All issues
Volume 6 (2014)
- Issue 4 Pages 43-
- Issue 3 Pages 37-
- Issue 2 Pages 15-
- Issue 1 Pages 1-
Volume 6, Issue 4
Displaying 1-2 of 2 articles from this issue
- |<
- <
- 1
- >
- >|
Paper
-
Chiaki Itoh, Yoshiyuki Higashi, Seiji Fujino2014 Volume 6 Issue 4 Pages 43-52
Published: 2014
Released on J-STAGE: November 29, 2014
JOURNAL FREE ACCESSWe consider parallelism of an extended IDR (s) method based on the secant method by means of the proposed Cache-Cache Balance technique. In this paper, we describe concept and outline of the Cache-Cache Balance technique. Through many numerical experiments, we will make clear effectiveness of the proposed Cache-Cache Balance implementation for parallelism.
View full abstractDownload PDF (1535K) -
George Abe, Yusuke Onoue, Seiji Fujino2014 Volume 6 Issue 4 Pages 53-59
Published: 2014
Released on J-STAGE: November 29, 2014
JOURNAL FREE ACCESSThis paper compares the performance of sparse Matrix-vector multiplication paralleled by the conventional Block-Cyclic distribution and its improved variant on parallel computer with shared memory. The underlying idea is to exchange nonzero entries of matrix assigned to each thread with block unit. Numerical results demonstrate that the proposed distribution using exchange nonzero entries of matrix with block unit gives or improves parallelism.
View full abstractDownload PDF (1506K)
- |<
- <
- 1
- >
- >|