A Prediction Framework for Fast Sparse Triangular Solves

Ahmad, Najeeb; Yilmaz, Buse; Unat, Didem

doi:10.1007/978-3-030-57675-2_33

Cited by 6 publications

(2 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The experimental evaluation carried out previously reveals that some procedures are able to attain predic-tions with values of 80% of accuracy. A similar approach has been recently followed by other authors [15]. This effort extends [13] in several lines with the purpose of improving the performance of the whole selection procedure.…”

Section: Introductionmentioning

confidence: 94%

Machine learning for optimal selection of sparse triangular system solvers on GPUs

Dufrechou

Ezzatti

Freire

et al. 2021

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 94%

Machine learning for optimal selection of sparse triangular system solvers on GPUs

Dufrechou

Ezzatti

Freire

et al. 2021

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

“…Nowadays, with the rapid development of computer hardware, adaptive method for data format or kernel selection has becoming a promising research direction, especially in applications that involve sparse matrix computations. Examples can be found in works for automatic selection of sparse matrix format in SpMV [40][41][42][43][44][45][46] and spGEMM [47,48], and the adaptive switch of computing kernels such as SpMV/SpMSpV [49] and SpTRSV [50]. Recently, the adaptive idea is also used in MTTKRP sequence computation arising in the CP decomposition of higher-order tensors [51].…”

Section: Related Workmentioning

confidence: 99%

a-Tucker: Input-Adaptive and Matricization-Free Tucker Decomposition for Dense Tensors on CPUs and GPUs

Li¹,

Xiao²,

Yang³

2020

Preprint

View full text Add to dashboard Cite

Tucker decomposition is one of the most popular models for analyzing and compressing large-scale tensorial data. Existing Tucker decomposition algorithms usually rely on a single solver to compute the factor matrices and core tensor, and are not flexible enough to adapt with the diversities of the input data and the hardware. Moreover, to exploit highly efficient GEMM kernels, most Tucker decomposition implementations make use of explicit matricizations, which could introduce extra costs in terms of data conversion and memory usage. In this paper, we present a-Tucker, a new framework for input-adaptive and matricization-free Tucker decomposition of dense tensors. A mode-wise flexible Tucker decomposition algorithm is proposed to enable the switch of different solvers for the factor matrices and core tensor, and a machine-learning adaptive solver selector is applied to automatically cope with the variations of both the input data and the hardware. To further improve the performance and enhance the memory efficiency, we implement a-Tucker in a fully matricization-free manner without any conversion between tensors and matrices. Experiments with a variety of synthetic and real-world tensors show that a-Tucker can substantially outperform existing works on both CPUs and GPUs.

show abstract

a-Tucker: fast input-adaptive and matricization-free Tucker decomposition of higher-order tensors on GPUs

Duan

Xiao

et al. 2022

CCF Trans. HPC

View full text Add to dashboard Cite

A Prediction Framework for Fast Sparse Triangular Solves

Cited by 6 publications

References 27 publications

Machine learning for optimal selection of sparse triangular system solvers on GPUs

Machine learning for optimal selection of sparse triangular system solvers on GPUs

a-Tucker: Input-Adaptive and Matricization-Free Tucker Decomposition for Dense Tensors on CPUs and GPUs

a-Tucker: fast input-adaptive and matricization-free Tucker decomposition of higher-order tensors on GPUs

Contact Info

Product

Resources

About