>Home >ffA Software >The Core >GPU Processing
ffA Software
GPU Accelerated Computing
SVI Pro 2009 is the first 3D Seismic Volume Analysis application capable of accelerating 3D seismic analysis
through simultaneous computation across multiple GPUs, on a Windows workstation.
Graphics Processing Unit (GPU) technology has to date provided significant increases in the graphics and visualisation
capabilities of computer workstations.
The GPU can now also be harnessed as a highly parallel processor for computational tasks other than computer graphics.
SVI Pro’s GPU enabled processes provide significant performance improvements over computation on modern Quad Core CPU’s
which translate to significant time savings and increases in productivity.
Recent benchmarks on ffA’s GPU enabled algorithms are showing overall performance improvements ranging
from 5x to 22x on a single GPU, depending on the algorithm applied.
Simultaneous use of multiple GPUs provides further step performance improvements, as illustrated in the example below.
To date a selection of key analysis algorithms are GPU enabled and within 2009, ffA will bring to market it complete suite of 3D Seismic Volume
Analysis tools on this multiple GPU enabled computation framework
See also the R&D High Performance Computing page for more information.
 |
Benchmarks for SVI Pro’s algorithms running on a modern multiple GPU workstation.
|
Last Modified: 19th May 2009