Cache Flow

Thursday, 3 August
3:45 - 5:30 pm
Room 253
Session Chair: Daniel Wexler, NVIDIA Corporation

A Scalable and Reconfigurable Shared-Memory Graphics Architecture
A scalable, tightly coupled cluster of custom-built boards that provide an AGP interface for commodity graphics accelerators.

Michael Manzke
Ross Brennan
Keith O'Conor
John Dingliana
Carol O'Sullivan
Trinity College Dublin
michael.manzke (at)
  Silhouette Management for Protruded Displacement Mapping
A new silhouette management technique using tangent space transformation represents the correct silhouette of protruded shapes by solving the overflow problem that frequently occurs in image-based displacement mapping.

Byounghyun Yoo
Soonhung Han
Korea Advanced Institute of Science and Technology
yoo (at)
A Performance-Oriented Data-Parallel Virtual Machine for GPUs
A new virtual machine abstraction for GPUs that provides policy-free, low-level access to the hardware and is designed for high-performance, data-parallel applications.

Mark Peercy
Mark Segal
Derek Gerstmann
ATI Research Inc.
dgerstma (at)
The Real-Time Reprojection Cache
A method that uses graphics hardware to cache and track surface information from previously rendered frames. This information is then used to significantly improve rendering cost and/or quality.

Diego Nehab
Princeton University

Pedro V. Sander
John Isidoro
ATI Research Inc.
pedro (at)