Combining Sparse Approximate Factorizations with Mixed-precision Iterative Refinement
نویسندگان
چکیده
The standard LU factorization-based solution process for linear systems can be enhanced in speed or accuracy by employing mixed-precision iterative refinement. Most recent work has focused on dense systems. We investigate the potential of refinement to enhance methods sparse based approximate factorizations. In doing so, we first develop a new error analysis LU- and GMRES-based under general model factorization that accounts approximation typically used modern solvers, such as low-rank approximations relaxed pivoting strategies. then provide detailed performance both execution time memory consumption different algorithms, selected set variants Our study uses multifrontal solver MUMPS, which exploit block static pivoting. evaluate algorithms large, problems coming from variety real-life industrial applications showing combined with lead considerable reductions consumption.
منابع مشابه
GPU-Accelerated Asynchronous Error Correction for Mixed Precision Iterative Refinement
In hardware-aware high performance computing, blockasynchronous iteration and mixed precision iterative refinement are two techniques that are applied to leverage the computing power of SIMD accelerators like GPUs. Although they use a very different approach for this purpose, they share the basic idea of compensating the convergence behaviour of an inferior numerical algorithm by a more efficie...
متن کاملLU-Decomposition with Iterative Refinement for Solving Sparse Linear Systems
In the solution of a system of linear algebraic equations Ax = b with a large sparse coefficient matrix A, the LU-decomposition with iterative refinement (LUIR) is compared with the LU-decomposition with direct solution (LUDS), which is without iterative refinement. We verify by numerical experiments that the use of sparse matrix techniques with LUIR may result in a reduction of both the comput...
متن کاملSparse CCA via Precision Adjusted Iterative Thresholding
Sparse Canonical Correlation Analysis (CCA) has received considerable attention in high-dimensional data analysis to study the relationship between two sets of random variables. However, there has been remarkably little theoretical statistical foundation on sparse CCA in high-dimensional settings despite active methodological and applied research activities. In this paper, we introduce an eleme...
متن کاملMixed Precision Iterative Refinement Techniques for the Solution of Dense Linear Systems
By using a combination of 32-bit and 64-bit floating point arithmetic, the performance of many dense and sparse linear algebra algorithms can be significantly enhanced while maintaining the 64-bit accuracy of the resulting solution. The approach presented here can apply not only to conventional processors but also to exotic technologies such as Field Programmable Gate Arrays (FPGA), Graphical P...
متن کاملIterative Refinement of the Approximate Posterior for Directed Belief Networks
Variational methods that rely on a recognition network to approximate the posterior of directed graphical models offer better inference and learning than previous methods. Recent advances that exploit the capacity and flexibility in this approach have expanded what kinds of models can be trained. However, as a proposal for the posterior, the capacity of the recognition network is limited, which...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: ACM Transactions on Mathematical Software
سال: 2023
ISSN: ['0098-3500', '1557-7295']
DOI: https://doi.org/10.1145/3582493