K
Keren Zhou
Researcher at Rice University
Publications - 23
Citations - 169
Keren Zhou is an academic researcher from Rice University. The author has contributed to research in topics: Computer science & Instrumentation (computer programming). The author has an hindex of 4, co-authored 19 publications receiving 93 citations. Previous affiliations of Keren Zhou include Yunnan University & Chinese Academy of Sciences.
Papers
More filters
Proceedings ArticleDOI
Understanding the GPU Microarchitecture to Achieve Bare-Metal Performance Tuning
TL;DR: The toolchain is an attempt to automatically crack different GPU ISA encodings and build an assembler adaptively for the purpose of performance enhancements to applications on GPUs.
Proceedings ArticleDOI
A performance analysis framework for exploiting GPU microarchitectural capability
TL;DR: A performance analysis framework at the assembly level that incorporates block partitions, occupancy, and the generated DAG to predict running cycles of the source code and presents its potential bottlenecks and can predict performance with an error as low as 2% in average.
Proceedings ArticleDOI
Tools for top-down performance analysis of GPU-accelerated applications
TL;DR: Extensions to Rice University's HPCToolkit performance tools to support measurement and analysis of GPU-accelerated applications and to support fine-grain analysis and tuning are described.
Proceedings ArticleDOI
GVPROF: A Value Profiler for GPU-Based Clusters
TL;DR: GVPROF as mentioned in this paper is the first value profiler that can detect temporal and spatial value redundancies in applications running on GPU-based clusters, which provides useful information to guide code optimization.
Proceedings ArticleDOI
Multi-Classes Feature Engineering with Sliding Window for Purchase Prediction in Mobile Commerce
TL;DR: This paper designs a feature engineering framework to extract 6 categories of features that aim to capture the buying potential of user-item pairs and applies Gradient Boost Decision Trees (GBDT) as the training model, and combines outputs of individual GBDT together by Logistic Regression to get the final predictions.