Stream compaction for deferred shading

Jared Hoberock, Victor Lu, Yuntao Jia, John C Hart

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The GPU leverages SIMD efficiency when shading because it rasterizes a triangle at a time, running the same shader on all of its fragments. Ray tracing sacrifices this shader coherence, and the result is that SIMD units often must run different shaders simultaneously resulting in serialization. We study this problem and define a new measure called heterogeneous efficiency to measure SIMD divergence among multiple shaders of different complexities in a ray tracing application. We devise seven different algorithms for scheduling shaders onto SIMD processors to avoid divergence. In all but simply shaded scenes, we show the expense of sortingshaders pays off with better overall shading performance.

Original languageEnglish (US)
Title of host publicationProceedings - High-Performance Graphics 2009
Pages173-180
Number of pages8
StatePublished - Dec 1 2009
EventHigh-Performance Graphics 2009 - New Orleans, LA, United States
Duration: Aug 1 2009Aug 3 2009

Publication series

NameProceedings of the SIGGRAPH/Eurographics Workshop on Graphics Hardware

Other

OtherHigh-Performance Graphics 2009
CountryUnited States
CityNew Orleans, LA
Period8/1/098/3/09

Fingerprint

Ray tracing
Compaction
Scheduling
Graphics processing unit

ASJC Scopus subject areas

  • Hardware and Architecture
  • Computer Science Applications

Cite this

Hoberock, J., Lu, V., Jia, Y., & Hart, J. C. (2009). Stream compaction for deferred shading. In Proceedings - High-Performance Graphics 2009 (pp. 173-180). (Proceedings of the SIGGRAPH/Eurographics Workshop on Graphics Hardware).

Stream compaction for deferred shading. / Hoberock, Jared; Lu, Victor; Jia, Yuntao; Hart, John C.

Proceedings - High-Performance Graphics 2009. 2009. p. 173-180 (Proceedings of the SIGGRAPH/Eurographics Workshop on Graphics Hardware).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hoberock, J, Lu, V, Jia, Y & Hart, JC 2009, Stream compaction for deferred shading. in Proceedings - High-Performance Graphics 2009. Proceedings of the SIGGRAPH/Eurographics Workshop on Graphics Hardware, pp. 173-180, High-Performance Graphics 2009, New Orleans, LA, United States, 8/1/09.
Hoberock J, Lu V, Jia Y, Hart JC. Stream compaction for deferred shading. In Proceedings - High-Performance Graphics 2009. 2009. p. 173-180. (Proceedings of the SIGGRAPH/Eurographics Workshop on Graphics Hardware).
Hoberock, Jared ; Lu, Victor ; Jia, Yuntao ; Hart, John C. / Stream compaction for deferred shading. Proceedings - High-Performance Graphics 2009. 2009. pp. 173-180 (Proceedings of the SIGGRAPH/Eurographics Workshop on Graphics Hardware).
@inproceedings{8209e3ad357549a4aef01c58b11a04fd,
title = "Stream compaction for deferred shading",
abstract = "The GPU leverages SIMD efficiency when shading because it rasterizes a triangle at a time, running the same shader on all of its fragments. Ray tracing sacrifices this shader coherence, and the result is that SIMD units often must run different shaders simultaneously resulting in serialization. We study this problem and define a new measure called heterogeneous efficiency to measure SIMD divergence among multiple shaders of different complexities in a ray tracing application. We devise seven different algorithms for scheduling shaders onto SIMD processors to avoid divergence. In all but simply shaded scenes, we show the expense of sortingshaders pays off with better overall shading performance.",
author = "Jared Hoberock and Victor Lu and Yuntao Jia and Hart, {John C}",
year = "2009",
month = "12",
day = "1",
language = "English (US)",
isbn = "9781605586038",
series = "Proceedings of the SIGGRAPH/Eurographics Workshop on Graphics Hardware",
pages = "173--180",
booktitle = "Proceedings - High-Performance Graphics 2009",

}

TY - GEN

T1 - Stream compaction for deferred shading

AU - Hoberock, Jared

AU - Lu, Victor

AU - Jia, Yuntao

AU - Hart, John C

PY - 2009/12/1

Y1 - 2009/12/1

N2 - The GPU leverages SIMD efficiency when shading because it rasterizes a triangle at a time, running the same shader on all of its fragments. Ray tracing sacrifices this shader coherence, and the result is that SIMD units often must run different shaders simultaneously resulting in serialization. We study this problem and define a new measure called heterogeneous efficiency to measure SIMD divergence among multiple shaders of different complexities in a ray tracing application. We devise seven different algorithms for scheduling shaders onto SIMD processors to avoid divergence. In all but simply shaded scenes, we show the expense of sortingshaders pays off with better overall shading performance.

AB - The GPU leverages SIMD efficiency when shading because it rasterizes a triangle at a time, running the same shader on all of its fragments. Ray tracing sacrifices this shader coherence, and the result is that SIMD units often must run different shaders simultaneously resulting in serialization. We study this problem and define a new measure called heterogeneous efficiency to measure SIMD divergence among multiple shaders of different complexities in a ray tracing application. We devise seven different algorithms for scheduling shaders onto SIMD processors to avoid divergence. In all but simply shaded scenes, we show the expense of sortingshaders pays off with better overall shading performance.

UR - http://www.scopus.com/inward/record.url?scp=78651337597&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78651337597&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:78651337597

SN - 9781605586038

T3 - Proceedings of the SIGGRAPH/Eurographics Workshop on Graphics Hardware

SP - 173

EP - 180

BT - Proceedings - High-Performance Graphics 2009

ER -