M
Matthew G. F. Dosanjh
Researcher at Sandia National Laboratories
Publications - 23
Citations - 222
Matthew G. F. Dosanjh is an academic researcher from Sandia National Laboratories. The author has contributed to research in topics: Multithreading & Matching (statistics). The author has an hindex of 9, co-authored 23 publications receiving 170 citations. Previous affiliations of Matthew G. F. Dosanjh include University of New Mexico.
Papers
More filters
Book ChapterDOI
Finepoints: Partitioned Multithreaded MPI Communication
TL;DR: In this paper, the MPI multithreading model has been shown to provide superior performance than current approaches and even outperform single-threaded MPI, and an interface designed specifically for threads was proposed.
Book ChapterDOI
Measuring Multithreaded Message Matching Misery
TL;DR: A framework for studying the effects of multithreading on MPI message matching is presented and surprising results on the challenge posed by message matching for multithreaded application performance are highlighted.
Proceedings ArticleDOI
RMA-MT: a benchmark suite for assessing MPI multi-threaded RMA performance
TL;DR: The design and demonstration of the first available proxy applications and micro-benchmark suite for multi-threaded RMA in MPI, a study of multi- threads RMA performance of different MPI implementations, and an evaluation of how these benchmarks can be used to test development for both performance and correctness.
Proceedings ArticleDOI
INCA: in-network compute assistance
TL;DR: INCA builds upon contemporary NIC offload capabilities to provide on-NIC, deadline-free, general-purpose compute capacities that can be utilized when the network is inactive, and is demonstrated to be Turing complete.
Proceedings ArticleDOI
Improving MPI Multi-threaded RMA Communication Performance
Nathan Hjelm,Matthew G. F. Dosanjh,Ryan E. Grant,Taylor Groves,Patrick G. Bridges,Dorian Arnold +5 more
TL;DR: This paper describes the design and implementation of a new RMA implementation for Open MPI that targets scalability and multi-threaded performance and offers an evaluation that demonstrates scaling to 524,288 cores, the full size of a leading supercomputer installation.