Divisible load applications occur in many fields of science and engineering, can be eas-ily parallelized in a master-worker fashion, but pose several scheduling challenges. While a number of approaches have been proposed that allocate work to workers in a single round, using multiple rounds improves overlap of computation with communication. Unfortunately, multi-round algorithms are difficult to analyze and have thus received only limited attention. In this paper we answer three open questions in the multi-round divisible load scheduling area: (i) How to account for latencies? (ii) How to account for heterogeneous platforms; and (iii) How many rounds should be used? To answer (i), we derive the first closed-form optimal schedule for a homogeneous platform with both computation and communication latencies, for a given number of rounds. To answer (ii) and (iii), we present a novel algorithm, UMR. We use simulation to evaluate UMR in a variety of realistic scenarios.

Recently, several external memory techniques have been developed for a wide variety of graphics and visualization problems, including surface simplification, volume rendering, isosurface generation, ray tracing, surface reconstruction, and so on. This work has had significant impact given that in recent years there has been a rapid increase in the raw size of datasets. Several technological trends are contributing to this, such as the development of high-resolution 3D scanners, and the need to visualize ASCI-size (Accelerated Strategic Computing Initiative) datasets. Another important push for this kind of technology is the growing speed gap between main memory and caches, which penalizes algorithms that do not optimize for coherence of access. Because of these reasons, much research in computer graphics focuses on developing out-of-core (and often cache-friendly) techniques. This paper surveys fundamental issues, current problems, and unresolved questions, and aims to provide graphics researchers and professionals with an effective knowledge of current techniques, as well as the foundation to develop novel techniques on their own. Keywords: Out-of-core algorithms, scientific visualization, computer graphics, interactive rendering, vol-ume rendering, surface simplification.

In this paper we describe a parallel and out-of-core view-dependent isocontour visualization algorithm that efficiently extracts and renders the visible portions of an isosurface from large datasets. The algorithm first creates an occlusion map using ray-casting and nearest neighbors. With the occlusion map constructed, the visible portion of the isosurface is extracted and rendered. All steps are in a single pass with minimal communication overhead.

In this paper, we propose a novel out-of-core isosurface extraction technique for large time-varying fields over irregular grids. We employ our meta-cell technique to explore the spatial coherence of the data, and our time tree algorithm to consider the temporal coherence as well. Our one-time preprocessing phase first partitions the dataset into meta-cells that cluster spatially neighboring cells together and are stored in disk. We then build a time tree to index the meta-cells for fast isosurface extraction. The time tree takes advantage of the temporal coherence among the scalar values at different time steps, and uses BBIO trees as secondary structures, which are stored in disk and support I/O-optimal interval searches. The time tree algorithm employs a novel meta-interval collapsing scheme and the buffer technique, to take care of the temporal coherence in an I/O-efficient way. We further make the time tree cache-oblivious, so that searching on it automatically performs optimal number of block transfers between any two consecutive levels of memory hierarchy (such as between cache and main memory and between main memory and disk) simultaneously. At run-time, we perform optimal cache-oblivious searches in the time tree, together with I/O-optimal searches in the BBIO trees, to read the active meta-cells from disk and generate the queried isosurface efficiently. The experiments demonstrate the effectiveness of our new technique. In particular, compared with the query-optimal mainmemory algorithm [Cignoni et al. 1997] (extended for time-varying fields) when there is not enough main memory, our technique can speed up the isosurface queries from more than 18 hours to less than 4 minutes.

Abstract not found

Abstract not found

We develop a new indexing structure and a new out-of-core scheme to extract and render isosurfaces for large scale time-varying 3-D volume data. The new algorithm enables the fast visualization of arbitrary isosurfaces cut by a user-specified hyperplane along any of the four dimensions. Our data structure makes use of the entropy measure to establish the relative resolutions of the spatial and temporal dimensions rather than treating the temporal dimension just as any other dimension. The preprocessing is very efficient and the resulting indexing structure is very compact. We have tested our scheme on 40GB subset of the Richtmyer-Meshkov instability data set and obtained very good performance for a wide range of isosurface extraction queries.

Visualization of multidimensional data presents special challenges for the design of efficient out-of-core data access. Elements that are nearby in the visualization may not be nearby in the underlying data file, which can severely tax the operating system's disk cache. The Granite Scientific Database System can address these problems because it is aware of the organization of the data on disk, and it knows the visualization method's pattern of access. The access pattern is expressed using a toolkit of iterators that both describe the access pattern and perform the iteration itself. Because our system has knowledge of both the data organization and the access pattern, we are able to provide significant performance improvements while hiding the details of out-of-core access from the visualization programmer.

Large datasets, on the order of GB and TB, are increasingly common as abundant computational resources allow practitioners to collect, produce and store data at higher rates. As dataset sizes grow, it becomes more challenging to interactively manipulate and analyze these datasets due to the large amounts of data that need to be moved and processed. Application-independent caches, such as operating system page caches and database buffer caches, are present throughout the memory hierarchy to reduce data access times and alleviate transfer overheads. We claim that an applicationaware cache with relatively modest memory requirements can effectively exploit dataset structure and application information to speed access to large datasets. We demonstrate this idea in the context of a system named the tree cache, to reduce query latency to large octree datasets by an order of magnitude.

Abstract Ray tracing a volume scene graph composed of multiple point-based volume objects (PBVO) can produce high quality images with effects such as shadows and constructive operations. A naive approach, however, would demand an overwhelming amount of memory to accommodate all point datasets and their associated control structures such as octrees. This paper describes an out-of-core system for rendering such a scene graph in a scalable manner. In order to address the difficulty in pre-determining the order of data caching, we introduce a technique based on a dynamic, in-core working set. We present a ray-driven algorithm for predicting the working set automatically. This allows both the data and the control structures required for ray tracing to be dynamically pre-fetched according to access patterns determined based on captured knowledge of ray-data intersection. We have conducted a series of experiments on the scalability of the technique using working sets and datasets of different sizes. With the aid of both qualitative and quantitative analysis, we demonstrate that this approach allows the rendering of multiple large PBVOs in a volume scene graph be performed on desktop computers. Keywords out-of-core · very large dataset visualization · octree · point-based modeling · point-based rendering · ray tracing · volume scene graph