The Experts below are selected from a list of 276 Experts worldwide ranked by ideXlab platform

Michael Greenacre - One of the best experts on this subject based on the ideXlab platform.

Alvin R. Lebeck - One of the best experts on this subject based on the ideXlab platform.

  • MICRO - Exploiting instruction level parallelism in geometry processing for three dimensional Graphics applications
    Proceedings. 31st Annual ACM IEEE International Symposium on Microarchitecture, 1
    Co-Authors: Chia-lin Yang, Barton J. Sano, Alvin R. Lebeck
    Abstract:

    Three dimensional (3D) Graphics applications have become very important workloads running on today's computer systems. A cost-effective Graphics solution is to perform geometry processing of 3D Graphics on the host CPU and have specialized hardware handle the rendering task. In this paper, we analyze microarchitecture and SIMD instruction set enhancements to a RISC superscalar processor for exploiting instruction level parallelism (ILP) in geometry processing for 3D computer Graphics. Our results show that 3D geometry processing has inherent parallelism. When ignoring cycle time effects, an 8-issue processor can achieve up to 60% performance improvement over a 4-issue. However, certain application attributes can hinder the exploitation of ILP on a superscalar processor. Adding SIMD operations improves performance from 8% to 28% on a 4-issue processor that can issue at most 2 floating-point operations. If processor cycle time scales with the number of ports to the register file, doubling only the floating-point issue width of a 4-issue processor with SIMD instructions gives the best performance among the architecture configurations that we examine (the most aggressive configuration is an 8-issue processor with SIMD instructions).

Chia-lin Yang - One of the best experts on this subject based on the ideXlab platform.

  • MICRO - Exploiting instruction level parallelism in geometry processing for three dimensional Graphics applications
    Proceedings. 31st Annual ACM IEEE International Symposium on Microarchitecture, 1
    Co-Authors: Chia-lin Yang, Barton J. Sano, Alvin R. Lebeck
    Abstract:

    Three dimensional (3D) Graphics applications have become very important workloads running on today's computer systems. A cost-effective Graphics solution is to perform geometry processing of 3D Graphics on the host CPU and have specialized hardware handle the rendering task. In this paper, we analyze microarchitecture and SIMD instruction set enhancements to a RISC superscalar processor for exploiting instruction level parallelism (ILP) in geometry processing for 3D computer Graphics. Our results show that 3D geometry processing has inherent parallelism. When ignoring cycle time effects, an 8-issue processor can achieve up to 60% performance improvement over a 4-issue. However, certain application attributes can hinder the exploitation of ILP on a superscalar processor. Adding SIMD operations improves performance from 8% to 28% on a 4-issue processor that can issue at most 2 floating-point operations. If processor cycle time scales with the number of ports to the register file, doubling only the floating-point issue width of a 4-issue processor with SIMD instructions gives the best performance among the architecture configurations that we examine (the most aggressive configuration is an 8-issue processor with SIMD instructions).

M. M. Kaila - One of the best experts on this subject based on the ideXlab platform.

  • Simulated Operational Characteristics of a High-T _c Terahertz (THz) Thermal Sensor
    Journal of Superconductivity, 2004
    Co-Authors: M. M. Kaila, M. M. Kaila
    Abstract:

    A solution of the time-dependent heat transfer differential equations for the two-temperature (electron and phonon) model for the sensor-lattice and the lattice-substrate interface system is used for a computational analysis of the operational characteristics, for a high- T _c heterodyne thermal sensor. Responsivity and conversion gain of the sensor are analyzed in particular reference to external control parameters, e.g. the bias currents and voltages. Some operational characteristics in the form of Three-Dimensional Graphics of the sensor are presented.

Tianshu Zhou - One of the best experts on this subject based on the ideXlab platform.

  • Accurate depth based post-processing for perception enhancement in real time Three-Dimensional Graphics
    2008
    Co-Authors: Jim X. Chen, Tianshu Zhou
    Abstract:

    Increased understanding of the importance of human visual perception within the computer Graphics community is leading to new approaches in the way in which computer generated scenes are created. New techniques are being developed to simulate important aspects of the human visual system, such as depth cueing, in an effort to improve realism, and the sense of ‘immersion’ of a participant. Yet many of these techniques require a trade-off between accuracy and performance, often resulting in specialized implementations that do not consider the need to integrate with other techniques or existing visualization systems. The main aim of the dissertation is to present and demonstrate new techniques for generating depth cueing effects for real-time rendering of high-quality interactive Graphics. Depth cueing helps perceiving distance and space. When these cues are missing or inaccurately presented, the realism of the computer generated images is degraded. For example, most computer generated Graphics lack depth of field; hence everything in the scene appears perfectly sharp. This dissertation applies new scene depth based techniques in generating real-time depth of field effect. We propose a new post-processing algorithm for accurate depth cueing effects in real time 3D Graphics. Compared to existing work, our method will perform scene depth reconstruction on per pixel basis, decoupling it from the scene rendering process, and making it easy to integrate into existing visualization applications – a key aspect overlooked by existing implementations. Built on this scene depth reconstruction, we explore a new technique for restoring the full 3D fragment position; new algorithms for generating depth of field and layered fog are developed respectively for improved performance and quality. The new algorithms are designed to be executed entirely on the Graphics processing unit (GPU), taking advantage of the parallelism of modern Graphics cards and permitting real time perception enhancement effects. Our new depth of field algorithm performs two pass adaptive filtering in the post-processing step in the GPU, achieving real time frame rates and eliminating leakage artifacts. The new layered fog algorithm performs Euclidean distance fog integration along the line of sight using specialized fog density functions, similarly in the post-processing step with hardware acceleration, achieving high real time frame rates (3 ms for fog processing). In addition to ideally suited to a GPU implementation for achieving real time frame rates, both algorithms share the new accurate scene depth reconstruction algorithm resulting high quality images. In order to demonstrate the depth cueing effectively, a new comprehensive scene management application framework has been built. This application framework is capable of converting standard 3D scene world from VRML, 3ds, max, lightwave, etc to application recognizable format, then load, display, navigate and archive the scene world. The experimentations demonstrated that the post-processing effects are activated through a hook in the standard rendering pipe line, with minimum impact to the existing application and easy integration. The dissertation explores the realism of computer Graphics from perception point of view and tackles the issues from a practical level. It provides ready to use algorithms for generating quality depth cueing effects in real time as well as integrated solutions to apply them to existing rendering systems.