Memory Considerations for Low Energy Ray Tracing
Daniel Kopta, Konstantin Shkurko, Josef Spjut, Erik Brunvand, and Al Davis
In Computer Graphics Forum (CGF), 2014
Abstract
We propose two hardware mechanisms to decrease energy consumption on massively parallel graphics processors for ray tracing. First, we use a streaming data model and configure part of the L2 cache into a ray stream memory to enable efficient data processing through ray reordering. This increases L1 hit rates and reduces off-chip memory energy substantially through better management of off-chip memory access patterns. To evaluate this model, we augment our architectural simulator with a detailed memory system simulation that includes accurate control, timing and power models for memory controllers and off-chip dynamic random-access memory. These details change the results significantly over previous simulations that used a simpler model of off-chip memory, indicating that this type of memory system simulation is important for realistic simulations that involve external memory. Secondly, we employ reconfigurable special-purpose pipelines that are constructed dynamically under program control. These pipelines use shared execution units that can be configured to support the common compute kernels that are the foundation of the ray tracing algorithm. This reduces the overhead incurred by on-chip memory and register accesses. These two synergistic features yield a ray tracing architecture that reduces energy by optimizing both on-chip and off-chip memory activity when compared to a more traditional approach.
BibTeX
@article{CGF14_koptaMemConsLowEnergyRT,
author = {Daniel Kopta and Konstantin Shkurko and Josef Spjut and Erik Brunvand and Al Davis},
journal = {Computer Graphics Forum},
title = {{Memory Considerations for Low Energy Ray Tracing}},
year = {2015},
volume = {34},
number = {1},
pages = {47--59},
issn = {1467-8659},
url = {http://dx.doi.org/10.1111/cgf.12458},
doi = {10.1111/cgf.12458}
}Acknowledgements
This material is based upon work supported by the National Science Foundation under Grant No. CNS-1017457. Manjunath Shevgoor, Niladrish Chatterjee and Tim Purcell provided many useful discussions. Sibenik Cathedral is from Marko Dabrovic, Fairy Forest is from the University of Utah, Crytek Sponza is from Frank Meinl at Crytek and Marko Dabrovic, Conference is from Anat Grynberg and Greg Ward, Dragon and Buddha are from the Stanford Computer Graphics Laboratory, Vegetation and Hairball are from Samuli Laine and San Miguel is from Guillermo Leal Laguno.