Filters








12,032 Hits in 5.5 sec

Automatic Selection of Sparse Matrix Representation on GPUs

Naser Sedaghati, Te Mu, Louis-Noel Pouchet, Srinivasan Parthasarathy, P. Sadayappan
2015 Proceedings of the 29th ACM on International Conference on Supercomputing - ICS '15  
We then build a decision model using machine learning to automatically select the best representation to use for a given sparse matrix on a given target platform, based on the sparse matrix features.  ...  Experimental results on three GPUs demonstrate that the approach is very effective in selecting the best representation.  ...  We thank NVIDIA Corporation for donating the K40c GPU. This work was supported in part by National Science Foundation awards CCF-0926127, CCF-1240651, CCF-1321147, CCF-1418265 and ACI-1404995.  ... 
doi:10.1145/2751205.2751244 dblp:conf/ics/SedaghatiMPPS15 fatcat:2nq77quj5fa2lngf2gzym6jwee

Characterizing dataset dependence for sparse matrix-vector multiplication on GPUs

Naser Sedaghati, Arash Ashari, Louis-Noël Pouchet, Srinivasan Parthasarathy, P. Sadayappan
2015 Proceedings of the 2nd Workshop on Parallel Programming for Analytics Applications - PPAA 2015  
Many GPU implementations of SpMV have been proposed, proposing different sparse matrix representations.  ...  We present some insights on the correlation between matrix features and the best choice for sparse matrix representation.  ...  Kernel Performance for HYB The HYB representation automatically selects a cut-off point (i.e. k) for separation of dense (i.e. ELL) and sparse (i.e. COO) parts of a matrix.  ... 
doi:10.1145/2726935.2726941 fatcat:4mnoyl4zljfm7o4wqmntah377u

Auto-Selection of an Optimal Sparse Matrix Format in the Neuro-Simulator ANNarchy

Helge Ülo Dinkelbach, Badr-Eddine Bouhlal, Julien Vitay, Fred H. Hamker
2022 Frontiers in Neuroinformatics  
We employ an artificial neural network to develop a predictive model to help the developer select the optimal sparse matrix format.  ...  We pinpoint the role of sparse matrix formats implemented in the neuro-simulator ANNarchy with respect to computation time.  ...  Sparse Matrix Formats Sparse matrix representations require a memory overhead to index the elements of a matrix (e.g., row pointers).  ... 
doi:10.3389/fninf.2022.877945 pmid:35676973 pmcid:PMC9169689 fatcat:freiv4kyobflvbdxz5xx3wmjra

Effective Sparse Matrix Representation For The GPU Architectures

B Neelima
2012 International Journal of Computer Science Engineering and Applications  
It also gives 10% to 133% improvements in memory transfer (of only access information of sparse matrix) between CPU and GPU.  ...  One such very widely used computation intensive kernel is sparse matrix vector multiplication (SPMV) in sparse matrix based applications.  ...  Table 2 represents the general characteristics of the selected matrices. Fig. 3 to Fig. 5 shows the undirected or bipartite graph representation of the selected matrices.  ... 
doi:10.5121/ijcsea.2012.2213 fatcat:mg6jop56m5b67dwghffght6h24

Deep Learning with Apache SystemML [article]

Niketan Pansare, Michael Dusenberry, Nakul Jindal, Matthias Boehm, Berthold Reinwald, Prithviraj Sen
2018 arXiv   pre-print
SystemML's novel compilation approach automatically generates runtime execution plans for machine/deep learning algorithms that are composed of single-node and distributed runtime operations depending  ...  Enterprises operate large data lakes using Hadoop and Spark frameworks that (1) run a plethora of tools to automate powerful data preparation/transformation pipelines, (2) run on shared, large clusters  ...  SystemML maintains the number of nonzeros for each intermediate matrix, decides upon dense or sparse formats, and selects appropriate runtime operators for combinations of dense and sparse inputs.  ... 
arXiv:1802.04647v1 fatcat:jkt6zv6tfjghnbzvvjhlrtioay

Sparse Matrix Classification on Imbalanced Datasets using Convolutional Neural Networks

Juan C. Pichel, Beatriz Pateiro-Lopez
2019 IEEE Access  
This paper deals with the class imbalance problem in the context of the automatic selection of the best storage format for a sparse matrix with the aim of maximizing the performance of the sparse matrix  ...  vector multiplication (SpMV) on GPUs.  ...  In this paper we address the automatic classification of sparse matrices to select the best SpMV performing storage format on GPUs using CNNs.  ... 
doi:10.1109/access.2019.2924060 fatcat:zxdjbblgzvaozcudjvbyibufty

A New Approach for Sparse Matrix Classification Based on Deep Learning Techniques

Juan C. Pichel, Beatriz Pateiro-Lopez
2018 2018 IEEE International Conference on Cluster Computing (CLUSTER)  
We focus on the selection of the proper format for the sparse matrixvector multiplication (SpMV), which is one of the most important computational kernels in many scientific and engineering applications  ...  In this paper, a new methodology to select the best storage format for sparse matrices based on deep learning techniques is introduced.  ...  In this paper we address the problem of the automatic selection of the best storage format for sparse matrices on GPUs.  ... 
doi:10.1109/cluster.2018.00017 dblp:conf/cluster/PichelP18 fatcat:fanbx4zzgjhzvjfsxto7oywg7u

Adaptive Optimization of Sparse Matrix-Vector Multiplication on Emerging Many-Core Architectures

Shizhao Chen, Jianbin Fang, Donglin Chen, Chuanfu Xu, Zheng Wang
2018 2018 IEEE 20th International Conference on High Performance Computing and Communications; IEEE 16th International Conference on Smart City; IEEE 4th International Conference on Data Science and Systems (HPCC/SmartCity/DSS)  
This paper provides the first comprehensive study on the impact of sparse matrix representations on two emerging many-core architectures: the Intel's Knights Landing (KNL) XeonPhi and the ARM-based FT-  ...  We show that the best sparse matrix representation depends on the underlying architecture and input.  ...  Further, they build a model based on decision tree to automatically select the best representation for a given sparse matrix on a given GPU platform.  ... 
doi:10.1109/hpcc/smartcity/dss.2018.00116 dblp:conf/hpcc/ChenFCXW18 fatcat:nvtcmxgdt5afzmbnfl4zq46fga

Optimizing Sparse Matrix-Vector Multiplication on Emerging Many-Core Architectures [article]

Shizhao Chen, Jianbin Fang, Donglin Chen, Chuanfu Xu, Zheng Wang
2018 arXiv   pre-print
This paper provides the first comprehensive study on the impact of sparse matrix representations on two emerging many-core architectures: the Intel's Knights Landing (KNL) XeonPhi and the ARM-based FT-  ...  We show that the best sparse matrix representation depends on the underlying architecture and the program input.  ...  Further, they build a model based on decision tree to automatically select the best representation for a given sparse matrix on a given GPU platform.  ... 
arXiv:1805.11938v1 fatcat:jefi2bu2nzewxgzyvjitiipu7i

Implementing Sparse Matrix-Vector Multiplication with QCSR on GPU

Jilin Zhang, Enyi Liu, Jian Wan, Yongjian Ren, Miao Yue, Jue Wang
2013 Applied Mathematics & Information Sciences  
Various data formats to store the sparse matrix have been implemented on GPUs to maximize the performance.  ...  In this paper, we propose and evaluate a new implementation of SPMV on GPU based on QCSR storage format which combines the quadtree storage format and CSR format.  ...  Figure 4 ELLP ACK storage format representation of sparse matrix A.  ... 
doi:10.12785/amis/070207 fatcat:okhgxfolzjbp5mb4tevbzf25bq

A Performance Modeling and Optimization Analysis Tool for Sparse Matrix-Vector Multiplication on GPUs

Ping Guo, Liqiang Wang, Po Chen
2014 IEEE Transactions on Parallel and Distributed Systems  
This paper presents a performance modeling and optimization analysis tool to predict and optimize the performance of sparse matrix-vector multiplication (SpMV) on GPUs.  ...  optimal solution auto-selection algorithm to automatically report an optimal solution (i.e., optimal storage strategy, storage format(s), and execution time) for a target sparse matrix.  ...  ACKNOWLEDGMENTS The work was supported in part by NSF under Grants 0941735, CAREER-1054834, and by the Graduate Assistantship of the School of Energy Resources at the University of Wyoming.  ... 
doi:10.1109/tpds.2013.123 fatcat:apzbps2fq5eitgvalrr6mqgd6a

Pyxis: An Open-Source Performance Dataset of Sparse Accelerators [article]

Linghao Song, Yuze Chi, Jason Cong
2022 arXiv   pre-print
Specialized accelerators provide gains of performance and efficiency in specific domains of applications. Sparse data structures or/and representations exist in a wide range of applications.  ...  Accelerator researchers rely on real execution to get precise feedback for their designs. In this work, we present PYXIS, a performance dataset for specialized accelerators on sparse data.  ...  An input sparse matrix/graph determines the M and K, but users select the N value for their applications.  ... 
arXiv:2110.04280v2 fatcat:zqhhxv5rozeehney5y2qjffeme

Exploiting dynamic sparse matrices for performance portable linear algebra operations [article]

Chris Stylianou, Michele Weiland
2022 arXiv   pre-print
and 7x on CPUs and GPUs respectively, through runtime selection of the best format on each MPI process.  ...  More than 70 sparse matrix storage formats have been developed over the years, targeting a wide range of hardware architectures and matrix types.  ...  ACKNOWLEDGMENT This research is part of the EPSRC project ASiMoV (EP/S005072/1).  ... 
arXiv:2209.06478v1 fatcat:isjkjmxapfdspcottlgcs2nocq

Compressed dynamic mode decomposition for background modeling

N. Benjamin Erichson, Steven L. Brunton, J. Nathan Kutz
2016 Journal of Real-Time Image Processing  
The key principal of cDMD is to obtain the decomposition on a (small) compressed matrix representation of the video feed.  ...  Selection of the optimal modes characterizing the background is formulated as a sparsity-constrained sparse coding problem.  ...  Once a mode is selected, the initial condition x 1 is orthogonally projected on the span of the previously selected set of modes.  ... 
doi:10.1007/s11554-016-0655-2 fatcat:76gq77wzxza4dd3fe3wc7rmto4

Automatically harnessing sparse acceleration

Philip Ginsbach, Bruce Collie, Michael F. P. O'Boyle
2020 Proceedings of the 29th International Conference on Compiler Construction  
Appropriate places for library insertion are detected in compiler intermediate representation, independent of source languages.  ...  In this paper, we develop a new approach based on our specification Language for implementers of Linear Algebra Computations (LiLAC).  ...  Previous work [25, 30] established systems for automatic management of CPU-GPU communication.  ... 
doi:10.1145/3377555.3377893 dblp:conf/cc/GinsbachCO20 fatcat:wf6utlth6na7ddronimzit5xzq
« Previous Showing results 1 — 15 out of 12,032 results