TY - GEN
T1 - A fully run-time auto-tuned sparse iterative solver with OpenATLib
AU - Naono, Ken
AU - Sakurai, Takao
AU - Igai, Mitsuyoshi
AU - Katagiri, Takahiro
AU - Ohshima, Satoshi
AU - Itoh, Shoji
AU - Nakajima, Kengo
AU - Kuroda, Hisayasu
N1 - Copyright:
Copyright 2012 Elsevier B.V., All rights reserved.
PY - 2012
Y1 - 2012
N2 - We propose a general application programming interface called OpenATLib for auto-tuning (AT). OpenATLib is carefully designed to establish the reusability of AT functions for sparse iterative solvers. Using APIs of OpenATLib, we develop a fully auto-tuned sparse iterative solver called Xabclib. Xabclib has several novel runtime AT functions. We also develop a numerical computation policy that can optimize memory space and computational accuracy. Using the above functions and policies, we obtain the following important findings: (1) an average memory space is reduced to 1/45 under lower memory policies, and (2) fault convergence, which the conventional solvers judges to be converged but actually not converged in the sense of the before-preconditioned matrix, is avoided under higher accuracy policies. The results imply policy-based runtime AT plays significant role in sparse iterative matrix computations.
AB - We propose a general application programming interface called OpenATLib for auto-tuning (AT). OpenATLib is carefully designed to establish the reusability of AT functions for sparse iterative solvers. Using APIs of OpenATLib, we develop a fully auto-tuned sparse iterative solver called Xabclib. Xabclib has several novel runtime AT functions. We also develop a numerical computation policy that can optimize memory space and computational accuracy. Using the above functions and policies, we obtain the following important findings: (1) an average memory space is reduced to 1/45 under lower memory policies, and (2) fault convergence, which the conventional solvers judges to be converged but actually not converged in the sense of the before-preconditioned matrix, is avoided under higher accuracy policies. The results imply policy-based runtime AT plays significant role in sparse iterative matrix computations.
UR - http://www.scopus.com/inward/record.url?scp=84867950606&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84867950606&partnerID=8YFLogxK
U2 - 10.1109/ICIAS.2012.6306176
DO - 10.1109/ICIAS.2012.6306176
M3 - Conference contribution
AN - SCOPUS:84867950606
SN - 9781457719677
T3 - ICIAS 2012 - 2012 4th International Conference on Intelligent and Advanced Systems: A Conference of World Engineering, Science and Technology Congress (ESTCON) - Conference Proceedings
SP - 143
EP - 148
BT - ICIAS 2012 - 2012 4th International Conference on Intelligent and Advanced Systems
T2 - 2012 4th International Conference on Intelligent and Advanced Systems, ICIAS 2012
Y2 - 12 June 2012 through 14 June 2012
ER -