Geography Reference
In-Depth Information
There are two directions we can solve the intensive computing problem. Firstly,
graphics hardware-accelerating is a good way to increase the computing capabil-
ity. It mainly depends on powerful GPU parallel computing which can provide
better rendering capacity and serve high performance application, such as CUDA
platform published by NVIDIA. In addition, Distributed visualization strategy
and hybrid computing strategy with CPUs and GPUs are also effective ways to
tremendously improve the performance of computing resources to achieve the
intensive-computing task. Secondly, we can reduce as much volume as possible in
current rendering pipe. A typical atmospheric data set can be more than Terabytes
in size. To render it fluently require a fast memory. Generally, these optimization
techniques are categorized as empty space skipping, early ray termination, Octree
and BSP space subdivision, multiple and adaptive resolution representation, and
user layers of data for rendering. For example, Li et al. (
2011a
) gave an Octree-
based view dependent LOD strategy to help efficiently visualize large volume data.
As an example, in climate studies, we developed a set of advanced rendering
solutions to visualize the dynamic phenomena with virtual globe(World Wind) to
handle big data by: (1) Designing a fresh volume render algorithm which use sphere
as basic splatting unit to represent each cell from 4D data array, (2) Designing and
implementing data processing and render techniques to deal with large amount of
rendering volume data, such as LOD, (3) Making full use of machine's computing
resources of both CPU and GPU to accelerate the speed of rendering pipeline. As a
result, the system (Fig.
18.10
) provided more intuitive visual effects and convenient
interactivity operations in visualizing dynamic environmental phenomena and helps
scientists explore the potential patterns and principles in scientific research.
Fig. 18.10
World Wind based 4-dimensional visualization is used to explore climate patterns
Search WWH ::
Custom Search