High-performance Graphics Workstation Research Articles

Launch of the I13-2 data beamline at the Diamond Light Source synchrotron

Users of the Diamond-Manchester Imaging Branchline I13-2 commonly spend many months analysing the large volumes of tomographic data generated in a single beamtime. This is due to the difficulties inherent in performing complicated, computationally-expensive analyses on large datasets with workstations of limited computing power. To improve productivity, a ‘data beamline’ was launched in January 2016. Users are scheduled for visits to the data beamline in the same way as for regular beamlines, with bookings made via the User Administration System and provision of financial support for travel and subsistence. Two high-performance graphics workstations were acquired, with sufficient RAM to enable simultaneous analysis of several tomographic volumes. Users are given high priority on Diamond’s central computing cluster for the duration of their visit, and if necessary, archived data are restored to a high-performance disk array. Within the first six months of operation, thirteen user visits were made, lasting an average of 4.5 days each. The I13-2 data beamline was the first to be launched at Diamond Light Source and, to the authors’ knowledge, the first to be formalised in this way at any synchrotron.

The Construction of Seismic and Geological Studies' Cloud Platform Using Desktop Cloud Visualization Technology

D visualization technology makes full use of 3D seismic information and properties. Using 3D graphics to de- scribe all kinds of complex data fields and the data relation can be easily perceived, it is an important method of seismic and geological research. In petroleum exploration, the high performance graphics workstations are deployed to complete the interpretation work. The traditional application mode leads to the following problems: the utilization of resource is not balanced between different workstations; the lack of hardware, software and data resources sharing; and collaborative re- search is not convenient between experts in different places. By constructing seismic and geological studies' cloud plat- form using desktop cloud visualization technology, the traditional C/S architecture software using mode change to B/S mode, the platform improve the utilization rate of resources. The cloud platform realizes the seismic data interpretation work based on the 3D remote visualization technology, and the user needs only a desktop computer to complete all jobs. Through the cloud platform, users can access the software and hardware resources needed during the research by network, collaborative research of experts in different places will be very convenient, and the platform also provide science basis and data support for the comprehensive decisions.

Hybrid architecture for 3D visualization of ultrasonic data

With the current proliferation of multi-core processors and hardware acceleration, large-data processing methods are increasingly being modified for use in parallel and distributed computing environments. In this paper, we present a hybrid architecture for the visualization and processing of large-scale volumetric data. Various hardware environments and technologies are integrated in this architecture to perform interactive operations on very large volumetric datasets. All of the datasets are stored in a data center with a gigabit network environment. Time-consuming data-processing tasks are allocated to compute nodes connected to the same network, while the visualization and interaction operations are executed on a high-performance graphics workstation. OpenCL and OpenMP are used to implement volume rendering algorithms to accelerate visualization of a hierarchical volume data structure using GPUs and multi-core CPUs. Various out-of-core algorithms are also presented to process the large dataset directly. The experimental results indicate that the proposed hybrid architecture and methods are effective and efficient in processing and visualizing very large-volumetric datasets.

English

Height field terrain rendering is an important aspect of GIS, outdoor virtual reality applicationssuch as flight simulation, 3-D games, etc. A polygonal model of very large terrain data requiresa large number of triangles. So, even most high-performance graphics workstations have greatdifficulty to display even moderately sized height fields at interactive frame rates. To bringphotorealism in visualisation, it is required to drape corresponding high-resolution satellite oraerial phototexture over 3-D digital terrain and also to place multiple collections of point-location-based static objects such as buildings, trees, etc and to overlay polyline vector objects suchas roads on top of the terrain surface. It further complicates the requirement of interactive framerates while navigation over the terrain. This paper describes a novel approach for objects andterrain visualisation by combination of two algorithms, one for terrain data and the other forobjects. The terrain rendering is accomplished by an efficient dynamic multiresolution view-dependent level-of-detail mesh simplification algorithm. It is augmented with out-of-corevisualisation of large-height geometry and phototexture terrain data populated with 3-D/2-Dstatic objects as well as vector overlays without extensive memory load. The proposedmethodology provides interactive frame rates on a general-purpose desktop PC with OpenGL-enabled graphics hardware. The software TREND has been successfully tested on different real-world height maps and satellite phototextures of sizes up to 16K*16K coupled with thousandsof static objects and polyline vector overlays.

Development of a PC‐based real‐time surgical simulator

AbstractPresently, in constructing a surgical simulator, a high‐performance graphics workstation (GWS) has generally been used because it is necessary that a highly faithful surgical image be generated in real time. There exist problems, however, in cost, operation, and working environment, which have prevented the widespread use of the simulator. In order to solve this problem, the authors constructed a real‐time PC‐based surgical simulator for the training of ophthalmologists in vitreoretinal surgery, based on the results of evaluation of a surgical simulator developed on GWS. From this study, the necessary conditions for the simulator have been identified, and a PC‐based surgical simulator satisfying these conditions has been successfully constructed. The constructed system can perform the following operations. The surgical image is displayed in stereo vision, through a binocular stereo vision display. The tactile sensation of surgical tool insertion is provided by using a simulated tool device. The surgery scene can be controlled by operating a foot switch. The constructed surgical simulator provides an environment which effectively simulates real surgery. If an increased number of ophthalmologists can use the proposed simulator more easily for repetitive surgery training, surgical techniques will be improved, and new surgical procedures may be developed. © 2002 Wiley Periodicals, Inc. Syst Comp Jpn, 33(7): 11–20, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.10076

リアルタイム手術シミューレタの構築

This paper describes the architecture of real-time surgical simulators. First of all, a high performance graphics workstation based real-time simulation so that it could supply the user demand; however, there remained one problem of the cost. Then, the next simulator was developed with a personal computer in order to reduce the system cost and also new software simulating another surgery case was developed. Especially, the cost of a stereo display and haptic devices, both of which were the core components of the system and occupied the large ratio of the cost, were reduced by specifying the functions of them for ocular surgery. As a result, we succeeded in establishing a simulator that could supply the user demand of boty functions and the cost. Therefore, it has been tried to develop a neurosurgery simulator with the same components as that of the ocular surgery one. Although it remains the problem of haptic display, it is found that the specification of the stereo view can apply to not only ocular surgery but aloso neurosurgery simulators.

Real-time 3D dose calculation and display: A tool for plan optimization

Purpose : Both human and computer optimization of treatment plans have advantages; humans are much better at global pattern recognition, and computers are much better at detailed calculations. A major impediment to human optimization of treatment plans by manipulation of beam parameters is the long time required for feedback to the operator on the effectiveness of a change in beam parameters. Our goal was to create a real-time dose calculation and display system that provides the planner with immediate (fraction of a second) feedback with displays of three-dimensional (3D) isodose surfaces, digitally reconstructed radiographs (DRRs), dose-volume histograms, and/or a figure of merit (FOM) (i.e., a single value plan score function). This will allow the experienced treatment planner to optimize a plan by adjusting beam parameters based on a direct indication of plan effectiveness, the FOM value, and to use 3D display of target, critical organs, DRRs, and isodose contours to guide changes aimed at improving the FOM value. Methods and Materials : We use computer platforms that contain easily utilized parallel processors and very tight coupling between calculation and display. We ported code running on a network of two workstations and an array of transputers to a single multiprocessor workstation. Our current high-performance graphics workstation contains four 150-MHz processors that can be readily used in a shared-memory multithreaded calculation. Results : When a 10 × 10-cm beam is moved, using an 8-mm dose grid, the full 3D dose matrix is recalculated using a Bentley—Milan-type dose calculation algorithm, and the 3D dose surface display is then updated, all in < 0.1 s. A 64 × 64-pixel DRR calculation can be performed in < 0.1 s. Other features, such as automated aperture calculation, are still required to make real-time feedback practical for clinical use. Conclusion : We demonstrate that real-time plan optimization using general pyrpose multiprocessor workstations is a practical goal. Parallel processing technology provides this capability for 3D planning systems, and when combined with objective plan ranking algorithms should prove effective for optimizing 3D conformal radiation therapy. Compared to our earlier transputer work, multiprocessor workstations are more easily programmed, making software development costs more reasonable compared with uniprocessor development costs. How the dose calculation is partitioned into parallel tasks on a multiprocessor work station can make a significant difference in performance. Shared-memory multiprocessor workstations are our first choice for future work, because they require minimum programming effort and continue to be driven to higher performance by competition in the workstation arena.

A dynamic finite element surface model for segmentation and tracking in multidimensional medical images with application to cardiac 4D image analysis

This paper presents a physics-based approach to anatomical surface segmentation, reconstruction, and tracking in multidimensional medical images. The approach makes use of a dynamic “balloon” model—a spherical thin-plate under tension surface spline which deforms elastically to fit the image data. The fitting process is mediated by internal forces stemming from the elastic properties of the spline and external forces which are produced from the data. The forces interact in accordance with Lagrangian equations of motion that adjust the model's deformational degrees of freedom to fit the data. We employ the finite element method to represent the continuous surface in the form of weighted sums of local polynomial basis functions. We use a quintic triangular finite element whose nodal variables include positions as well as the first and second partial derivatives of the surface. We describe a system, implemented on a high performance graphics workstation, which applies the model fitting technique to the segmentation of the cardiac LV surface in volume (3D) CT images and LV tracking in dynamic volume (4D) CT images to estimate its nonrigid motion over the cardiac cycle. The system features a graphical user interface which minimizes error by affording specialist users interactive control over the dynamic model fitting process.

The Determination of Environmental Accessibility and ADA Compliance Through Virtual Wheelchair Simulation

The widespread use of power wheelchairs has greatly increased the requirements for the accessibility of buildings and other architectural structures to handicapped persons. In addition, recent advances in microcomputer technology have made possible increasingly sophisticated power wheelchair interfaces, such as halo, puff and sip, and muscle control mechanisms, which can provide mobility for an even larger portion of the handicapped population. Finally, the ADA (the Americans with Disabilities Act of 1990) requires handicapped accessibility for (almost) all public structures. We have developed a virtual structure prototyping system that allows navigation by a person using a power wheelchair. The system is a tool for three groups of people: (1) for architects and designers, it provides structure visualization that can both improve the handicapped accessibility of building designs and test a structure for ADA compliance; (2) for wheelchair users, it provides more appropriate device fitting and training with wheelchair control mechanisms; and (3) for health care professionals, it provides evaluations of wheelchair users. The system consists of an instrumented, joystick-driven power wheelchair connected to a high-performance graphics workstation; it simulates the actual speed and maneuverability of the particular wheelchair within a virtual structure. The display generates realistic interiors containing multiple light sources and surface textures and is viewed in stereo through lightweight polarized glasses. The system maintains a hierarchical data structure which detects collisions between the virtual wheelchair and the environment. In this paper we discuss (1) the system's user interface, (2) the system's hardware and software configuration, (3) the impact of the system on the architectural design process, and (4) future system additions. In the last section we also discuss virtual manipulation for enabling technology.

The harsh reality of large, integrated data volumes

The general availability of 3-D data interpolation and display tools, and recent reductions in the cost of high performance graphics workstations to run them on, has helped introduce such tools into the mainstream of geoscience interpretation. Many types of data (such as formation tops, fluid contacts, faults, porosity, and seismic data attributes) can be modeled quickly on a full or partial field basis, and visually analyzed by the armchair interpreter.

Cooperative Visualization of Computational Fluid Dynamics

AbstractTempus Fugit/Interview is a computational fluid dynamics visualization application for which processing is distributed between high performance graphics workstations and supercomputers. Facilities are provided in the application for more than one user to view shared images creating a cooperative visualization environment. The way in which the computation is partitioned between the super computer and the workstations is critical to the capability of the application to present simultaneous, identical, animated images of fluid dynamics to more than one user.

Use of transputers for real time dose calculation and presentation for three-dimensional radiation treatment planning

Purpose : Real-time 3-dimensional dose calculation will allow display of isodose contours and other metrics for a planner to assess plan effectiveness during plan development, facilitating optimization. Methods and Materials : Parallel processing provides an effective means to calculate 3-dimensional dose distribution in real-time while plan parameters are being chosen and adjusted. An array of 20 transputers and a high performance graphics workstation have demonstrated the feasibility of real-time 3-dimensional beam parameter specification, dose calculation, and dose-distribution presentation for evaluation. A mesh connected set of processors using surface processors to generate and terminate rays, and ray processors to calculate ray attenuation and dose distribution has been developed to efficiently utilize large numbers of processors and provide good load sharing, even for small beams that intersect only a small part of the volume. Results : Our feasibility study has calculated dose distribution by the Effective Path Length method in about one second per beam for a treatment volume of 56,400 voxels. We expect to reduce the total time for computation, communication, and display, with even larger volumes, to less than one second. The number of processors can easily be increased for larger treatment volumes or more accurate and computation-intensive dose-calculation algorithms. Transputers provide an elegant and economical method for harnessing up to hundreds of powerful general-purpose processors for computational tasks including dose calculation and isodose contour generation. The same distributed-memory parallel-processing configuration is also suitable for calculation of isodose contours and dose-volume histograms for plan evaluation, automatic calculation of apertures and filters as beam parameters are manipulated, and more accurate dose calculation algorithms that incorporate the effects of scatter. Conclusion : Parallel processors can efficiently provide real-time calculation of the information necessary to evaluate treatment plans as they are developed allowing the planner to optimize the plan based on dose distribution and its effects on tumor control and complications.

Computer‐generated three‐dimensional reconstruction of the mediastinm correlated with sectional and radiological anatomy

AbstractWe describe a three‐dimensional computer‐generated model of the human mediastinum. Contours were manually traced from photographs of a serially sectioned, unfixed frozen cadaver block, Using software developed in‐house, the contours were digitized, automatically assembled and placed in registration, and displayed on a high performance graphics workstation as surfaced images. The model thus produced has advantages over traditional representations of anatomical structures; it is realistic, accurate, efficient, and flexible. Each anatomical component of the mediastinum is stored as a separate file in the computer and can be selectively accessed, displayed, and manipulated, allowing the model to be disassembled, reassembled, dissected, and viewed in any orientation. The model can be divided into a series of transversely cut “slabs,” the surfaces of which correlated with the anatomy displayed in the original cryosections and with radiological scans, thus permitting 2‐D sectional anatomy to be viewed in the context of the 3‐D model. Animation of the model can be accomplished by recording pre‐planned scripts onto videodiscs. The videodiscs serve as image archives that can be viewed in an interactive, self‐paced manner or accessed from a computer‐based multimedia system that is being developed by the Digital Anatomist Program. Anatomical reconstructions based on datasets of x‐, y‐, z‐coordinates have potential clinical applications as models that can be displayed to match the sectional planes of radiological images. The mediastinum model we describe is part of a large image database being developed by the Digital Anatomist Program that will form the spatial component of a knowledge base in structural biology. © 1992 Wiley‐Liss, Inc.

Myofiber orientation in the weanling mouse heart

This study provides a quantitative description at the cellular level of myofiber orientation throughout the ventricles of the mouse heart. We employed computer-based methods of three-dimensional reconstruction from 3 microns plastic-embedded serial sections. Registration marks were introduced by drilling minute holes into each plastic block. Subfields of selected sections were photographed at 20 x magnification, using a computer-controlled microscope. The 35-mm film frames were projected onto a digitizer tablet and the epi- and endocardial boundaries were digitized manually. The "heads" and "tails" of linear segments of a representative myofiber sample present in each projected image were digitized in point mode. The many x-, y-, z-coordinate tables generated by digitization were reassembled automatically, giving a numerical description of the myofiber pattern. This pattern was studied interactively on a high-performance graphics workstation. We find that the heart wall is, to a first approximation, a "sandwich," in which the myofibers in the middle layer run mainly circumferentially, whereas those in the inner and outer layers run parallel or oblique to the apical-basal axis, a variant of the classical model of the myofiber pattern. We observed a "sleeve" in the interventricular septum, formed by longitudinal and oblique myofibers, a feature which apparently has not been described previously. Myofibers not running parallel to the transverse or longitudinal planes were not resolved in this study. We conclude that three-dimensional reconstruction of the cardiac myofiber pattern at the light-microscopic level, while laborious, is technically feasible and scientifically worthwhile.

Scientific visualization in high-speed network environments

In several cases, new visualization techniques have vastly increased the researchers' ability to analyze and comprehend data. Similarly, the role of networks in providing an efficient supercomputing environment has become more critical and continues to grow and at a faster rate than the increase in the processing capabilities of supercomputers. A close relationship between scientific visualization and high-speed networks in providing an important link to support efficient supercomputing is identified. The two technologies are driven by the increasing complexities and volume of supercomputer data. The interactions of scientific visualization and high-speed networks in a Computational Fluid Dynamics simulation/visualization environment are given. Current capabilities supported by high-speed networks, supercomputers, and high-performance graphics workstations at the Numerical Aerodynamic Simulation Facility (NAS), NASA Ames Research Center are described. Applied research in providing a supercomputer visualization environments to support future computational requirements are summarized.

A SIMD parallel trapezoid rasterization algorithm

An algorithm for the rasterization of trapezoids on the Connection Machine TM (CM) is described. The input consists of an array of trapezoids, with two horizontal sides, arranged with one trapezoid per processor. (Unless otherwise indicated, “processor” should be taken to mean virtual processor.) Each trapezoid is converted to an edge record and the edge records are then distributed to enough processors so that each processor is responsible for one scanline of one trapezoid. Each processor computes a scan record for its scanline, and the scan records are then distributed to enough processors so that one processor is responsible for a single pixel. Final interpolation of position, and possibly shading information, is performed in parallel for all pixels thus created, and the pixels are then broadcast into a frame buffer array, with depth comparisons being performed at the receiving end to ensure that the nearest pixel appears in the array. The Connection Machine-specific features used by the algorithm are logarithmic time cumulative summing, general communication with comparisons for collision arbitration, and virtual processor sets. Performance is similar to that of good graphics workstations. The intended application is to display data already resident in the machine as the result of some previous computation, when a high-performance graphics workstation is not available.

Meaningful graphics workstation performance measurements

We present how graphics workstation performance is currently advertised by workstation manufactures. We suggest that the manufacturers' performance numbers are not the best indicators of how well the system will perform. Four levels of graphics system performance measurement have been defined by the NCGA Graphics Performance Characterization group[1]. These levels are low-level primitives (points, lines, and polygons per second), pictures, systems (input and action response times), and applications. We suggest that the best indicator of a graphics workstation's performance capabilities is that of an application running on the system. For applications level performance measurements, there are many potential applications from which to choose. We suggest that three-dimensional visual simulators are one of the better applications to study as that type of system has a realistic mix of processor and graphics computational requirements. We describe one such system, the Moving Platform Simulator (MPS), developed at the Naval Postgraduate School with the goal of fully listing its capabilities and utilized graphics techniques. We then present how that system performs on a currently available, high-performance graphics workstation.

Numerical flow field visualization

The application of computer-generated imagery to analyze numerical solutions of fluid flow is presented. High-performance graphics workstations, their current capabilities, and their use in flow field analysis are discussed. Interactive flow field computations about complex three-dimensional bodies applying computer-generated imagery are projected for the future.

Computer-aided research into a natural neural network

While artificial neural networks are being applied to solving problems in signal processing and other domains, much remains to be discovered about how even the simpler biological neural networks function. Studying simpler examples of natural systems promises to advance our understanding of principles of organization of neural tissue wherever it occurs. It also may prove useful in the development of new computer architectures. Accordingly, the authors have begun the study of the macular linear bioaccelerometers, or balance organs of mammals, using the rat as the model for the class. This effort includes creating a computer-based workbench that a scientist can use to generate geometric reconstructions of neural tissue from electron microscope serial sections, to create a functional model of information flow within the neural tissue, and ultimately to generate computer animations to visualize how the network functions. Their work to data is based upon the study of long series of sections in a transmission electron microscope. The sections are photographed and the photographs are assembled into montages. Selected nerves and receptor units synapsing with them (their receptive fields) are next traced onto transparencies from the montages. The tracings, which are cross-sectional contours, are digitized with a tablet and stored inmore » data files on a personal computer. The files are transferred to a high performance graphics workstation, where software has been developed to reconstruct these sets of contours as polygonal objects, display them in wireframe or solid form, and create sequence files that can be used to produce a computer animation on videotape.« less

Coordinated three-dimensional reconstruction from serial sections at macroscopic and microscopic levels of resolution: the human heart.

This report describes procedures we have developed for obtaining correlated quantitative structural information at two very different levels of resolution. Accurate reconstructions of entire organs and samples of tissue within organs were produced in correct scalar and topographical relationship using computer-assisted techniques. A specially designed sectioning apparatus, a macrovibratome, was used to section serially the ventricles of the human heart macroscopically. Photographs were taken of every slice. A tissue block excised from a slice at a specified locus in the left ventricular wall was embedded in plastic; serial-3 micron sections were cut in each of two orthogonal orientations. Photomicrographs were taken by semi-automated microscopy. Images of both macroscopic and microscopic sections were projected onto a bitpad and manually digitized. The resulting tables of x-, y-, and z-coordinates were reassembled on a VAX 11/750 computer, then transferred to a high-performance graphics workstation and displayed as three-dimensional images. Microscopic images were shown in the correct reference frame with respect to the macroscopic (parent) structure.