A
An-Chow Lai
Researcher at Universiti Tunku Abdul Rahman
Publications - 35
Citations - 814
An-Chow Lai is an academic researcher from Universiti Tunku Abdul Rahman. The author has contributed to research in topics: Distributed shared memory & Photovoltaic system. The author has an hindex of 8, co-authored 33 publications receiving 740 citations. Previous affiliations of An-Chow Lai include National Cheng Kung University & Purdue University.
Papers
More filters
Proceedings ArticleDOI
Dead-block prediction & dead-block correlating prefetchers
TL;DR: The Dead-Block Predictors (DBPs) are proposed, trace-based predictors that accurately identify “when” an Ll data cache block becomes evictable or “dead”, and a DBCP enables effective data prefetching in a wide spectrum of pointer-intensive, integer, and floating-point applications.
Proceedings ArticleDOI
Selective, accurate, and timely self-invalidation using last-touch prediction
An-Chow Lai,Babak Falsafi +1 more
TL;DR: Results from running shared-memory applications on a simulated DSM to evaluate LTPs indicate that LTP enables selective, accurate, and timely self-invalidation in DSM, speeding up program execution on average by 11%.
Proceedings ArticleDOI
Memory sharing predictor: the key to a speculative coherent DSM
An-Chow Lai,Babak Falsafi +1 more
TL;DR: The Memory Sharing Predictors (MSPs) are proposed, pattern-based predictors that significantly improve prediction accuracy and implementation cost over general message predictors and identify simple techniques and mechanisms to trigger prediction timely and perform speculation for remote read accesses.
Proceedings ArticleDOI
Load balancing in distributed shared memory systems
TL;DR: This paper proposes and experimentally evaluates a load balancing method called Dependence-Driven Load Balancing (DDLB) that is dedicated for multithreaded DSM systems, and its most attractive feature is to take thread dependence into account in making decisions for migration.
Patent
Method and an apparatus for interleaving read data return in a packetized interconnect to memory
TL;DR: In this article, a cache line of each of a number of read data returns is packed into one or more packets, and each of the packets are then split into a plurality of flits.