Advanced

Auto-tuning Interactive Ray Tracing using an Analytical GPU Architecture Model

Ganestam, Per LU and Doggett, Michael LU (2012) GPGPU5 In The ACM International Conference Proceedings Series
Abstract
This paper presents a method for auto-tuning interactive ray tracing on GPUs using a hardware model. Getting full performance from modern GPUs is a challenging task. Workloads which require a guaranteed performance over several runs must select parameters for the worst performance of all runs. Our method uses an analyti- cal GPU performance model to predict the current frame’s render- ing time using a selected set of parameters. These parameters are then optimised for a selected frame rate performance on the partic- ular GPU architecture. We use auto-tuning to determine parameters such as phong shading, shadow rays and the number of ambient oc- clusion rays. We sample a priori information about the current ren- dering load to estimate the... (More)
This paper presents a method for auto-tuning interactive ray tracing on GPUs using a hardware model. Getting full performance from modern GPUs is a challenging task. Workloads which require a guaranteed performance over several runs must select parameters for the worst performance of all runs. Our method uses an analyti- cal GPU performance model to predict the current frame’s render- ing time using a selected set of parameters. These parameters are then optimised for a selected frame rate performance on the partic- ular GPU architecture. We use auto-tuning to determine parameters such as phong shading, shadow rays and the number of ambient oc- clusion rays. We sample a priori information about the current ren- dering load to estimate the frame workload. A GPU model is run iteratively using this information to tune rendering parameters for a target frame rate. We use the OpenCL API allowing tuning across different GPU architectures. Our auto-tuning enables the render- ing of each frame to execute in a predicted time, so a target frame rate can be achieved even with widely varying scene complexities. Using this method we can select optimal parameters for the cur- rent execution taking into account the current viewpoint and scene, achieving performance improvements over predetermined parame- ters. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
GPU Model, Ray Tracing, Auto-tuning, OpenCL
in
The ACM International Conference Proceedings Series
pages
7 pages
conference name
GPGPU5
external identifiers
  • Scopus:84858791144
language
English
LU publication?
yes
id
d0ec480a-349c-47a8-b893-adfb01879f1e (old id 2374593)
date added to LUP
2012-03-21 10:57:14
date last changed
2016-10-13 04:52:55
@misc{d0ec480a-349c-47a8-b893-adfb01879f1e,
  abstract     = {This paper presents a method for auto-tuning interactive ray tracing on GPUs using a hardware model. Getting full performance from modern GPUs is a challenging task. Workloads which require a guaranteed performance over several runs must select parameters for the worst performance of all runs. Our method uses an analyti- cal GPU performance model to predict the current frame’s render- ing time using a selected set of parameters. These parameters are then optimised for a selected frame rate performance on the partic- ular GPU architecture. We use auto-tuning to determine parameters such as phong shading, shadow rays and the number of ambient oc- clusion rays. We sample a priori information about the current ren- dering load to estimate the frame workload. A GPU model is run iteratively using this information to tune rendering parameters for a target frame rate. We use the OpenCL API allowing tuning across different GPU architectures. Our auto-tuning enables the render- ing of each frame to execute in a predicted time, so a target frame rate can be achieved even with widely varying scene complexities. Using this method we can select optimal parameters for the cur- rent execution taking into account the current viewpoint and scene, achieving performance improvements over predetermined parame- ters.},
  author       = {Ganestam, Per and Doggett, Michael},
  keyword      = {GPU Model,Ray Tracing,Auto-tuning,OpenCL},
  language     = {eng},
  pages        = {7},
  series       = {The ACM International Conference Proceedings Series},
  title        = {Auto-tuning Interactive Ray Tracing using an Analytical GPU Architecture Model},
  year         = {2012},
}