### Abstract

Cumulative reaction probability (CRP) calculations provide a viable computational approach to estimate reaction rate coefficients. However, in order to give meaningful results these calculations should be done in many dimensions (ten to fifteen). This makes CRP codes memory intensive. For this reason, these codes use iterative methods to solve the linear systems, where a good fraction of the execution time is spent on matrix-vector multiplication. In this paper, we discuss the tensor product form of applying the system operator on a vector. This approach shows much better performance and provides huge savings in memory as compared to the explicit sparse representation of the system matrix.

Original language | English |
---|---|

Title of host publication | High Performance Computing - HiPC 2008 - 15th International Conference, Proceedings |

Publisher | Springer Verlag |

Pages | 120-130 |

Number of pages | 11 |

Volume | 5374 LNCS |

ISBN (Print) | 354089893X, 9783540898931 |

DOIs | |

Publication status | Published - 2008 |

Externally published | Yes |

Event | 15th International Conference on High Performance Computing, HiPC 2008 - Bangalore, India Duration: 17 Dec 2008 → 20 Dec 2008 |

### Publication series

Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|

Volume | 5374 LNCS |

ISSN (Print) | 03029743 |

ISSN (Electronic) | 16113349 |

### Other

Other | 15th International Conference on High Performance Computing, HiPC 2008 |
---|---|

Country | India |

City | Bangalore |

Period | 17/12/08 → 20/12/08 |

### Fingerprint

### ASJC Scopus subject areas

- Theoretical Computer Science
- Computer Science(all)

### Cite this

*High Performance Computing - HiPC 2008 - 15th International Conference, Proceedings*(Vol. 5374 LNCS, pp. 120-130). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5374 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-540-89894-8_14

**Improving the performance of tensor matrix vector multiplication in cumulative reaction probability based quantum chemistry codes.** / Kaushik, Dinesh; Gropp, William; Minkoff, Michael; Smith, Barry.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

*High Performance Computing - HiPC 2008 - 15th International Conference, Proceedings.*vol. 5374 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5374 LNCS, Springer Verlag, pp. 120-130, 15th International Conference on High Performance Computing, HiPC 2008, Bangalore, India, 17/12/08. https://doi.org/10.1007/978-3-540-89894-8_14

}

TY - GEN

T1 - Improving the performance of tensor matrix vector multiplication in cumulative reaction probability based quantum chemistry codes

AU - Kaushik, Dinesh

AU - Gropp, William

AU - Minkoff, Michael

AU - Smith, Barry

PY - 2008

Y1 - 2008

N2 - Cumulative reaction probability (CRP) calculations provide a viable computational approach to estimate reaction rate coefficients. However, in order to give meaningful results these calculations should be done in many dimensions (ten to fifteen). This makes CRP codes memory intensive. For this reason, these codes use iterative methods to solve the linear systems, where a good fraction of the execution time is spent on matrix-vector multiplication. In this paper, we discuss the tensor product form of applying the system operator on a vector. This approach shows much better performance and provides huge savings in memory as compared to the explicit sparse representation of the system matrix.

AB - Cumulative reaction probability (CRP) calculations provide a viable computational approach to estimate reaction rate coefficients. However, in order to give meaningful results these calculations should be done in many dimensions (ten to fifteen). This makes CRP codes memory intensive. For this reason, these codes use iterative methods to solve the linear systems, where a good fraction of the execution time is spent on matrix-vector multiplication. In this paper, we discuss the tensor product form of applying the system operator on a vector. This approach shows much better performance and provides huge savings in memory as compared to the explicit sparse representation of the system matrix.

UR - http://www.scopus.com/inward/record.url?scp=58449097645&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=58449097645&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-89894-8_14

DO - 10.1007/978-3-540-89894-8_14

M3 - Conference contribution

AN - SCOPUS:58449097645

SN - 354089893X

SN - 9783540898931

VL - 5374 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 120

EP - 130

BT - High Performance Computing - HiPC 2008 - 15th International Conference, Proceedings

PB - Springer Verlag

ER -