On GPU acceleration of common solvers for (quasi-) triangular generalized Lyapunov equations
详细信息 查看全文
- 作者：Martin Kö ; hler ; koehlerm@mpi-magdeburg.mpg.de" class="auth_mail" title="E-mail the corresponding author ; Jens Saak saak@mpi-magdeburg.mpg.de" class="auth_mail" title="E-mail the corresponding author
- 关键词：Lyapunov equations ; BLAS level-3 ; Accelerator device
- 刊名：Parallel Computing
- 出版年：2016
- 出版时间：September 2016
- 年：2016
- 卷：57
- 期：Complete
- 页码：212-221
- 全文大小：535 K
文摘
- •
We develop a GPU accelerated variant of the blocked Bartels–Stewart algorithm for the generalized Lyapunov equation.
- •
We analyze the influence of optimized data alignment on the GPU and discuss the mathematical problem behind it.
- •
Reordering eigenvalues avoids alignment problems on the GPU and accelerates the computation.