咨询与建议

限定检索结果

文献类型

  • 3 篇 期刊文献
  • 3 篇 会议

馆藏范围

  • 6 篇 电子文献
  • 0 种 纸本馆藏

日期分布

学科分类号

  • 5 篇 工学
    • 5 篇 计算机科学与技术...
    • 2 篇 软件工程
    • 1 篇 电气工程
  • 2 篇 理学
    • 2 篇 数学

主题

  • 6 篇 mixed precision ...
  • 2 篇 rounding error a...
  • 2 篇 low-rank approxi...
  • 2 篇 tensor cores
  • 2 篇 singular value d...
  • 2 篇 gpu computing
  • 2 篇 numerical linear...
  • 2 篇 lu factorization
  • 1 篇 parallel archite...
  • 1 篇 central processi...
  • 1 篇 powerpack
  • 1 篇 libraries
  • 1 篇 matlab
  • 1 篇 rapl
  • 1 篇 architecture
  • 1 篇 active libraries
  • 1 篇 algorithms
  • 1 篇 data sparse matr...
  • 1 篇 gpu-based scient...
  • 1 篇 graphics process...

机构

  • 1 篇 oak ridge natl l...
  • 1 篇 king abdullah un...
  • 1 篇 natl sci fdn & m...
  • 1 篇 univ paris sacla...
  • 1 篇 univ paris panth...
  • 1 篇 univ tennessee i...
  • 1 篇 inria f-91405 or...
  • 1 篇 cnrs irit 2 rue ...
  • 1 篇 univ paris 11 f-...
  • 1 篇 sorbonne univ cn...
  • 1 篇 univ paris sacla...
  • 1 篇 sorbonne univ cn...
  • 1 篇 ansys co livermo...
  • 1 篇 univ manchester ...
  • 1 篇 kaust supercompu...
  • 1 篇 edf r&d f-91120 ...
  • 1 篇 ens lyon mumps t...
  • 1 篇 univ paris sacla...
  • 1 篇 inst dev & resso...
  • 1 篇 edf r&d f-75005 ...

作者

  • 3 篇 mary theo
  • 2 篇 baboulin marc
  • 2 篇 ltaief hatem
  • 1 篇 jezequel fabienn...
  • 1 篇 dongarra jack
  • 1 篇 robeyns matthieu
  • 1 篇 sukkari dalal
  • 1 篇 falcou joel
  • 1 篇 keyes david
  • 1 篇 buttari alfredo
  • 1 篇 donfack simplice
  • 1 篇 masliah ian
  • 1 篇 luszczek piotr
  • 1 篇 kaya oguz
  • 1 篇 boiteau olivier
  • 1 篇 lopez florent
  • 1 篇 amestoy patrick
  • 1 篇 gerest matthieu
  • 1 篇 l'excellent jean...
  • 1 篇 weaver vincent m...

语言

  • 6 篇 英文
检索条件"主题词=mixed precision algorithms"
6 条 记 录,以下是1-10 订阅
排序:
mixed precision LU factorization on GPU tensor cores: reducing data movement and memory footprint
收藏 引用
INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS 2023年 第2期37卷 165-179页
作者: Lopez, Florent Mary, Theo ANSYS Co Livermore Software Technol Canonsburg PA 15317 USA Sorbonne Univ CNRS 4 Pl JussieuLIP6 Paris France
Modern GPUs equipped with mixed precision tensor core units present great potential to accelerate dense linear algebra operations such as LU factorization. However, state-of-the-art mixed half/single precision LU fact... 详细信息
来源: 评论
mixed precision low-rank approximations and their application to block low-rank LU factorization
收藏 引用
IMA JOURNAL OF NUMERICAL ANALYSIS 2023年 第4期43卷 2198-2227页
作者: Amestoy, Patrick Boiteau, Olivier Buttari, Alfredo Gerest, Matthieu Jezequel, Fabienne L'excellent, Jean-Yves Mary, Theo ENS Lyon Mumps Technol 46 Allee Italie F-69007 Lyon France EDF R&D F-91120 Palaiseau France CNRS IRIT 2 Rue Charles Camichel F-31071 Toulouse France EDF R&D F-75005 Paris France Sorbonne Univ CNRS LIP6 F-75005 Paris France Univ Paris Pantheon Assas F-75005 Paris France
We introduce a novel approach to exploit mixed precision arithmetic for low-rank approximations. Our approach is based on the observation that singular vectors associated with small singular values can be stored in lo... 详细信息
来源: 评论
mixed precision Randomized Low-Rank Approximation with GPU Tensor Cores  30th
Mixed Precision Randomized Low-Rank Approximation with GPU T...
收藏 引用
30th European Conference on Parallel and Distributed Processing (Euro-Par)
作者: Baboulin, Marc Donfack, Simplice Kaya, Oguz Mary, Theo Robeyns, Matthieu Univ Paris Saclay CNRS ENS Paris Saclay LMF Gif Sur Yvette France Univ Paris Saclay UVSQ INRIA CNRSCEAMaison Simulat Gif Sur Yvette France Univ Paris Saclay CNRS LISN Orsay France Sorbonne Univ CNRS LIP6 Paris France Inst Dev & Ressources Informat Sci Rue John von Neumann F-91403 Orsay France
Randomized projection methods have been shown to be very efficient at computing low-rank approximations (LRA) of large matrices. In this work, we investigate the design and development of such methods capable of explo... 详细信息
来源: 评论
Energy Footprint of Advanced Dense Numerical Linear Algebra using Tile algorithms on Multicore Architectures
Energy Footprint of Advanced Dense Numerical Linear Algebra ...
收藏 引用
2nd International Conference on Cloud and Green Computing / 2nd International Conference on Social Computing and its Applications (CGC/SCA)
作者: Dongarra, Jack Ltaief, Hatem Luszczek, Piotr Weaver, Vincent M. Univ Tennessee Innovat Comp Lab Knoxville TN 37996 USA Oak Ridge Natl Lab Comp Sci & Math Div Oak Ridge TN USA Univ Manchester Sch Math Sch Comp Sci Manchester NH USA Natl Sci Fdn & Microsoft Research Manchester NH USA KAUST Supercomputing Lab Thuwal Saudi Arabia
We propose to study the impact on the energy footprint of two advanced algorithmic strategies in the context of high performance dense linear algebra libraries: (1) mixed precision algorithms with iterative refinement... 详细信息
来源: 评论
Metaprogramming dense linear algebra solvers Applications to multi and many-core architectures  14
Metaprogramming dense linear algebra solvers Applications to...
收藏 引用
13th IEEE International Symposium on Parallel and Distributed Processing with Applications
作者: Masliah, Ian Baboulin, Marc Falcou, Joel Univ Paris 11 F-91405 Orsay France Inria F-91405 Orsay France
The increasing complexity of new parallel architectures has widened the gap between adaptability and efficiency of the codes. As high performance numerical libraries tend to focus more on performance, we wish to addre... 详细信息
来源: 评论
A High Performance QDWH-SVD Solver Using Hardware Accelerators
收藏 引用
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE 2016年 第1期43卷 6-6页
作者: Sukkari, Dalal Ltaief, Hatem Keyes, David King Abdullah Univ Sci & Technol Extreme Comp Res 4700 King Abdullah Blvd Thuwal Jeddah 23955 Saudi Arabia
This article describes a new high performance implementation of the QR-based Dynamically Weighted Halley Singular Value Decomposition (QDWH-SVD) solver on multicore architecture enhanced with multiple GPUs. The standa... 详细信息
来源: 评论