Sparsetir: Composable abstractions for sparse compilation in deep learning
Sparse tensors are rapidly becoming critical components of modern deep learning
workloads. However, developing high-performance sparse operators can be difficult and …
workloads. However, developing high-performance sparse operators can be difficult and …
Sparse supernodal solver using block low-rank compression: Design, performance and analysis
This paper presents two approaches using a Block Low-Rank (BLR) compression technique
to reduce the memory footprint and/or the time-to-solution of the sparse supernodal solver …
to reduce the memory footprint and/or the time-to-solution of the sparse supernodal solver …
Some new techniques to use in serial sparse Cholesky factorization algorithms
MO Karsavuran, EG Ng, BW Peyton… - arXiv preprint arXiv …, 2024 - arxiv.org
We present a new variant of serial right-looking supernodal sparse Cholesky factorization
(RL). Our comparison of RL with the multifrontal method confirms that RL is simpler, slightly …
(RL). Our comparison of RL with the multifrontal method confirms that RL is simpler, slightly …
Improving predication efficiency through compaction/restoration of simd instructions
A Barredo, JM Cebrian, M Moretó… - … symposium on high …, 2020 - ieeexplore.ieee.org
Vector processors offer a wide range of unexplored opportunities to improve performance
and energy efficiency. However, despite its potential, vector code generation and execution …
and energy efficiency. However, despite its potential, vector code generation and execution …
Sparse supernodal solver using block low-rank compression
This paper presents two approaches using a Block Low-Rank (BLR) compression technique
to reduce the memory footprint and/or the time-to-solution of the sparse supernodal solver …
to reduce the memory footprint and/or the time-to-solution of the sparse supernodal solver …
Optimizing partitioned CSR-based SpGEMM on the Sunway TaihuLight
General sparse matrix-sparse matrix (SpGEMM) multiplication is one of the basic kernels in
a great many applications. Several works focus on various optimizations for SpGEMM. To …
a great many applications. Several works focus on various optimizations for SpGEMM. To …
Sparsity analysis and optimization for state-space-based simulation of power electronic systems
Z Yu, B Shi, S Jia, H Xu, Y Xiao… - CSEE Journal of Power …, 2024 - ieeexplore.ieee.org
Computer-aided analysis tool is playing an important role in the design of power electronics
converters. However, with the increase of system scale and complexity, the existing …
converters. However, with the increase of system scale and complexity, the existing …
Improved sparsity techniques for solving network equations in transient stability simulations
T Xiao, J Wang, Y Gao, D Gan - IEEE Transactions on Power …, 2018 - ieeexplore.ieee.org
When solving network algebraic equations during power system transient stability
simulations, the nonzero elements in the independent vector and the elements needed in …
simulations, the nonzero elements in the independent vector and the elements needed in …
Blocking Sparse Matrices to Leverage Dense-Specific Multiplication
Research to accelerate matrix multiplication, pushed by the growing computational
demands of deep learning, has sprouted many efficient architectural solutions, such as …
demands of deep learning, has sprouted many efficient architectural solutions, such as …
On the use of low-rank arithmetic to reduce the complexity of parallel sparse linear solvers based on direct factorization techniques
G Pichon - 2018 - inria.hal.science
Solving sparse linear systems is a problem that arises in many scientific applications, and
sparse direct solvers are a time consuming and key kernel for those applications and for …
sparse direct solvers are a time consuming and key kernel for those applications and for …