Software Open Access

Intersection Prediction for Accelerated GPU Ray Tracing

Liu, Lufei; Chang, Wesley; Demoullin, Francois; Chou, Yuan Hsi; Saed, Mohammadreza; Pankratz, David; Nowicki, Tyler; Aamodt, Tor M.

Dublin Core Export

<?xml version='1.0' encoding='utf-8'?>
<oai_dc:dc xmlns:dc="" xmlns:oai_dc="" xmlns:xsi="" xsi:schemaLocation="">
  <dc:creator>Liu, Lufei</dc:creator>
  <dc:creator>Chang, Wesley</dc:creator>
  <dc:creator>Demoullin, Francois</dc:creator>
  <dc:creator>Chou, Yuan Hsi</dc:creator>
  <dc:creator>Saed, Mohammadreza</dc:creator>
  <dc:creator>Pankratz, David</dc:creator>
  <dc:creator>Nowicki, Tyler</dc:creator>
  <dc:creator>Aamodt, Tor M.</dc:creator>
  <dc:description>This artifacts contains a Docker image, in which the simulator used in our MICRO 2021 conference paper is installed. The system also includes a ray tracing application and scripts to generate results to match the paper. 

Ray tracing techniques can generate highly realistic images by traversing rays through hierarchical acceleration structures and testing for intersections. Raster-based shading techniques have historically been preferred for real-time graphics due to the computational intensity of ray tracing. However, recent Graphics Processing Units (GPUs) now incorporate hardware support for ray tracing, accelerating the ray traversal and intersection process. We observe that rays which follow similar paths execute many redundant ray-box intersection tests in their traversals. We propose a ray intersection predictor to speculatively elide redundant operations in traversals and proceed directly to necessary intersection tests.</dc:description>
  <dc:description>This version uses L1 data cache instead of specialized RT cache</dc:description>
  <dc:title>Intersection Prediction for Accelerated GPU Ray Tracing</dc:title>
All versions This version
Views 358149
Downloads 11274
Data volume 353.1 GB235.3 GB
Unique views 290137
Unique downloads 8352


Cite as