Low-Order Finite Element Solver with Small Matrix-Matrix Multiplication Accelerated by AI-Specific Hardware for Crustal Deformation Computation
Low-Order Finite Element Solver with Small Matrix-Matrix Multiplication Accelerated by AI-Specific Hardware for Crustal Deformation Computation
复制标题
具有小矩阵-矩阵乘法的低阶有限元求解器,由 AI 专用硬件加速,用于地壳变形计算
DOI:
10.1145/3394277.3401860
复制
发表时间:
2020
期刊:
影响因子:
--
通讯作者:
Ueda Naonori
中科院分区:
文献类型:
--
作者:
Yamaguchi Takuma;Fujita Kohei;Ichimura Tsuyoshi;Naruse Akira;Wells Jack C.;Zimmer Christopher J.;Straatsma Tjerk P.;Hori Muneo;Maddegedara Lalith;Ueda Naonori
This study proposes a fast low-order finite element solver for crustal deformation computations by applying Tensor Core, AI-specific hardware on a Volta GPU. Tensor Core can compute large matrix-matrix multiplications rapidly in half precision. We redesign a state-of-the-art solver algorithm so that lower-precision data types can be used and memory access costs can be reduced even when we use small matrices. With the proposed solver, we solved 13 billion degrees-of-freedom two-layered problems that mimicked the Earth's crust and mantle using 36 compute nodes of Summit. In the matrix-vector kernel, we obtained a 4.1-fold speedup over a standard kernel in a single-precision format. Our proposed solver increased the FLOP count of the entire solver; however, we reduced the time-to-solution by 1.7-fold since the Tensor Core provided a high effective performance.
登录
查看更多内容
影响因子:
3
作者:
T. Ichimura;M. Hori;Hiroyuki Kuwamoto
通讯作者:
T. Ichimura;M. Hori;Hiroyuki Kuwamoto
DOI:
10.1137/1.9780898718003
发表时间:
2003-05
期刊:
--
影响因子:
--
作者:
Y. Saad
通讯作者:
Y. Saad
DOI:
--
发表时间:
2008
期刊:
影响因子:
--
作者:
J. Parker;G. Lyzenga;C. Norton;C. Zuffada;M. Glasscoe;J. Lou;A. Donnellan
通讯作者:
A. Donnellan
DOI:
10.1109/sc.2018.00052
发表时间:
2018-11
期刊:
SC18: International Conference for High Performance Computing, Networking, Storage and Analysis
影响因子:
--
作者:
T. Ichimura;K. Fujita;Takuma Yamaguchi;Akira Naruse;J. Wells;T. Schulthess;T. Straatsma;Christopher Z
通讯作者:
T. Ichimura;K. Fujita;Takuma Yamaguchi;Akira Naruse;J. Wells;T. Schulthess;T. Straatsma;Christopher Z
影响因子:
64.8
作者:
Aage, Niels;Andreassen, Erik;Sigmund, Ole
通讯作者:
Sigmund, Ole