Q
Qinyi Luo
Researcher at University of Southern California
Publications - 8
Citations - 112
Qinyi Luo is an academic researcher from University of Southern California. The author has contributed to research in topics: Speedup & Overhead (computing). The author has an hindex of 5, co-authored 8 publications receiving 75 citations.
Papers
More filters
Proceedings ArticleDOI
Prague: High-Performance Heterogeneity-Aware Asynchronous Decentralized Training
TL;DR: The proposed Prague, a high-performance heterogeneity-aware asynchronous decentralized training approach, achieves the above goal with intensive synchronization optimization by exploring the interplay between algorithm and system implementation, or statistical and hardware efficiency.
Proceedings ArticleDOI
Hop: Heterogeneity-aware Decentralized Training
TL;DR: Hop as discussed by the authors proposes a queue-based synchronization mechanism that can efficiently implement backup workers and bounded staleness in the decentralized setting to cope with deterministic slowdown, and skip iterations so that the effect of slower workers is further mitigated.
Proceedings ArticleDOI
SympleGraph: distributed graph processing with precise loop-carried dependency guarantee
TL;DR: SympleGraph is proposed, a novel distributed graph processing framework that precisely enforces loop-carried dependency, i.e., when a condition is satisfied by a neighbor, all following neighbors can be skipped and achieves a good trade-off between precise semantics and parallelism.
Posted Content
Heterogeneity-Aware Asynchronous Decentralized Training
TL;DR: This paper proposes Ripples, a high-performance heterogeneity-aware asynchronous decentralized training approach that achieves the above goal with intensive synchronization optimization, emphasizing the interplay between algorithm and system implementation.
Proceedings ArticleDOI
Hop: Heterogeneity-Aware Decentralized Training
TL;DR: This paper proposes Hop, the first heterogeneity-aware decentralized training protocol, a queue-based synchronization mechanism that can efficiently implement backup workers and bounded staleness in the decentralized setting and proposes skipping iterations so that the effect of slower workers is further mitigated.