scheduler, DREAM, which effectively handles various dynamicity in RTMMworkloads targeting multi-accelerator systems. DREAM quantifies the uniquerequirements for RTMM workloads and utilizes the quantified scores to drivescheduling decisions, considering the current system load and other inferencejobs on different models and input frames. DREAM utilizes tunable parametersthat provide fast and effective adaptivity to dynamic workload changes. In ourevaluation of five scenarios of RTMM workload, DREAM reduces the overallUXCost, which is an equivalent metric of the energy-delay product (EDP) forRTMM defined in the paper, by 32.2% and 50.0% in the geometric mean (up to80.8% and 97.6%) compared to state-of-the-art baselines, which shows theefficacy of our scheduling methodology.

DREAM: A Dynamic Scheduler for Dynamic Real-time Multi-model ML Workloads