As the year’s end fast approaches, now may be a good time to look at some recent announcements related to computer use in the oil & gas industry. The industry’s place at the forefront of high-performance computing highlights its need for innovative hardware solutions.

To take one example, Total’s Houston Geophysical Research Group (HGRG) is tasked with researching and developing better seismic imaging/migration algorithms. Jing Wen, GeoScience/Technical Systems Supervisor, has noted that almost every year new algorithms require another order of magnitude of processing power.

The 3-D wave equation RTM algorithm is a case in point. According to Appro, which provides supercomputing solutions, although RTM theory and method has been developed for some time, use was limited by the cost and availability of computing power. For this and other reasons, Total is said to be considering acquisition of a petra-flops scale machine in the near future.

Recently, Appro delivered to the Lawrence Livermore, Los Alamos, and Sandia National laboratories 21 scalable units — 438 tera-flops, Linux high-performance clusters. Lawrence Livermore Labs alone has up to 495 tera-flops of computing power spread over 18 Linux clusters.

The gold standard when it comes to visual computing, in O&G and other industries, is NVIDIA. Its Quadro ultra-high-end graphics solutions are said to transcend the limits facing today’s visualization applications. Armed with a new graphics architecture, NVIDIA ultra high-end products deliver up to 102 giga-byte/second memory bandwidth, 4 giga-byte GDDR3 frame buffer memory, and 512 -bit memory interface.

The recently introduced Quadro FX 5800 is said to give geophysicists visual supercomputing from their desktops. As a second-generation solution delivers up to a 50% increased performance over the first generation through 240 processor cores. Its feature set includes up to 4 gigabytes of frame buffer leading to greater interactivity with large datasets.

More down to earth, a persistent problem for those wishing to access seismic files and archived project data is that they must often wait weeks to access this information due to cumbersome procedures related to tape storage at remote sites. Moving those files to disk storage means the information can be made immediately available. The advantages are clear, but one good question is how to get it done.

Landmark, working with Storwize for its compression and decompression technology and NetApp for enterprise storage, has introduced a solution that includes the deployment, data migration, and disaster recovery services needed to make the transition from tape to disk, including indexing files and incorporating metadata for quick retrieval. Landmark says that archival storage costs can thereby be lowered to less than $1,000 per terabyte, as well as reducing associated infrastructure footprint, and
cooling, power, and administrative costs. Called DeepStor, the solution is available in 600 terabyte and 1.2 petabyte versions, assuming a 30% compression rate.

Rackable Systems, a provider of servers and storage to large-scale data centers, can deliver a self-contained, fully portable data center installed in a shipping container. With a network connection and a power source provided, the unit is ready to use when it arrives at its destination. In the oil & gas industry, companies can use the data center to move extreme computing power to the data collection site for real-time analysis of large data sets. A 40 ft. by 8 ft. shipping container can contain up to 22,400 cores or storage capacity of up to 7.1 peta-bytes. By combining water- and air-cooling technology, the company says it can reduce cooling costs by as much as 80% over a traditional data center.

Another provider of modular data centers, Verari Systems, says that its FOREST Container is 110% more efficient than a traditional data center and is suitable for extreme temperature and rugged terrain environments. While some competing solutions may not support other vendors, its solution can house networking equipment or even existing nodes from manufacturers using less efficient traditional front-to-back cooling.

If you’ve been good, maybe some of these advanced toys for engineers and scientists will show up in your holiday stocking.