Swizzle Inventor: Data Movement Synthesis for GPU Kernels
Swizzle Inventor: Data Movement Synthesis for GPU Kernels
复制标题
Swizzle Inventor:GPU 内核的数据移动综合
DOI:
10.1145/3297858.3304059
复制
发表时间:
2019
期刊:
影响因子:
--
通讯作者:
Torlak, Emina
中科院分区:
文献类型:
--
作者:
Phothilimthana, Phitchaya Mangpo;Bodik, Rastislav;Elliott, Archibald Samuel;Wang, An;Jangda, Abhinav;Hagedorn, Bastian;Barthels, Henrik;Kaufman, Samuel J.;Grover, Vinod;Torlak, Emina
Utilizing memory and register bandwidth in modern architectures may require swizzles --- non-trivial mappings of data and computations onto hardware resources --- such as shuffles. We develop Swizzle Inventor to help programmers implement swizzle programs, by writing program sketches that omit swizzles and delegating their creation to an automatic synthesizer. Our synthesis algorithm scales to real-world programs, allowing us to invent new GPU kernels for stencil computations, matrix transposition, and a finite field multiplication algorithm (used in cryptographic applications). The synthesized 2D convolution and finite field multiplication kernels are on average 1.5--3.2x and 1.1--1.7x faster, respectively, than expert-optimized CUDA kernels.
登录
查看更多内容
DOI:
10.1145/305138.305197
发表时间:
1999-06
期刊:
--
影响因子:
--
作者:
Amy W. Lim;Gerald I. Cheong;M. Lam
通讯作者:
Amy W. Lim;Gerald I. Cheong;M. Lam
DOI:
--
发表时间:
2017
期刊:
International Carnahan Conference on Security Technology
影响因子:
--
作者:
W. Lee;Xian;B. Goi;R. Phan
通讯作者:
R. Phan
DOI:
10.1109/ipdps.2017.79
发表时间:
2017
期刊:
2017 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
影响因子:
--
作者:
Jie Wang;Xinfeng Xie;J. Cong
通讯作者:
J. Cong
DOI:
--
发表时间:
2016
期刊:
International Conference on Supercomputing
影响因子:
--
作者:
Eli Ben;Matan Hamilis;M. Silberstein;Eran Tromer
通讯作者:
Eran Tromer
DOI:
--
发表时间:
2014
期刊:
International Conference for High Performance Computing, Networking, Storage and Analysis
影响因子:
--
作者:
Zhilei Xu;Shoaib Kamil;Armando Solar
通讯作者:
Armando Solar