MAGMA — GPU-accelerated dense linear algebra

MaRDI QIDQ6825364swMATHFDO

Community MathAlgoDB

Described by source

• Dense linear algebra solvers for multicore with GPU accelerators
• MAGMA embedded: Towards a dense linear algebra library for energy efficient extreme computing

Cited in

(15)

• QR Decomposition
• Cholesky (LL^T) factorization for SPD matrices
• Cholesky Decomposition
• Singular Value Decomposition
• Conjugate Gradients
• LU decomposition
• Biconjugate Gradients
• Biconjugate Gradients Stabilized
• LU decomposition with (partial) pivoting for solving Ax=b and computing determinants
• QR decomposition (Householder/Givens) for least-squares and rank-revealing factorizations
• Generalized Minimal Residual
• Strassen-style fast matrix multiplication/inversion (subcubic complexity)
• Classical Gaussian elimination with pivoting plus back-substitution for dense linear systems
• Iterative refinement (iterative improvement) to enhance solutions of direct factorizations
• Singular Value Decomposition (SVD) for least-squares, pseudoinverses, and low-rank analysis

This page was built for software: MAGMA — GPU-accelerated dense linear algebra