IBM and Stone Ridge Technology claim a ‘performance milestone’ for a reservoir simulation run. The test run was performed using Stone Ridge Technology’s Echelon reservoir simulator to run a billion-cell job in ‘only 92 minutes.’ The hardware was cluster of 30 IBM Power Systems S822LC for HPC servers equipped with 60 Power processors and 120 Nvidia Tesla P100 GPU accelerators. IBM compares this with the previously published ‘record’ run reported by ExxonMobil on the NCSA’s Blue Waters supercomputer that used over 700,000 cores (Oil ITJ Vol 22 N°2). IBM worked with Nvidia to port the code to the GPU architecture and achieve ‘10x the performance and efficiency over legacy CPU codes.’
Stone Ridge president Vincent Natoli said, ‘This demonstrates the computational capability and density of a GPU-based solution. By increasing compute performance and efficiency by over an order of magnitude, we’re democratizing HPC for the reservoir simulation community.’ Sumit Gupta, IBM VP HPC added, ‘By running on IBM Power Systems, users can achieve faster run-times using a fraction of the hardware. The previous record used a supercomputer that occupies nearly half a football field. Stone Ridge did this calculation on a system that could fit in the space of half a ping-pong table.’ IBM and Stone Ridge present their achievement as a victory for ‘fully GPU-based codes’ over ‘legacy CPU codes.’ Your mileage may vary…
© Oil IT Journal - all rights reserved.