skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Rendering and Compositing Infrastructure Improvements to VisIt for Insitu Rendering

Abstract

Compared to posthoc rendering, insitu rendering often generates larger numbers of images, as a result rendering performance and scalability are critical in the insitu setting. In this work we present improvements to VisIt's rendering and compositing infrastructure that deliver increased performance and scalability in both posthoc and insitu settings. We added the capability for alpha blend compositing and use it with ordered compositing when datasets have disjoint block domain decomposition to optimize the rendering of transparent geometry. We also made improvements that increase overall efficiency by reducing communication and data movement and have addressed a number of performance issues. We structured our code to take advantage of SIMD parallelization and use threads to overlap communication and compositing. We tested our improvements on a 20 core workstation using 8 cores to render geometry generated from a $256^3$ cosmology dataset and on a Cray XC31 using 512 cores to render geometry generated from a $$2000^2 \times 800$$ plasma dataset. Our results show that ordered compositing provides a speed up of up to $$4 \times$$ over the current sort first strategy. The other improvements resulted in modest speed up with one notable exception where we achieve up to $$40 \times$$ speed up of rendering and compositing of opaque geometry when both opaque and transparent geometry are rendered together. We also investigated the use of depth peeling, but found that the implementation provided by VTK is substantially slower,both with and without GPU acceleration, than a local camera order sort.

Authors:
 [1];  [1]
  1. Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
Publication Date:
Research Org.:
Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
Sponsoring Org.:
USDOE Office of Science (SC)
OSTI Identifier:
1237686
Report Number(s):
LBNL-1004236
ir:1004236
DOE Contract Number:
AC02-05CH11231
Resource Type:
Technical Report
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING

Citation Formats

Loring, Burlen, and Ruebel, Oliver. Rendering and Compositing Infrastructure Improvements to VisIt for Insitu Rendering. United States: N. p., 2016. Web. doi:10.2172/1237686.
Loring, Burlen, & Ruebel, Oliver. Rendering and Compositing Infrastructure Improvements to VisIt for Insitu Rendering. United States. doi:10.2172/1237686.
Loring, Burlen, and Ruebel, Oliver. 2016. "Rendering and Compositing Infrastructure Improvements to VisIt for Insitu Rendering". United States. doi:10.2172/1237686. https://www.osti.gov/servlets/purl/1237686.
@article{osti_1237686,
title = {Rendering and Compositing Infrastructure Improvements to VisIt for Insitu Rendering},
author = {Loring, Burlen and Ruebel, Oliver},
abstractNote = {Compared to posthoc rendering, insitu rendering often generates larger numbers of images, as a result rendering performance and scalability are critical in the insitu setting. In this work we present improvements to VisIt's rendering and compositing infrastructure that deliver increased performance and scalability in both posthoc and insitu settings. We added the capability for alpha blend compositing and use it with ordered compositing when datasets have disjoint block domain decomposition to optimize the rendering of transparent geometry. We also made improvements that increase overall efficiency by reducing communication and data movement and have addressed a number of performance issues. We structured our code to take advantage of SIMD parallelization and use threads to overlap communication and compositing. We tested our improvements on a 20 core workstation using 8 cores to render geometry generated from a $256^3$ cosmology dataset and on a Cray XC31 using 512 cores to render geometry generated from a $2000^2 \times 800$ plasma dataset. Our results show that ordered compositing provides a speed up of up to $4 \times$ over the current sort first strategy. The other improvements resulted in modest speed up with one notable exception where we achieve up to $40 \times$ speed up of rendering and compositing of opaque geometry when both opaque and transparent geometry are rendered together. We also investigated the use of depth peeling, but found that the implementation provided by VTK is substantially slower,both with and without GPU acceleration, than a local camera order sort.},
doi = {10.2172/1237686},
journal = {},
number = ,
volume = ,
place = {United States},
year = 2016,
month = 1
}

Technical Report:

Save / Share:
  • This report is about a work effort where the overall objectives were to establish a methodology and approach for selected transmission and distribution (T&D) grid modernization; monitor the results; and report on the findings, recommendations, and lessons learned. The work reported addressed T&D problems and solutions, related reliability issues, equipment and operation upgrades, and respective field testing.
  • Recent advances in high throughput/automated compositing with robotics/field-screening methods offer seldom-tapped opportunities for achieving cost reduction in ground-water-quality monitoring programs. An economic framework is presented for the evaluation of sample compositing as a screening tool in ground-water-quality monitoring. When the likelihood of occurrence of a contaminant in a well is very small, the use of sample compositing instead of routine exhaustive sampling will lead to reduction in analytical efforts. Such reduction will be maximum when there are no contaminated wells in the network. An N-fold reduction will result when none of the wells in a network of N wells aremore » contaminated. When 25% or more wells in a network are contaminated, the use of sample compositing will require, at the most, an additional 50% analytical effort compared to exhaustive sampling. A quantitative ratio (fl) of laboratory analytical cost to that of well installation and field sampling costs and a ratio (f2) of the expected number of contaminated wells to that of the total number of wells were determined. Several useful mathematical results of primary interest are derived and illustrated with case examples in the paper. Selected areas for further research are also outlined.« less
  • A probabilistic strategy for sample compositing is developed to minimize the radiometric analyses required for classifying non-radioactive waste for transport. Such waste must have concentrations of total radioactivity that are below an acceptance limit of L = 2 nCi/g. A composite of N different samples must have a radioactive concentration below L/N to assure that no individual sample is unacceptable. Unacceptable samples are eventually identified by analyses of successive splits composited with N/2, N/4, {hor ellipsis} of the original samples. The probable number of such analyses is derived using Gaussian distributions for the composite concentrations, per invoking the Central Limitmore » Theorem. A preliminary compositing strategy, based only on the average concentration {mu}, uses N = L/{mu} to yield a minimum fraction of {approx}2{mu}/L analyses per total samples. These approximations are useful for L/{mu} > 4. Refined strategies, based on both the {mu} and {sigma} for the concentration distribution, define the optimization more precisely. Experimental data from composites of 880 samples of low-level radiometric waste are consistent with the calculated predictions. 11 refs., 4 figs., 1 tab.« less
  • Some statistical aspects of compositing field samples of soils for determining the content of Pu are discussed. Some of the potential problems involved in pooling samples are reviewed. This is followed by more detailed discussions and examples of compositing designs, adequacy of mixing, statistical models and their role in compositing, and related topics.
  • Data from the Advanced Very High Resolution Radiometer (AVHRR) were used in a cooperative project, sponsored by the U.S. Department of Agriculture, Forest Service, Southern Forest Experiment Station, and the United Nations, Food and Agriculture Organization (FAO), to map Mexico's forest cover types. To provide satisfactory AVHRR data sets for the project, the sensor scan angle needed to be calculated for data points in the composite image data, and the clouds had to be removed and composite image data created from individual data sets. Techniques used to accomplish those two tasks are described here. Concepts illustrated here should be applicablemore » to other similar projects.« less