K. Ma, J. Painter, C. Hansen, and M. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59--68, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the ray casting algorithm is known to produce the highest quality of rendered images, although it is one of the most compute and memory intensive tasks for volume visualization. There have been several previous works to parallelize the ray casing algorithm on various supercomputing environments [1, 7, 8, 12]. In attempt to ray cast large volume datasets at interactive rates, we have developed an effective parallel distributed ray casting scheme that can be used to visualize very large volume data in the distributed environments. In particular, we are concerned with two platforms. First, we consider ....

K.-L. Ma, J. Painter, C. Hansen, and M. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59--68, July 1994.

....30] this approach requires custom designed components, which are expensive. Also, latency is induced if compositing is done in multiple stages [16] There has been considerable work over the last decade on reducing the overheads of image composition in sort last systems. For instance, Ma et al. [15] reduced pixel redistribution costs by keeping track of the screen space bounding box of the graphics primitives rendered by each processor. Cox and Hanrahan [9] went further by keeping track of which pixels were updated by each processor (the active pixels) and only composited them instead of ....

....pixels) and only composited them instead of compositing the entire screen. Ahrens and Painter [1] compressed pixel color and depth values using run length encoding before transmitting them for composition. Other researchers have investigated composition schemes for specific network topologies [15, 20, 29] or viewing conditions [21] However, some of these methods are effective only if the viewpoint does not zoom in to view a portion of the model, and none of them has achieved low enough communication bandwidths for interactive rendering on a PC cluster. Most encouraging speedup results have been ....

[Article contains additional citation context not shown here]

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59--68, 1994.

....volume rendering algorithms, and show how different taxonomies from the developed classification provide the ability to compare, contrast, and show potential for future research results. Keywords: volume visualization, splatting, ray casting, taxonomy 1 References [98] 49] 16, 48, 113] 44] [52] [107] 111, 81] 94] 16, 15] 34] 101] 40] 66] 73] 90] 116] 26] 20] 68] 6] 54, 52, 63, 28, 67] 7, 1, 42, 4, 62] 31, 75, 30, 38, 86] 41, 21, 119, 112, 118, 45, 58] 55, 65, 14, 29, 79] 115, 53, 70, 74, 87, 19, 117, 32] 72, 57, 92, 36, 7] 111, 12, 90, 97, 78, 5] 15, 27, 35, ....

....provide the ability to compare, contrast, and show potential for future research results. Keywords: volume visualization, splatting, ray casting, taxonomy 1 References [98] 49] 16, 48, 113] 44] 52] 107] 111, 81] 94] 16, 15] 34] 101] 40] 66] 73] 90] 116] 26] 20] 68] 6] [54, 52, 63, 28, 67] [7, 1, 42, 4, 62] 31, 75, 30, 38, 86] 41, 21, 119, 112, 118, 45, 58] 55, 65, 14, 29, 79] 115, 53, 70, 74, 87, 19, 117, 32] 72, 57, 92, 36, 7] 111, 12, 90, 97, 78, 5] 15, 27, 35, 43, 87, 28, 51, 22, 23, 59, 60, 64, 63, 66, 68, 71, 80, 81, 91, 93, 99, 47, 108] 103, 116, 83, 106, 107, 105, ....

[Article contains additional citation context not shown here]

K. L. Ma, J. Painter, C. D. Hansen, and M. F. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59--67, 1994.

....graphics port interface specification, May 1998. http: www.intel.com pc supp platform agfxport. 4] G. Abram and H. Fuchs. Vlsi architectures for computer graphics. In G. Enderle, editor, Advances in Computer Graphics I, pages 6 21, Berlin, Heidelberg, New York, Tokyo, 1986. Springer Verlag. [5] A.V. Aho, R. Sethi, and J.D. Ullman. Compilers Principles, Techniques, and Tools. Addison Wesley, 1988. 6] K. Akeley. Realityengine graphics. In Computer Graphics (Proc. Siggraph) pages 109 116, August 1993. 7] K. Akeley and T. Jermoluk. High performance polygon rendering. Computer ....

....volume are divided into multiple processors, each computes a sub image, that is later regrouped. The main shortcomings are the higher need for communication and synchronization among the processors (parallel machines still have slow communications with respect to processor speed) Ma et al. [5] describes an efficient algorithm for re grouping the images back together to form a single correct image. In statically partitioning the volume dataset, one has to be careful to give every processor the same amount of work (see [11] Time Space Parallelism. In time space parallelism, ....

K. Ma, J. Painter, C. Hansen, and M. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59-- 68, 1994.

....field and isosurfaces the result. Then it reads the temperature field and colors the isosurface based on the temperature. The isourface is then rendered using a software renderer and the resulting image is composited with other image results using a sort last binary swap compositing algorithm[11] 4 . In addition, an isosurface of the ocean floor is created using the temperature field data and rendered as well. The dataset is partitioned to the processors using the method described in this section. Color Figure 10 provides a visual example of block partitioning of the ocean data, by ....

K.L. Ma, J. Painter, C. Hansen, and M. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics, pages 59--67, July 1994.

....structures evolve over time. For visualizing large data sets, parallel processing is often used to speed up the expensive volumetric rendering process. Although the subject of rendering a single volumetric data set using a parallel computer has been studied extensively by numerous researchers [17, 16, 14, 22, 10], parallel animation of TVVD, in contrast, has received relatively little attention. Compared to parallel volume rendering of a single data set, rendering TVVD in parallel poses a di#erent set of design tradeo#s. First, because TVVD typically consists of a sequence of data volumes, the I O ....

....volume animation requires re thinking of the types of parallelism one should exploit to achieve the optimal performance. In particular, I O overlap and resource utilization e#ciency play a crucial role in the parallelization strategy. We start with a generic parallel volume rendering program [14], modify it to experiment with di#erent approaches for parallel volume animation of time varying data sets, and analyze the performance tradeo# among various partitioning strategies. Although the results and analysis are based on implementations on an Intel Paragon, we believe that the conclusions ....

[Article contains additional citation context not shown here]

K.-L. Ma, J. S. Painter, C. Hansen, and M. Krogh, Parallel Volume Rendering Using BinarySwap Compositing, IEEE Computer Graphics and Applications, 14 (1994), pp. 59--68.

....strategy depends critically on the composition stage. Various methods have been proposed so far to perform the composition. The simplest method is to send the sub images to a single compositing processor [MEP92] Other schemes proposed are binary tree composition [SGS91] binary swap composition [MPHK94, Kar94] and parallel pipeline 14 composition [LRN96] Mitra and Chiueh [MC98] showed that all previously proposed sub image compositing methods can be uni ed in a single framework. In general, in sort last, a processor sends all the pixels of the relevant image space to another processor. This ....

K. Ma, J. S. Painter, C. D. Hansen, and M. F. Krogh. Parallel Volume Rendering using Binary-Swap Compositing. IEEE Computer Graphics and Applications, 14(4), 1994.

....This pixel redistribution overhead compares favorably with previous sort last methods. For instance, several volume rendering systems employ static partitions which result in an approximate depth complexity of , since the volume is divided into blocks along each of its three axes [15, 21]. If we assume each server is responsible for a tile covering pixels, the total number of pixels composited by each server is: # (5) a) Hand (643 objects, 654,666 polygons) b) Dragon (1,257 objects, 871,414 polygons) c) Buddha (1,476 objects, 1,087,716 polygons) Figure ....

....redundantly. In (b) and (c) brighter pixel intensities represent more image composition overheads. servers during each frame is related to the scene depth complexity (D) multipled by the resolution of the screen (P) where D is approximately ) and the total overhead is 0 21 [15, 21]. In contrast, for the hybrid algorithm, the number of pixels transferred is related to 0 # , where B is the average size of an object s bounding box (see Section 6) Empirically, the pixel redistribution overheads of the hybrid algorithm are much smaller than for sort last in ....

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59--68, 1994.

....version of the visible human dataset. Each image is 250 by 250. 15 DRAFT 4 Conclusions In this article, we have introduced the PVR system. The idea of developing PVR started out of frustration from trying to use a network of workstations and the Paragon as rendering engines for VolVis [2]. It was always clear that a pure distributed approach to building rendering environments would be much more powerful than special rendering tools with parallel capabilities. Here are some of the key features in our system: ffl Transparency PVR hides most of the hardware dependencies from the ....

....supported by the National Science Foundation under grants CCR 9205047 and DCA 9303181 and by the Department of Energy under the PICS grant. References [1] V. Anupam, C. Bajaj, D. Schikore, and M. Schikore. Distributed and collaborative visualization. IEEE Computer, 27(7) 37 43, 1994. [2] R. Avila, T. He, L. Hong, A. Kaufman, H. Pfister, C. Silva, L. Sobierajski, and S. Wang. Volvis: A diversified volume visualization system. In IEEE Visualization 94, pages 31 38. IEEE CS Press, October 1994. 17 DRAFT [3] D. Hillis. The Connection Machine. MIT Press, 1985. 4] A. E. Kaufman. ....

[Article contains additional citation context not shown here]

K. Ma, J. Painter, C. Hansen, and M. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59--68, 1994.

....compositing stage plays an important role in the effectiveness of the sort last strategy.The simplest method is to send the sub images to a single compositing processor [12] or to do binary tree composition [16] where more and more processors become idle as the composition progresses. Ma et al. [10] and Karia [8] described similar parallel compositing techniques for volume rendering applications that can be easily adapted for polygon rendering, and are now referred to as binary swapping. Lee et al. 9] showed that parallel pipeline compositing is more efficient that binary swapping in ....

K. L. Ma et al. Parallel Volume Rendering using Binaryswap Compositing. IEEE Computer Graphics and Applications, 14(4), 1994.

No context found.

K. Ma, J. Painter, C. Hansen, and M. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59--68, 1994.

No context found.

Ma, K.-L., Painter, J. S., Hansen, C., and Krogh, M. Parallel Volume Rendering Using Binary-Swap Compositing. IEEE Computer Graphics Applications 14, 4 (July 1994), 59--67.

No context found.

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh, "Parallel Volume Rendering using Binary-Swap Compositing," IEEE Comput. Graphics and Appl., vol. 14, no. 4, pp. 59--68, July 1993.

....[19] accelerating or eliminating ray voxel intersection tests and parallelization. Acceleration is usually accomplished by a combination of spatial subdivision and early ray termination [9] 8] 24] Ray tracing for volume visualization naturally lends itself towards parallel implementations [14], 16] The computation for each pixel is independent of all other pixels, and the data structures used for casting rays are usually read only. These properties have resulted in many parallel implementations. A variety of techniques have been used to make such systems parallel, and many successful ....

....pixel is independent of all other pixels, and the data structures used for casting rays are usually read only. These properties have resulted in many parallel implementations. A variety of techniques have been used to make such systems parallel, and many successful systems have been built (e.g. [14], 27] 22] 17] These techniques are surveyed by Whitman [29] B. Methods of Volume Visualization There are several ways that scalar volumes can be made into images. The most popular simple volume visualization techniques that are not based on cutting planes are isosurfacing, ....

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh. Parallel Volume Rendering using Binary-Swap Compositing. IEEE Comput. Graphics and Appl., 14(4):59--68, July 1993.

....this data set. Finally, a 265 time steps, 640#256#256 data set was used. Figure 5 displays a time step in this data set. These last two data sets were used to study selected aspects of the compression results and the parallel visualization pipeline. We use a parallel ray casting volume renderer [16] for our study. This renderer is reasonably optimized and capable of generating high quality images. There are other volume rendering algorithms such as the shear warp algorithm [12] which can not only deliver superior rendering rates but is also highly parallelizable [11] Since our task is to ....

K.-L. Ma, J. S. Painter, C. Hansen, and M. Krogh. Parallel Volume Rendering Using Binary-Swap Compositing. IEEE Computer Graphics Applications, 14(4):59-- 67, July 1994.

....Thus, this software framework pro TOP Left Right Bottom Figure 1: k D Tree Subdivision of a Data Volume vides an excellent platform for exploring different rendering techniques. 3 Binary Swap Volume Rendering The algorithm described here has previously been published in two different papers [10, 11]. Here we describe a brief overview of the algorithm. The basic idea behind our algorithm, like other similar methods, is very simple: divide the data up into smaller subvolumes distributed to multiple processors, render them separately and locally, and combine the resulting images in an ....

....with the half image they receive. A similar exchange and compositing of partial images is done between Processor 3 and 4. The key thing to note is that the Binary Swap algorithm sends more data than other parallel compositing algorithms but can exploit the fast interconnection network of MPPs [11]. 4 T3D Implementation The focus of this research was to implement a fast, hopefully interactive, volume renderer on the CRI T3D. We will not describe the T3D architecture in this paper. The reader is referred to the wealth of documentation from CRI and other CUG papers[6, 3] As noted, the ....

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh. Parallel Volume Rendering using BinarySwap Compositing. IEEE Comput. Graphics and Appl., 14(4):59--68, July 1993.

....a normal 64 bit integer maximum reduction to accomplish the z buffer composition. This new z buffer reduction uses our Binary Swap compositing scheme which runs in O(N=2 N=4 N=8 Delta Delta Delta N=P ) time where N is the number of pixels per processor and P is the number of processors 2 [14]. In the limit as P goes to infinity, this reduces to O(N) time. Since the total problem size grows by the product N Theta P , assuming a dense image distribution on each processor, an O(N) algorithm on P processors is optimal. 6 Results To gauge our algorithm s performance we timed several ....

K. Ma, J. Painter, C. Hansen, and M. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4):59-- 68, July 1994.

....the ray casting [7] projection [3] splatting [15] and shear warp [6] methods. Various optimization and acceleration techniques for volume rendering have also been developed including encoding object space coherence [8] encoding image space coherence [17] hardware assisted [13, 2] and parallel [10, 5] methods. To take advantage of these improvements in visualization algorithms, we need an e#ective user interface for volume visualization. Because volume data exploration often involves a trial and error process of parameter specification, an important part of a user interface is a structured ....

K.-L. Ma, J. S. Painter, C. Hansen, and M. Krogh, Parallel Volume Rendering Using BinarySwap Compositing, IEEE Computer Graphics & Applications, 14 (1994), pp. 59--67.

....[19] accelerating or eliminating ray voxel intersection tests and parallelization. Acceleration is usually accomplished by a combination of spatial subdivision and early ray termination [9] 8] 24] Ray tracing for volume visualization naturally lends itself towards parallel implementations [14], 16] The computation for each pixel is independent of all other pixels, and the data structures used for casting rays are usually read only. These properties have resulted in many parallel implementations. A variety of techniques have been used to make such systems parallel, and many successful ....

....pixel is independent of all other pixels, and the data structures used for casting rays are usually read only. These properties have resulted in many parallel implementations. A variety of techniques have been used to make such systems parallel, and many successful systems have been built (e.g. [14], 27] 22] 17] These techniques are surveyed by Whitman [29] B. Methods of Volume Visualization There are several ways that scalar volumes can be made into images. The most popular simple volume visualization techniques that are not based on cutting planes are isosurfacing, ....

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh. Parallel Volume Rendering using Binary-Swap Compositing. IEEE Comput. Graphics and Appl., 14(4):59--68, July 1993.

....[7] accelerating or eliminating ray voxel intersection tests and parallelization. Acceleration is usually accomplished by a combination of spatial subdivision and early ray termination [4] 8] 9] Ray tracing for volume visualization naturally lends itself towards parallel implementations [10], 11] The computation for each pixel is independent of all other pixels, and the data structures used for casting rays are usually read only. These properties have resulted in many parallel implementations. A variety of techniques have been used to make such systems parallel, and SUBMITTED TO ....

....for casting rays are usually read only. These properties have resulted in many parallel implementations. A variety of techniques have been used to make such systems parallel, and SUBMITTED TO IEEE TRANSACTIONS ON COMPUTER GRAPHICS AND VISUALIZATION 2 many successful systems have been built (e.g. [10], 12] 13] 14] These techniques are surveyed by Whitman [15] B. Methods of Volume Visualization There are several ways that scalar volumes can be made into images. The most popular simple volume visualization techniques that are not based on cutting planes are isosurfacing, ....

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh, "Parallel Volume Rendering using Binary-Swap Compositing," IEEE Comput. Graphics and Appl., vol. 14, no. 4, pp. 59--68, July 1993.

....[7] accelerating or eliminating ray voxel intersection tests and parallelization. Acceleration is usually accomplished by a combination of spatial subdivision and early ray termination [4] 8] 9] Ray tracing for volume visualization naturally lends itself towards parallel implementations [10], 11] The computation for each pixel is independent of all other pixels, and the data structures used for casting rays are usually read only. These properties have resulted in many parallel implementations. A variety of techniques have been used to make such systems parallel, and IEEE ....

....used for casting rays are usually read only. These properties have resulted in many parallel implementations. A variety of techniques have been used to make such systems parallel, and IEEE TRANSACTIONS ON COMPUTER GRAPHICS AND VISUALIZATION 2 many successful systems have been built (e.g. [10], 12] 13] 14] These techniques are surveyed by Whitman [15] B. Methods of Volume Visualization There are several ways that scalar volumes can be made into images. The most popular simple volume visualization techniques that are not based on cutting planes are isosurfacing, ....

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh, "Parallel Volume Rendering using Binary-Swap Compositing," IEEE Comput. Graphics and Appl., vol. 14, no. 4, pp. 59--68, July 1993.

No context found.

MA K. L., PAINTER J. S., HANSEN C. D., KROGH M. F.: Parallel Volume Rendering Using Binary-Swap Compositing. IEEE Computer Graphics and Applications 14, 4 (1994), 59--68. 2

No context found.

K.L. Ma, J.S. Painter, C.D. Hansen, and M.F. Krogh. Parallel Volume Rendering using Binary-Swap Compositing. IEEE Comput. Graphics and Appl., 14(4):59--68, July 1993.

No context found.

Kwan-Lui Ma, James S. Painter, Charles D. Hansen, and Michael F. Krogh. Parallel volume rendering using binary-swap compositing. IEEE Computer Graphics and Applications, 14(4), July 1994.

No context found.

K. Ma, J. Painter, C. Hansen, and M. Krogh, "Parallel volume rendering using binary-swap compositing," IEEE Computer Graphics and Applications 14(4), pp. 59--68, 1994.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC