3D and beyond. What graphics cards can do for you
Open Journal System Proceedings of TDWG
MetadataShow full item record
Biological mapping and visualisation tools like the ones developed withinBiodiversity Information Standards (TDWG) and GBIF (Global Biodiversity Information Facility) context are presented to a public accustomed to gimmicks and eye candy. Compare any map viewer with the rich fly-through world of Google Earth for example, and you'll notice we're missing a dimension. There's something unnatural about our current 2D presentation of biological datasets, about not being able to grab it and look at it from all possible directions - be it a bunch of specimen gathering spots, or a river trajectory, or measuring points on an ape skull. This is more than just eye candy:This is crucial for a good interpretation. And why are we always registering heights for our specimens, if not for using them in 3D? A bold Star-Trek style prediction (this is the "wild ideas" section after all): "We need to conquer the 3rd dimension, or we condemn ourselves to oblivion". 3D in visualisation, but 3D in data space as well: Spatial dataset calculations are an order of magnitude more complex than our current "distances" and "nearest neighbour" questions. How to get the 3D power for this brave new world? A tantalising new approach is GPGPU, or General Processing GPU (Graphical Processing Unit) processing. In normal language: Look for the processing power where it can be found. No other line of computer hardware is under such rapid development as the GPU (Graphical Processing Unit) on the graphics card, with throngs of gamers crying out for ever faster, more spectacular games. GPUs come relatively cheap as well, compared to your normal processor (CPU or Central Processing Unit). And you can string together more than one graphics card nowadays, to get one 3D super-number cruncher. An "idiot savant" supercruncher that is, because GPU's traditionally are built for a single task: Rendering as many frames per seconds as possible on your screen. They excel at shading, rendering 3D scenes and ray tracing (all things 3D), but don't ask them to calculate a square root. Clever software engineers have been working at just that: Tricking 3D graphics cards into "multi-purpose" processing, thus extracting an enormous processing power "on the cheap". Different frameworks exist for programming general science problems into the single-minded graphics card. But rather than losing ourselves in the technical titbits, we'd like to present some examples of what people are doing with this technology: 3D scenery rendering for movie special effects 3D physics animations (effects of gravity and wind on vegetation and constructions ) 3D particle behaviour and collision detection (fluids, clouds, sand grains) Realistic texture rendering (e.g. skin texture) for movies Software virus signature matching Encryption and decryption Random-number generation Earth subsurface imaging from seismic data We'll also address how GPGPUs could possibly serve in a biological/biodiversity context - as "wild idea" and food for thought: 3D river trajectory calculations for extracting biological parameters (e.g. speed of flow, total length of rivers) Random-number generating for Monte Carlo algorithms for biological statistics 3D calculations of (nearest neighbour) patterns within a specimen gathering dataset 3D correlation search between parameter layers (e.g. height) and specimen data points Biological virus signature matching / DNA pattern matching (e.g. in biogeography) Fast 3D rendering for online GIS (Geographical Information Systems) applications 3D skull and skeleton reconstructions and calculations.
CitationMeganck, B.; Mergen, P. (2008). 3D and beyond. What graphics cards can do for you. , Biodiversity Information Standards (TDWG) Annual Conference 2008, Vol. 2008, Open Journal System Proceedings of TDWG, DOI: http://www.tdwg.org/proceedings/article/view/335.