Innovation, Opportunity and You
I'm looking for a few exceptional people to augment my team to assemble the best and brightest in high performance numerical simulation for reservoir studies. If you’d like to be included please be in touch with me immediately.
Stone Ridge Technology is growing! Since the downturn subsided at the end of 2016, SRT has fielded steadily increasing interest in ECHELON, its innovative high-performance GPU based reservoir simulator. The capstone example is the major deal recently signed with Eni. In addition to Eni, ECHELON has been adopted by several other companies and is in trial in numerous large and small organizations. Interest has spiked upwards as word of mouth has spread about ECHELON’s speed, accuracy and scalability and there is genuine excitement about the use of leading edge GPU hardware which has gained rapid acceptance recently as a next generation HPC platform.
As companies enter this new era of oil and gas exploration and development, leaner and more focused, they are looking for new technologies that can give them an edge; technologies that will scale with new hardware; technologies that can complement their digital strategies with big data and machine learning; technologies that can dramatically accelerate their workflows making better use of their most valuable resources…their employees. ECHELON fits the bill.
As a consequence, we have a lot of work to do at SRT and we need more people to do it! We have immediate opportunities for core simulator developers, a product manager, user interface developers, and a devops engineer.
Simulator developers require deep experience in C++ and large scale parallel coding(CUDA/MPI/OMP). They typically have PhD’s in applied math, physics or engineering. The ideal candidate would be a highly skilled individual that has experience developing a numerical reservoir simulator.
The Product manager is a key position and we are looking for someone who has deep experience in reservoir simulation and engineering. The individual should have experience with simulator workflows in reservoir studies as well as an informed vision for their evolution. He or she should have excellent organizational and communication skills and the ability to work with a diverse, highly capable group of developers. The ideal candidate will also have experience developing high performance technical software.
I believe that SRT is at a transitional moment with a unique opportunity. There are prominent cracks in the legacy workflow software that has dominated the energy industry for decades. We hear it all the time from clients in the industry who lament performance shortcomings, usability issues, inflexible licensing models and old technologies. We sense an industry wide hunger for new solutions and novel ideas based on current and emerging technologies. Reservoir simulation is just one example, albeit one with which SRT is very familiar. For many years now, practitioners have wanted to run ensembles of models to more accurately capture the inherent uncertainty of subsurface properties. They have also wanted to run larger and higher resolution models. However, it’s difficult to imagine running thousands of realizations of a model in an ensemble when a single model takes three days or for that matter to consider resolutions in the hundreds of millions of cells when software grinds to a halt at ten million cells. ECHELON’s speed and scalability change this equation. Recently Eni announced a breakthrough calculation they did using their HPC4 supercomputer and ECHELON. They ran 100,000 realizations of a 5.7 million cell model on HPC4 in 15 hours and 28 minutes, a feat that would have taken the industry leading simulator 30 days. A year ago, we ran a billion cell model in 90 minutes on 30 IBM Power nodes forecasting 45 years of production. Imagine what can be done on models with sizes that are more typically used in the industry.
Figure 1 SpecINT(x1000) benchmark trend vs CPU and GPU generations. (http://bit.ly/karl-rupp-performance-trend)
While ECHELON has reaped the benefit of Moore’s law progression on GPUs (see Fig. 1) its competition on CPU architectures has largely languished with weak to modest gains. In three to four years I expect that same billion cell calculation to be down to 4 nodes and 15 minutes. GPU’s, once used only in visualization are playing important roles in seismic processing, deep learning and now, with ECHELON, in reservoir simulation.
SRT has the mission and resources to create the next generation reservoir simulator. We are starting from a foundation that is rooted in the newest and most rapidly advancing hardware technology available. I'm looking for a few exceptional people to augment my team to assemble the best and brightest in high performance numerical simulation for reservoir studies. If you’d like to be included please be in touch with me immediately.