40 data processing & interpretation
further performance boost for most processes. NVIDIA
are also planning to double the speed of their processors
every 18 months.
Using GPUs for spectral inversion
OpenGeoSolutions, a seismic data analysis company
that pioneered the use of spectral decomposition and
inversion, has reported a 55 times performance increase
after upgrading to the Tesla C1060 GPU.
“We are measuring speedups from two hours to two
minutes using CUDA and the Tesla C1060,” says James
Allison, president of OpenGeoSolutions. “This kind of
performance increase is totally unprecedented and in a
market where there is great economic value in being able
to determine these fine sub-surface details, this is a game
Fig. 3. Using GPUs in the Interactive Facies Classification workflow.
changer.”
Left: three seismic attribute volumes, terrace amplitude, envelope
On a CPU-based cluster, the seismic inversion process
and chaos (top to bottom) are computed using all GPUs. Middle: the
could take anywhere from 2 hours to several days, Allison
interpreter blends the volumes interactively and picks classes using the
says. In an effort to improve this, the team acquired a
visualisation GPU. Right: the final facies volume is calculated using all
workstation equipped with an NVIDIA Tesla C1060 GPU.
GPUs.
Over six weeks, the OpenGeoSolutions team converted a
key portion of their application to CUDA.
➠
the software to build CUDA into this step. “The Tesla products essentially give us all a personal
“We expect to have tools out in March next year which supercomputer,” says Allison. “Just one Tesla C1060
uses CUDA in this interactive way. That will allow us to delivers the same performance as our 64 CPU cluster, and
do much more, it’s much easier to write sophisticated this was a resource we had to share. This is a huge cost
algorithms using CUDA,” says Purves. and time saving that has transformed our workflow and
Finally volume processing, which can often be a slow boosted our productivity.”
process on a data set of several gigabytes, is performed
again using multiple GPUs to calculate the final volume.
NVIDIA and CUDA
The IFC has isolated 3 classes (right): the gas chimney NVIDIA is best known for producing graphics cards for
(red), the high amplitude peaks (yellow) and high gaming and CAD (Computer Aided Design) applications.
amplitude troughs (blue). But the humble graphics card in many a home gaming
The complexity of the gas chimney and the PC can be put to use to speed up common computational
distribution of the amplitude anomalies are now easy to problems many times over.
interpret from the detailed 3D representation. GPUs (Graphics Processing Units) are basically
multiple core, massively parallel processing chips
Benchmarks
that can perform many operations at the same time,
FfA is working on a series of benchmarks for various something that is not possible with the CPU that runs
common processes, comparing GPU enabled calculations a PC. They also use dedicated fast communications
with fast CPUs, and also looking at the speed interfaces, meaning the data can get in and out fast
improvements from upgrading older machines (Fig. 4). enough to take advantage of the processing power.
This will give customers a guide as to how much they CUDA is the first C language environment that enables
need to spend to take full advantage of the GPU enabled programmers and developers to write software to solve
processes, given their current hardware configuration complex computational problems in a fraction of the time
Fig. 5 shows the speed improvement for three typical by tapping into the many-core parallel processing power
processes: TDiffusion, a complex noise cancellation of GPUs.
calculation; DipAzimuth, a structural attribute; and Trace It makes it easier to code sophisticated algorithms
Attribute, a standard trace envelope. to use the graphics processing chip, and performance
The comparison is with a base system with two benefits compared to standard workstation of over one
quad core Intel Xeon processors and 16Gb of memory. hundred times are not unusual, according to NVIDIA.
As you add more processor cards, the performance It has been developed by NVIDIA to enable wider use
improvements continue to ramp up. of its GPUs for a range of applications, including those
A twelve times improvement might not sound that outside its traditional gaming and visualisation market.
impressive, but it means for example a 36 minute job
coming down to a three minute job.
The latest processors
Once all the algorithms are CUDA enabled, ffA will CUDA can be used on a wide range of NVIDIA products,
start to work on optimising them, which should give a from home gaming graphics cards to rack-mounted server
www.engineerlive.com
Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44