Demystifying GPU microarchitecture through microbenchmarking

Author

Wong, Henry ; Papadopoulou, Misel-Myrto ; Sadooghi-Alvandi, Maryam ; Moshovos, Andreas

Author_Institution

Dept. of Electr. & Comput. Eng., Univ. of Toronto, Toronto, ON, Canada

fYear

2010

fDate

28-30 March 2010

Firstpage

235

Lastpage

246

Abstract

Graphics processors (GPU) offer the promise of more than an order of magnitude speedup over conventional processors for certain non-graphics computations. Because the GPU is often presented as a C-like abstraction (e.g., Nvidia´s CUDA), little is known about the characteristics of the GPU´s architecture beyond what the manufacturer has documented. This work develops a microbechmark suite and measures the CUDA-visible architectural characteristics of the Nvidia GT200 (GTX280) GPU. Various undisclosed characteristics of the processing elements and the memory hierarchies are measured. This analysis exposes undocumented features that impact program performance and correctness. These measurements can be useful for improving performance optimization, analysis, and modeling on this architecture and offer additional insight on the decisions made in developing this GPU.

Keywords

computer graphics; coprocessors; GPU microarchitecture; Nvidia GT200 GPU; graphics processors; microbenchmarking; Clocks; Computer architecture; Delay; Hardware; Kernel; Microarchitecture; Performance analysis; Registers; Samarium; Yarn;

fLanguage

English

Publisher

ieee

Conference_Titel

Performance Analysis of Systems & Software (ISPASS), 2010 IEEE International Symposium on

Conference_Location

White Plains, NY

Print_ISBN

978-1-4244-6023-6

Electronic_ISBN

978-1-4244-6024-3

Type

conf

DOI

10.1109/ISPASS.2010.5452013

Filename

5452013