Case Study Intel® Xeon Phi™ Coprocessors High-Performance Computing Energy

Case Study
Intel® Xeon Phi™ Coprocessors
High-Performance Computing
Energy

Case Study Intel® Xeon Phi™ Coprocessors High-Performance Computing Energy

Taking oil and gas exploration to the next level DownUnder GeoSolutions harnesses the power of high-performance computing using a platform based on Intel® Xeon® processor E5-2600 v2 product family and Intel® Xeon Phi™ coprocessors to fast-track their clients’ exploration “Deploying SGI® ...Rackable® cluster based on Intel® Xeon Phi™ coprocessors and the Intel® Xeon® processor E5- 2600 v2 product family allowed us to use the power of high-performance computing to achieve faster turnaround time in seismic processing while saving significantly on IT infrastructure costs and maintain our existing software investments.” – Dr. Matthew Lamont Managing Director DownUnder GeoSolutions With its head office located in West Perth, Australia, DownUnder GeoSolutions (DUG) is an innovative geosciences company offering a suite of geophysical processing solutions and services to global customers in the oil and gas industry. The company’s products and services continue to be at the cutting edge of exploration and production services in the industry, with a core strength that is defined by an integrated approach across its comprehensive service offering, which includes seismic illumination studies, seismic data processing, depth imaging, petrophysical processing and interpretation, quantitative interpretation services, geostatistical depth conversation, and the complete range of DUG software. DUG is a global enterprise serving customers around the world through offices in Perth, Brisbane, Kuala Lumpur, Jakarta, Singapore, Houston, and London. Looking to take its business to the next level, the company needed improved performance for its software solutions to better serve clients. DUG chose the Intel® Xeon Phi™ coprocessors, which deliver the compute performance to handle its highly parallel technical computing workloads. Challenges • Increase compute performance. Keep pace with rising business demand and increasingly complex seismic algorithms by enabling an increase of 5 to 10 times in compute performance to improve seismic processing, particularly Kirchhoff migrations, reverse time migrations, data interpolation and 3D surface-related multiple elimination (SRME). • Build on current Intel® architecture software investments. Employ advanced compute solutions in seismic processing and imaging without sacrificing existing software investments. • Scale compute without scaling IT infrastructure costs. Lower costs of maintaining IT infrastructure by having a platform that can scale compute resources utilizing existing IT investments. Solution • Deploy SGI® Rackable® cluster powered by Intel® Xeon Phi™ coprocessors and Intel® Xeon® processor E5-2600 v2 product family. Utilizing a platform based on Intel Xeon processor E5-2600 v2 product family and Intel Xeon Phi coprocessors, DUG was able to scale its software to the latest highly parallel architecture for faster seismic processing and imaging while enhancing algorithmic gains in existing software. Technology Results • Utilized high-performance computing. Amplify high-performance optimization techniques of the current software through the high scalability of the Intel Xeon Phi coprocessors. • Harnessed the common code base of Intel architecture. Utilizing Intel Xeon Phi coprocessor-based platform allowed DUG to harness its existing sophisticated library infrastructure without changing the standard programming language. • Enabled transparent and efficient deployment. Utilize the integrated compiler support and execution model to allow end-user processing staff to benefit from transparent deployment. Business Value • Improved node-to-node performance. Achieved an up to 8 times node-to-node 1 performance improvement that enabled faster seismic processing and imaging to meet growing business demands. • Reduced IT maintenance costs. Utilizing highly parallel processing offered by multi-core, multi-threaded CPUs allowed DUG to maintain its existing software Read the full Case Study Intel® Xeon Phi™ Coprocessors High-Performance Computing Energy.