LAKE FOREST, CALIFORNIA (PRWEB) JUNE 20, 2017
PSSC Labs, a developer of custom High-Performance Computing (HPC) and Big Data computing solutions, today announced its work with University of Dayton’s Intelligent Optics Laboratory (IOL) to provide a powerful, turn-key HPC Cluster solution for its atmospheric optics system research.
The IOL investigates the effects of atmospheric disturbances on various laser systems. These include turbulence and other phenomena that cause changes in the refractive index of air, including solar-induced thermal gradients, humidity, precipitation, and heating induced by the laser itself. To accomplish this research, the IOL relies on “Weather Research and Forecasting” (WRF) software paired with input of weather conditions from the US National Oceanic and Atmospheric Administration (NOAA) or similar agencies to provide predictions of temperature, pressure, wind, humidity and related conditions for any area. The widely used software, developed by a cadre of international research teams, is enhanced by techniques developed at North Carolina State University (NCSU) and University of Dayton (UD) to derive very high resolution (“micro-scale”) estimations of these conditions. These calculated conditions are used as inputs to wave optics modeling software developed at UD to provide high-fidelity estimates of the effects of lasers propagated through real-world atmospheric conditions. It is critical that the calculation time is reduced as much as possible to accommodate the high number of input variables.
After receiving a DURIP (Defense University Research Instrumentation Program) grant, the University selected PSSC Labs to manufacture the HPC Cluster system. “It was important for us to select a vendor familiar with the systems requirements of this type of advanced research, that can simultaneously offer the processing power we need while staying within the budget permitted by our research grants,” said Morris Maynard, Sr. Software Engineer at UD. “PSSC offered the best combination of price, software and support available to help advance our research goals.”
The PowerWulf Cluster consists of 768 Intel® Xeon® Processor Cores, 4 Nvidia® Tesla® GPU Adapters, 2.1 TB System Memory, and 40TB+ Storage, all connected via Mellanox® InfiniBand® Interconnects. The PowerWulf Cluster includes PSSC Labs’ CBeST Cluster Management Toolkit to simplify the management, monitoring and maintenance. PSSC Labs will continue to support the HPC Cluster by providing operating system upgrades in order to maintain compatibility with newer versions of software and continued system maintenance.
The new PowerWulf Cluster will effectively reduce the time it takes to run the WRF models from 1-3 weeks to just a few days. In addition, the University will be able to perform analysis locally without transmitting the gigabytes of output from WRF to another location. This improvement will allow the IOL to improve the modeling technique considerably, due to quick turnaround from modification of our code to the output. In addition, it will also improve confidence in the results as researchers are able to observe consistent trends when parameters for a scenario are adjusted across several runs – none of which would be practical on a smaller or slower system.
PSSC Labs’ PowerWulf HPC Cluster offer a reliable, flexible, high performance computing platform for a variety of applications in the following verticals: Design & Engineering, Life Sciences, Physical Science, Financial Services and Machine/Deep Learning.
Every PowerWulf HPC Cluster includes a three-year unlimited phone / email support package (additional year support available) with all support provided by our US based team of experienced engineers. Prices for a custom built PowerWulf HPC Cluster solution start at $20,000. For more information see http://www.pssclabs.com/solutions/hpc-cluster/.
Pulsepoint Saves $300K With PSSC Labs
Learn how leading ad tech company, Pulsepoint, is leveraging PSSC Labs to grow their business and lower their TCO.
Loading a Time Series Database at 100 Million Points Per Second
There are many use cases for time series data, and they usually require handling a decent data ingest rate.