Throughput-oriented analytical models for performance estimation on programmable hardware accelerators

Throughput-oriented analytical models for performance estimation on programmable hardware accelerators

Show other versions (1)

In this thesis work, we have mainly worked on two topics of GPU performance analysis. First, we have developed an analytical method and a timing estimation tool (TEG) to predict CUDA application's performance for GT200 generation GPUs. TEG can predict GPU applications' performance in cycle...

Full description

Bibliographic Details
Main Author: Lai, Junjie
Language:ENG
Published: Université Rennes 1 2013
Subjects:
[INFO:INFO_OH] Computer Science/Other
GPGPU
CUDA
Fermi GPU
Kepler GPU
Performance upper bound
Performance Prediction
Performance Analysis
Online Access:http://tel.archives-ouvertes.fr/tel-00854019
http://tel.archives-ouvertes.fr/docs/00/85/40/19/PDF/LAI_Junjie.pdf
Description
Summary:In this thesis work, we have mainly worked on two topics of GPU performance analysis. First, we have developed an analytical method and a timing estimation tool (TEG) to predict CUDA application's performance for GT200 generation GPUs. TEG can predict GPU applications' performance in cycle-approximate level. Second, we have developed an approach to estimate GPU applications' performance upper bound based on application analysis and assembly code level benchmarking. With the performance upper bound of an application, we know how much optimization space is left and can decide the optimization effort. Also with the analysis we can understand which parameters are critical to the performance.

Similar Items