Algorithm 586: ITPACK 2C: A FORTRAN Package for Solving Large Sparse Linear Systems by Adaptive Accelerated Iterative Methods

Kincaid, David R.; Respess, John R.; Young, David M.; Grimes, Rober R.

doi:10.1145/356004.356009

Cited by 134 publications

(54 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The run-time redistribution of a sparse matrix used repeatedly in a matrix vector multiplication should become a common option in the standard library of sparse matrix operations, such as ITPACK [10]. We hope that the algorithms presented in this paper will enlarge the repertoire of run-time distribution algorithms available to the users of parallel machines.…”

Section: Resultsmentioning

confidence: 99%

“…In Fortran programs, the JA array is usually accessed via an indirection in a loop. As an example, consider the Fortran code (Figure 2(a)) for matrix-vector multiplication taken from ITPACK [10] which is used in many sparse linear solvers. If both VA and JA are distributed by rows and aligned with the multiplied vector x, then each processor is responsible for multiplying the rows allocated to it.…”

Section: Preprocessing and Data Distribution Of Sparse Matricesmentioning

confidence: 99%

See 1 more Smart Citation

Run-time optimization of sparse matrix-vector multiplication on SIMD machines

Ziantz

Özturan

Szymański

1994

PARLE'94 Parallel Architectures and Languages Europe

View full text Add to dashboard Cite

Abstract. Sparse matrix-vector multiplication forms the heart of iterative linear solvers used widely in scientific computations (e.g., finite element methods). In such solvers, the matrix-vector product is computed repeatedly, often thousands of times, with updated values of the vector until convergence is achieved. In an SIMD architecture, each processor has to fetch the updated off-processor vector elements while computing its share of the product. In this paper, we report on run-time optimization of array distribution and offprocessor data fetching to reduce both the communication and computation time. The optimization is applied to a sparse matrix stored in a compressed sparse row-wise format. Actual runs on test matrices produced up to a 35 percent relative improvement over a block distribution with a naive multiplication algorithm while simulations over a wider range of processors indicate that up to a 60 percent improvement may be possible in some cases.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Preprocessing and Data Distribution Of Sparse Matricesmentioning

confidence: 99%

Run-time optimization of sparse matrix-vector multiplication on SIMD machines

Ziantz

Özturan

Szymański

1994

PARLE'94 Parallel Architectures and Languages Europe

View full text Add to dashboard Cite

show abstract

“…We applied ACC to the SOR (Successive Over-Relaxation) routine found in ITPACK 2C [11], a package of Fortran subroutines for solving linear systems by adaptive iterative methods. Like many iterative solvers, the convergence rate of SOR is heavily influenced by a parameter, in this case the Figure 2: Adaptive SOR system through interception of ITSOR calls using ACC over-relaxation parameter ω. ITPACK includes an internal algorithm for automatically adapting ω.…”

Section: Case Study 1: Adaptive Sormentioning

confidence: 99%

Adaptive Code Collage: A Framework to Transparently Modify Scientific Codes

Kang

Ramakrishnan²,

Ribbens

et al. 2012

Comput. Sci. Eng.

View full text Add to dashboard Cite

Legacy scientific codes are often re-purposed to fit adaptive needs, such as to dynamically alter parameters to improve convergence behavior, or to switch algorithms at runtime for greater accuracy of modeling. Given a legacy scientific code, how can we make it adaptive without making changes to the original source program(s)? We present an approach-Adaptive Code Collage (ACC)-to achieve this goal using function call interception in a language-neutral way at link time. ACC transparently 'catches' function calls and redirects them so that an existing program can be made adaptive without causing a significant performance overhead. We demonstrate the appliction of ACC to designing adaptive SOR algorithms for solving linear systems and to improving the performance, stability, and accuracy of GenIDLEST, a large parallel computational fluid dynamics code.

show abstract

“…The reduced-system conjugate-gradientmethod provides an acceieratediterative solution that is 2-55 = WHC-EP-0445 well suited to sparse matrixes. In PORMC, the reduced-systemconjugategradient algorithm developed by Kincaid et al (1982) is used. Application of this method requires that the algebraicequations be reordered into a "red-black"system (Hageman and Young 1981).…”

Section: 924mentioning

confidence: 99%

PORMC: A model for Monte Carlo simulation of fluid flow, heat, and mass transport in variably saturated geologic media

1991

View full text Add to dashboard Cite

Algorithm 586: ITPACK 2C: A FORTRAN Package for Solving Large Sparse Linear Systems by Adaptive Accelerated Iterative Methods

Cited by 134 publications

References 4 publications

Run-time optimization of sparse matrix-vector multiplication on SIMD machines

Run-time optimization of sparse matrix-vector multiplication on SIMD machines

Adaptive Code Collage: A Framework to Transparently Modify Scientific Codes

PORMC: A model for Monte Carlo simulation of fluid flow, heat, and mass transport in variably saturated geologic media

Contact Info

Product

Resources

About