P
Peng Jiang
Researcher at University of Iowa
Publications - 29
Citations - 357
Peng Jiang is an academic researcher from University of Iowa. The author has contributed to research in topics: Computer science & Graph (abstract data type). The author has an hindex of 7, co-authored 22 publications receiving 242 citations. Previous affiliations of Peng Jiang include Ohio State University & Chinese Academy of Sciences.
Papers
More filters
Proceedings Article
A linear speedup analysis of distributed deep learning with sparse and quantized communication
Peng Jiang,Gagan Agrawal +1 more
TL;DR: This paper studies the convergence rate of distributed SGD for non-convex optimization with two communication reducing strategies: sparse parameter averaging and gradient quantization and proposes a strategy called periodic quantized averaging (PQASGD) that further reduces the communication cost while preserving the O(1/√MK) convergence rate.
Proceedings ArticleDOI
Exploiting recent SIMD architectural advances for irregular applications
TL;DR: This paper designs an optimization methodology, which includes three sub-steps: 1) locality enhancement through tiling, 2) data access pattern identification, and 3) write conflict removal at both SIMD and MIMD levels, to effectively optimize different subclasses of irregular applications.
Proceedings ArticleDOI
A novel data transformation and execution strategy for accelerating sparse matrix multiplication on GPUs
TL;DR: This work proposes a novel row-reordering technique to improve data locality for SpMM and SDDMM on GPUs by using a hierarchical clustering procedure optimized by locality-sensitive hashing.
Proceedings ArticleDOI
Accelerating Sparse CNN Inference on GPUs with Performance-Aware Weight Pruning
TL;DR: The experimental results with five real-world pruned CNN models show that the techniques can significantly improve the layer-wise performance of sparse convolution operations as well as the end-to-end performance of CNN inference.
Proceedings ArticleDOI
Combining SIMD and Many/Multi-core Parallelism for Finite State Machines with Enumerative Speculation
Peng Jiang,Gagan Agrawal +1 more
TL;DR: This paper develops a novel strategy for combining SIMD and multi/many-core parallelism for FSMs that speculates transitions from several possible states, reducing the prediction overheads of speculation approach and the large amount of redundant work in the enumerative approach.