Break Down GPU Execution Time with an Analytical Method - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2012

Break Down GPU Execution Time with an Analytical Method

Junjie Lai
  • Function : Author
  • PersonId : 913983
André Seznec


Because modern GPGPU can provide significant computing power and has very high memory bandwidth, and also, developer-friendly programming interfaces such as CUDA have been introduced, GPGPU becomes more and more accepted in the HPC research area. Much research has been done to help developers to better optimize GPU applications. But to fully understand GPU performance behavior remains a hot research topic. We developed an analytical tool called TEG (Timing Estimation tool for GPU) to estimate GPU performance. Previ- ous work shows that TEG has good approximation and can help us to quantify bottlenecks' performance effects. We have made some improvement to the tool and in this paper, we use TEG to analyze the GPU performance scaling be- havior. TEG takes the dis-assembly output of CUDA kernel binary code and instruction trace as input. It does not ex- ecute the codes, but try to model the execution of CUDA codes with timing information. Because TEG takes the native GPU assembly code as input, it can estimate the execution time with a small error and it allows us to get more insight into GPU performance result.
No file

Dates and versions

hal-00764874 , version 1 (13-12-2012)



Junjie Lai, André Seznec. Break Down GPU Execution Time with an Analytical Method. Rapido '12, Jan 2012, Paris, France. ⟨10.1145/2162131.2162136⟩. ⟨hal-00764874⟩
265 View
0 Download



Gmail Facebook X LinkedIn More