Direct send compositing for parallel sort-last rendering

Eilemann, Stefan; Pajarola, Renato

doi:10.1145/1508044.1508083

Cited by 25 publications

(23 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Among the three basic parallel rendering approaches, namely, sort-first, sortmiddle, and sort-last, defined by Molar et al [9], sort-last parallel rendering has been widely used by visualization researchers [5,1,20,19,17,16,12,2,3] due to its simple task decomposition for achieving load balancing. With the sort-last approach, the rendering stage scales very well since no communication overhead is involved.…”

Section: Introductionmentioning

confidence: 99%

Massively parallel volume rendering using 2--3 swap image compositing

Wang

Liu

2008

ACM SIGGRAPH ASIA 2008 Courses on - SIGGRAPH Asia '08

View full text Add to dashboard Cite

The ever-increasing amounts of simulation data produced by scientists demand high-end parallel visualization capability. However, image compositing, which requires interprocessor communication, is often the bottleneck stage for parallel rendering of large volume data sets. Existing image compositing solutions either incur a large number of messages exchanged among processors (such as the direct send method), or limit the number of processors that can be effectively utilized (such as the binary swap method). We introduce a new image compositing algorithm, called 2-3 swap, which combines the flexibility of the direct send method and the optimality of the binary swap method. The 2-3 swap algorithm allows an arbitrary number of processors to be used for compositing, and fully utilizes all participating processors throughout the course of the compositing. We experiment with this image compositing solution on a supercomputer with thousands of processors, and demonstrate its great flexibility as well as scalability.

show abstract

Section: Introductionmentioning

confidence: 99%

Massively parallel volume rendering using 2--3 swap image compositing

Wang

Liu

2008

ACM SIGGRAPH ASIA 2008 Courses on - SIGGRAPH Asia '08

View full text Add to dashboard Cite

show abstract

“…A more efficient method Scatter decomposition [2] distributes tasks randomly (pseudo-randomly) to each node in an attempt to load balance each node [3]. The most straightforward version of this approach is to use the mapping function F (i) = (i mod p).…”

Section: B Static Load Balancing Algorithmmentioning

confidence: 99%

Dynamic Load Balancing Method Based on PBT Tree

Lin¹,

Ren²,

Huang³

et al. 2018

Proceedings of the 2017 2nd International Conference on Electrical, Control and Automation Engineering (ECAE 2017)

View full text Add to dashboard Cite

Abstract-With the development of computer technology and GPU technology, hardware and software technology has been greatly developed, but a single PC machine still cannot meet the real-time computing requirements of high-complexity scenes. In this paper, we propose a dynamic load balancing strategy based on PBT tree for the load balancing problem of Sort-first architecture cluster computing system. The dynamic load balancing strategy is dynamically adjusted in real time according to the change of computing scenarios, Balance caused by the system performance cannot be improved, make full use of system resources and improve the overall performance of the system.

show abstract

“…In addition, GPUs can be employed to further accelerate rendering on each node, for example in sort-last approach with load balancing [27] or with multiple GPUs in a single physical node [21]. Typically, sort-last decompositions scale well with large data, but have to deal with more advanced compositing techniques such as binary swap [17], direct send [7], or radix-k [31]. More recently, distributed sort-first volume rendering has been employed [26] in conjunction with GPU acceleration and load balancing, avoiding a final compositing step but sacrificing the ability to scale easily with large data.…”

Section: Related Workmentioning

confidence: 99%

Visualization of Astronomical Nebulae via Distributed Multi-GPU Compressed Sensing Tomography

Wenger

Ament

Guthe

et al. 2012

IEEE Trans. Visual. Comput. Graphics

View full text Add to dashboard Cite

(a) (b) (c) (d) Fig. 1. The planetary nebula M2-9 is a typical example of a bipolar nebula. Its quasi-symmetric twin lobes of ionized material emanate from a binary star system in its center. Assuming axial symmetry, our reconstruction algorithm uses a single input image (a) to produce a high-resolution 3D visualization that closely resembles the original image when rendered from the same viewpoint (b). From a novel vantage point, the emission along the principal axis of the nebula accumulates and creates a luminous halo (c). As the vantage point approaches the symmetry axis, the received intensity further increases and the perceived shape of the nebula changes toward two entangled rings (d). The resolution of the reconstructed volume is 512 3 voxels. Original image: Bruce Balick (University of Washington), Vincent Icke (Leiden University, The Netherlands), Garrelt Mellema (Stockholm University), and NASA.Abstract-The 3D visualization of astronomical nebulae is a challenging problem since only a single 2D projection is observable from our fixed vantage point on Earth. We attempt to generate plausible and realistic looking volumetric visualizations via a tomographic approach that exploits the spherical or axial symmetry prevalent in some relevant types of nebulae. Different types of symmetry can be implemented by using different randomized distributions of virtual cameras. Our approach is based on an iterative compressed sensing reconstruction algorithm that we extend with support for position-dependent volumetric regularization and linear equality constraints. We present a distributed multi-GPU implementation that is capable of reconstructing high-resolution datasets from arbitrary projections. Its robustness and scalability are demonstrated for astronomical imagery from the Hubble Space Telescope. The resulting volumetric data is visualized using direct volume rendering. Compared to previous approaches, our method preserves a much higher amount of detail and visual variety in the 3D visualization, especially for objects with only approximate symmetry.

show abstract

Direct send compositing for parallel sort-last rendering

Cited by 25 publications

References 5 publications

Massively parallel volume rendering using 2--3 swap image compositing

Massively parallel volume rendering using 2--3 swap image compositing

Dynamic Load Balancing Method Based on PBT Tree

Visualization of Astronomical Nebulae via Distributed Multi-GPU Compressed Sensing Tomography

Contact Info

Product

Resources

About