SAH KD-tree construction on GPU

Wu, Zhefeng; Zhao, Fukai; Liu, Xinguo

doi:10.1145/2018323.2018335

Cited by 36 publications

(26 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The comparison of construction time is reported in Table 1. By Table 1, the conclusion is that our algorithm is slightly faster than algorithm of literature [7] on the small testing scenes, such like Bunny, but significantly faster on the larger testing scenes, such like the Happy Buddha and the Dragon. Figure 10 can show this contrast intuitively.…”

Section: Experimental Results and Analysesmentioning

confidence: 83%

“…The experimental results were compared with the SAH kd-tree construction on GPU proposed by Wu et al [7]. The comparison of construction time is reported in Table 1.…”

Section: Experimental Results and Analysesmentioning

confidence: 99%

“…In this way, a mass of computation are avoided. However, if using ray tracing which is based on spatial median splitting only, the quality of the kd-tree it generates is too poor to reach a high efficiency rendering result [7], especially for the scene with highly random distributed primitives.…”

Section: The Present and Problems Of Kd-treementioning

confidence: 99%

“…The algorithm aimed at solving the difference parallelism between large nodes and small nodes, and was proved that it outperforms previous SAH kd-tree construction [7]. Unfortunately the algorithm didn't provide methods for the case of a large space scene, the settled strategy of the space splitting may cause the quality degrading of kd-tree.…”

Section: The Present and Problems Of Kd-treementioning

confidence: 99%

See 3 more Smart Citations

Research on Parallel KD-Tree Construction for Ray Tracing

Zhang¹,

Xu²,

Minjie³

et al. 2016

IJGDC

View full text Add to dashboard Cite

Many computer graphics rendering algorithms and techniques use ray tracing for generation of photo-realistic images, and kd-tree is the most popular acceleration data structure for ray tracing. In order to avoid the inefficient parallel performance of kd-tree construction based on surface area heuristic (SAH), an algorithm using Morton code and path compression was present in this paper. Instead of building a kd-tree layer-by-layer, the proposed approach can be performed in parallel from bottom of a conceptual perfect binary tree. Experimental results on GPU show that our work achieves a faster kd-tree construction procedure.

show abstract

Section: Experimental Results and Analysesmentioning

confidence: 83%

“…The experimental results were compared with the SAH kd-tree construction on GPU proposed by Wu et al [7]. The comparison of construction time is reported in Table 1.…”

Section: Experimental Results and Analysesmentioning

confidence: 99%

Section: The Present and Problems Of Kd-treementioning

confidence: 99%

Section: The Present and Problems Of Kd-treementioning

confidence: 99%

See 2 more Smart Citations

Research on Parallel KD-Tree Construction for Ray Tracing

Zhang¹,

Xu²,

Minjie³

et al. 2016

IJGDC

View full text Add to dashboard Cite

show abstract

“…PDACRT also performs much better compared to the best parallel CPU k-d tree method, which needs 654ms and 835ms respectively to build the kd tree for Dragon and Buddha models on a 32 core machine [Choi et al 2010]. The best GPU k-d tree construction method also needs 511ms and 645ms respectively for Dragon and Buddha [Wu et al 2011]. PDACRT is 4-5 times faster to construct-and-trace these models.…”

Section: Performancementioning

confidence: 99%

Parallel divide and conquer ray tracing

Ravichandran

Narayanan

2013

SIGGRAPH Asia 2013 Technical Briefs

View full text Add to dashboard Cite

Figure 1: Buddha (1.08M, 315 ms), Dragon (871K, 283 ms) and Fairy (174K, 276ms) models rendered with shadows using our approach. Right: Buddha model with 8 ambient occlusion rays per pixel (410 ms). Resolution 1024 × 1024. AbstractDivide and Conquer Ray Tracing (DACRT) is a recent technique which constructs no explicit acceleration structure. It creates and traverses an implicit hierarchy in a depth-first fashion recursively and is suited for dynamic scenes that change constantly. In this paper, we present a parallel version of DACRT that runs entirely on the GPU, which exploits efficient primitives like sort and reduce. Our approach suits the GPU well, with a low memory footprint. Our implementation outperforms the serial CPU algorithm for both primary and secondary ray passes. We show good performance on primary pass and on advanced effects.

show abstract

DKD: a fast k‐d tree update design for dynamic scenes

Yang

Liu

Zhang

et al. 2016

Computer Animation & Virtual

View full text Add to dashboard Cite

We design dynamic k-d (DKD) tree based on classical k-d tree for animated scene rendering. Our method can inherit the benefit of efficient traversal of k-d tree and minimize time cost to update DKD tree, making it well suited for animated geometry. DKD employs primitive reset, redistribution to reflect the updated positions of geometry, and leaf node incremental growing to avoid the deterioration of hierarchy quality due to refitting. Our experiments show that DKD has a significant rendering performance improvement than selected existing methods.

show abstract

SAH KD-tree construction on GPU

Cited by 36 publications

References 17 publications

Research on Parallel KD-Tree Construction for Ray Tracing

Research on Parallel KD-Tree Construction for Ray Tracing

Parallel divide and conquer ray tracing

DKD: a fast k‐d tree update design for dynamic scenes

Contact Info

Product

Resources

About