Understanding Scientific Visualization in High Performance Computing: Process & Advantages, Slides of Computational Techniques

Download Understanding Scientific Visualization in High Performance Computing: Process & Advantages and more Slides Computational Techniques in PDF only on Docsity!

Scientific

Visualization

in

High

Performance

Computing

What

Is

Scientific

Visualization?

Scientific visualization is the study and practice of representing scientific data in a visual, pictorial form that is easier for humans to interpret than raw data.

Special tools used to calibrate, debug, and interpret results of large scale computer simulations.

Some

Advantages

of

Scientific

Visualization

Gives a new perspective on the data

Allows a viewer to quickly grasp the "big picture" - Can be used to interpolate any missing data - Can be made "user friendly" for non

technical viewers

Advantages

of

Visualization

(cont)

Quickly create or modify data sets by manipulating graphic objects on computer screen

Easier to spot errors in inconsistencies in massive data sets

Performance

Analysis

Subroutine histograms created during program execution can be used to determine flow of logic and time spent computing subroutine.

Useful for advanced architectures. Help balance computational load across large parallel processor machine.

Interpretation

of

Data

The human brain possesses a “narrow bandwidth” for processing raw numbers, but a surprisingly “wide bandwidth” for processing visual data.

Creative use of color, texture, and translucency can create large scale models that are relatively easy to understand.

History

(Definition)

Previously defined as “formation of visual images; the act or process of interpreting in visual terms or of putting into visual form.”

Recent addition

“A

tool or method for interpreting image data fed into a computer for generating images from complex multi

dimensional data sets.

History

(cont)

This is not a new phenomenon.

Used in maps, scientific drawings, and data plots for over a thousand years. - Examples: map of China

ad) and the famous map of Napoleon’s invasion of Russia in

by Jacque Minard.

Advances

in

the

Field

The fields of graphics hardware and software are advancing at a rapid pace.

Computational scientist should be careful not to become implementation oriented. - In the past twenty years, speeds of high performance computers have risen for just over one megaflop to more than a gigaflop.

Advances

in

the

Field

(cont)

Memory has also increased at a rapid pace, from about a million words (8MB) to over a billion words (GB) - Algorithms have improved on similar scale, combined million times more efficient - Currently, UNIX workstations still offer better performance than a typical PC - However, gap is rapidly closing

Implementation

Use an existing visualization software package

Develop some new visualization software that is specifically designed to aid analysis of problem under study.

Using

Scientific

Visualization

Software

User needs to:

Produce graphic file containing the results to be viewed - Operate an appropriate software package - Navigate the

OS

and user interfaces of the selected workstation

Many

Types

of

Visualization

2D

vs.

3D

Static vs. animated

Big picture vs. narrow view - etc.

2D

Visualization

The simplest visualizations are business ‐ style two ‐ dimensional charts.