Cuda-based parallel line integral convolution on gpu For high-performance visualization of large flow data

Journal of Computer Science & Systems Biology

ISSN: 0974-7230

Open Access

Cuda-based parallel line integral convolution on gpu For high-performance visualization of large flow data

International Conference on Big Data Analysis and Data Mining

May 04-05, 2015 Kentucky, USA

Shuyi Liu

Posters-Accepted Abstracts: J Comput Sci Syst Biol

Abstract :

Scientific visualization is an important application of computer graphics to scientific computing by providing deep insight into the complex pattern underlying big data, while flow visualization is crucial for visual data analysis in oceanographicatmospheric modeling, computational fluid dynamics simulation, and electro-magnetic field study that involve in-depth exploration and interpretation of directional information. Line Integral Convolution (LIC) is a texture-based flow visualization technique for creating a realistic high-resolution representation of the global structure and local detail of the flow. One major issue with LIC is the high computational cost, which restricts its applicability to visualization of massive flow data. The coherence between the intrinsic parallelism of LIC (with image synthesis on a per-pixel basis) and the parallel computing capability of CUDA (Compute-Unified Device Architecture) for GPU (Graphics Processing Unit) allows for CUDA-based parallel generation of an LIC image on GPU, with all pixels synthesized simultaneously as opposed to the serial (one-by-one) mechanism on CPU. This research addresses CUDA-based parallel LIC on GPU for high-performance visualization of large flow data. Built on a thorough investigation of the pipeline and implementation detail of CPU-based LIC, the innovative work consists in the design and implementation of a GPU-based LIC pipeline through CUDA. In addition, a Win32 program called cudaLIC has been developed and applied to the visualization of two large real flows (the Northeast Pacific ocean flow and Hurricane Isabel wind flow) for effective as well as efficient pattern analysis, with up to 30X speedup over CPU-based LIC yet with the same image quality.

Biography :

Shuyi Liu is a junior at Paul Laurence Dunbar (PLD) High School, Lexington, KY. As a member of the prestigious MSTC (Math, Science, and Technology Center) program of PLD, he has accomplished a variety of AP classes such as ?AP Computer Science AB? (including Java Programming), while he currently participates in the PLD Robotics club. With a strong passion for computer science, Mr. Liu is particularly interested in computer graphics, scientific visualization, image processing, parallel computing (through CUDA-based GPU programming), and GUI programming. Since the 2014 fall semester, he has been doing research on image processing under the supervision of Dr. Ruigang Yang (Professor of Computer Science at the University of Kentucky). Recently Mr. Liu won a second-place award at the Central Kentucky Regional Science and Engineering Fair (Feb 2015) under the ?Computer Software? category and has been invited to attend the Kentucky State Science and Engineering Fair (March 2015).

50+ Million Readerbase

Journal Highlights

Google Scholar citation report
Citations: 2279

Journal of Computer Science & Systems Biology received 2279 citations as per Google Scholar report

Journal of Computer Science & Systems Biology peer review process verified at publons

Indexed In

arrow_upward arrow_upward