A Supercomputer for the Field
Posted on: June 8, 2015
Combining cutting-edge hardware and software to create a mobile oil and gas exploration workstation
Analyzing what is underneath the surface of the earth is incredibly data intensive, which is one of the reasons oil and gas companies are interested in big data solutions. During exploration, scientists often work with datasets that would make a traditional computer or even professional-grade computer crash or slow to a crawl.
For scale, imagine a beach where each grain of sand is a single Excel cell. Now take a 10-mile stretch of beach and have bulldozers dig down 1,000 feet. Finally, try taking all of that sand and dump it into a funnel the size of a basketball hoop.
Your computer would either crash immediately or would slow to a crawl. To avoid this problem, most datasets recording subsurface terrain are brought back to supercomputer labs to be analyzed and viewed. As you guessed, this is quite an inconvenience.
What if oil and gas companies could analyze datasets onsite?
Together with Lenovo, Nvidia and CGG, Magma demonstrated that exact solution at EAGE 2015 in Madrid, Spain; a supercomputer powerful enough to handle 10 terabyte datasets and small enough to take just about anywhere in the world.
Building a supercomputer to GO
- Professional-grade workstation
- Multicore processors
- Versatile i/o
- Designed to work with heavy processing software suites
- Designed to run the world’s fastest GPUs
- Advanced PCIe Gen 3 backplane
- Performance cooling and power capabilities
- Versatile sub-terrain software
- Ultra-fast visualizations on huge datasets
- Designed to take advantage of GPU architecture
- Fastest Nvidia Quadro GPU
- 12 GB of memory each
- 3,072 CUDA cores each
The implications of onsite computing for oil and gas
- Fewer resources allocated to moving datasets
- Instant visibility to data
- Less time waiting for transfers and processing
- An overall savings in hardware cost and data transfer