Search

We describe a framework for direct volume rendering of adaptive mesh refinement (AMR) data that operates directly on the hierarchical grid structure, without the need to resample data onto a single, uniform rectilinear grid. The framework can be used for a range of renderers optimized for particular hardware architectures: a hardware-assisted renderer for single-processor graphics workstations, and a massively parallel software-only renderer for supercomputers. It is also possible to use the framework for distributed rendering servers. By exploiting the multiresolution structure of AMR data, the hardware-assisted renderers can render large AMR data sets at interactive rates, even if the data is stored remotely.

A mantra we've heard over the years is that the consumer gaming market drives graphics technology. Our thesis is that the same does not hold true for networking technology. Instead, it is applications from high performance graphics and visualization efforts, not games, that push networks past their breaking point, which in turn stimulates work that results in better networks. This case study shows how a high performance computational science and visualization application has set new levels of networking performance by winning the SC2001 Application Bandwidth Challenge by novel use of existing network protocols.

We discuss a web-based portal for the exploration, encapsulation, and dissemination of visualization results over the Grid. This portal integrates three components: an interface client for structured visualization exploration, a visualization web application to manage the generation and capture of the visualization results, and a centralized portal application server to access and manage grid resources. We demonstrate the usefulness of the developed system using an example for Adaptive Mesh Refinement (AMR) data visualization.

Adaptive mesh refinement (AMR) is a numerical simulation technique used in computational fluid dynamics (CFD). It permits the efficient simulation of phenomena characterized by substantially varying scales in complexity of local behavior of certain variables. By using a set of nested grids at different resolutions, AMR combines the simplicity of structured rectilinear grids with the possibility to adapt to local changes in complexity and spatial resolution. Hierarchical representations of scientific data pose challenges when isosurfaces are extracted. Cracks can arise at the boundaries between regions represented at different resolutions. We present a method for the extraction of isosurfaces from AMR data that avoids cracks at the boundaries between levels of different resolution.

The complexity of physical phenomena often varies substantially over space and time. There can be regions where a physical phenomenon/quantity varies very little over a large extent. At the same time, there can be small regions where the same quantity exhibits highly complex variations. Adaptive mesh refinement (AMR) is a technique used in computational fluid dynamics (CFD) to simulate phenomena with drastically varying scales concerning the complexity of the simulated variables. Using multiple nested grids of different resolutions, AMR combines the topological simplicity of structured-rectilinear grids, permitting efficient computation and storage, with the possibility to adapt grid resolutions in regions of complex behavior. We present methods for direct volume rendering of AMR data. Our methods utilize AMR grids directly for efficiency of the visualization process. We apply a hardware-accelerated rendering method to AMR data supporting interactive manipulation of color-transfer functions and viewing parameters. We also present a cell-projection-based rendering technique for AMR data.

Adaptive mesh refinement (AMR) is a numerical simulation technique used in computational fluid dynamics (CFD). By using a set of nested grids of different resolutions, AMR combines the simplicity of structu red rectilinear grids with the ability to adapt to local changes in complexity in th e domain. Without proper interpolation on the boundaries of grids of different lev els of a hierarchy, discontinuities can arise. Treating locations of data values giv en at cell centers of AMR grids as vertices of a dual grid creates gaps between hierarchy levels. Using an index-based tessellation approach that fills these gaps with ``stitch cells'' we define an interpolation scheme that avoids discontinuities at level boundaries. We use the resulting interpolation scheme to generate volume-rendered images. Modifying transfer functions on a per-level basis allows us to emphasize (or de-emphasize) a specific level and gain a better understanding of the underlying hierarchical structure.

The performance growth rate of high-performance computing (HPC) systems has fallen from 1000× to just 10× every eleven years. The HPC world, like large cloud service provider data centers, has turned to heterogeneous acceleration to deliver continued performance growth through specialization. Chiplets offer a new, compelling approach to scaling performance through adding workload-specific processors and massive bandwidth to memory into computing systems. If design and manufacturing challenges are resolved, chiplets can offer a cost-effective path for combining die from multiple function-optimized process nodes, and even from multiple vendors, into a single application-specific integrated circuit (ASIC). This article explores opportunities for building and improving the performance of bespoke HPC architectures using open-modular "chiplet" building blocks. The hypothesis developed is to use chiplets to extend the functional and physical modularity of modern HPC systems to within the semiconductor package. This planning can reduce the complexity and cost of assembling chiplets into an ASIC product and make it easier to build multiple product variants. [ABSTRACT FROM AUTHOR]

The long, steady advance of supercomputing capabilities historically makes it tempting to assume that similar future improvements are inevitable. In fact, the successful path to exascale systems was far from certain, requiring a shift to accelerator-based designs, an associated end-to-end rethinking of traditional computational approaches, and a fundamental change in how scientists collaborate and interact. In this article, we share lessons from the U.S. Department of Energy Exascale Computing Project along with perspectives on how to design and implement scientific applications in the exascale and post-exascale eras. [ABSTRACT FROM AUTHOR]