Page not found

Ceph in HPC Environments at SC17

Overview Individuals from MSI , RedHat Inc. , Indiana University (Jetstream), Laureate Institute for Brain Research , and Monash University came together at SC17: The International Conference for High Performance Computing, Networking, Storage and Analysis on Tuesday, November 14, 2017 in Denver,...

Ceph in HPC Environments at SC18

Overview Individuals from several organizations will come together at SC18: The International Conference for High Performance Computing, Networking, Storage and Analysis in November, 2018 in Dallas, Texas to share their experiences with Ceph in HPC Environments. This "Birds of a Feather (BoF)"...

Ceph in HPC Environments at SC16

Overview Individuals from MSI , RedHat Inc. , Intel Corp ., Indiana University (Jetstream), Laureate Institute for Brain Research , University of Michigan Advanced Research Computing Technology Services came together at SC16: The International Conference for High Performance Computing, Networking,...

Computational Fellowships Available

posted on December 4, 2014 Two opportunities in high-performance computing research are now accepting applications for their programs in 2015. The Computational Physics Student Summer Workshop will be hosted by Los Alamos National Laboratory’s Computational Physics Division, June 8 - August 14,...

Materials Processing Fundamentals

<h3 class="red">Materials Processing Fundamentals</h3><p>These researchers develop, implement, and apply large-scale numerical simulations to understand processes that are used for the production of advanced solid-state materials. Of particular interest are processes employed for the growth of large, single crystals. These studies aim to understand factors that directly affect materials properties and production costs. The researchers have studied the growth of electronic and photonic inorganic materials via a number of melt and flux growth methods, as well as organic crystallization from the solution phase. Current materials of interest include silicon, sapphire, and bulk gallium nitride that are important for energy applications (photovoltaic and LED devices), as well as semiconductor and scintillator crystals for radiation detectors employed in national security and medical imaging applications. Process models are developed using the mathematical depiction of continuum transport (incompressible flows, heat and mass transfer), phase-change, and other interfacial phenomena. These models are solved using finite element methods, using both two-dimensional and three-dimensional implementations for steady-state and transient analyses. Bifurcation analyses, employing continuation methods, form a paradigm to understand process behavior. A continuing challenge is the development of effective and robust parallel implementations for the accurate solution of three-dimensional, incompressible flows that are characteristic of the crystal growth systems of interest. The researchers are developing and employing parallel, finite element models for these purposes.</p><p>Return to this PI&#39;s <a href="">main page</a>.</p>
Group name: 

Sparse Coding Algorithm for Image Searching

Nearest Neighbors (NN) is a fundamental operation in many areas of scientific computing, including computer vision, machine learning, robotics, and data mining. It is the backbone of applications people use every day, such as Google Images. Images tend to be high-dimensional, and as the...

The Impact of Credit and Assets on Unemployment

<h3 class="red">The Impact of Credit and Assets on Employment</h3><p>Unemployed households&#39; access to unsecured revolving credit (credit cards) increased from 13 percent to 45 percent over the last three decades. This project analyzes how this large increase in revolving credit has impacted the business cycle. The main quantitative result is that business cycles are deeper and employment recoveries are exacerbated in periods of persistently increasing credit access, but in the long run, when credit growth slows, business cycles return to normal when measured by the timing and depth of the trough of employment. &nbsp;</p><p>This project includes dynamic programming problems that are designed to capture the decision process of an unemployed household. The researchers then use Monte Carlo methods to recover the underlying distributions. Both the dynamic programming (value function iteration) and the Monte Carlo simulations require the use of the supercomputers.</p><p>Return to this PI&#39;s <a href="">main page</a>.</p>
Group name: 

Computational Fluid Dynamics of Wind and Water Waves for Environmental and Energy Applications

<h3 class="red">Computational Fluid Dynamics of Wind and Water Waves for Environmental and Energy Applications</h3><p>These researchers use MSI computing resources to perform high-resolution simulations of wind and water wave flows using high-fidelity computational fluid dynamics. Water waves, wind, and their interactions are important to many applications, including atmosphere-ocean CO<sub>2</sub> exchange in the study of global climate change, offshore wind energy and wave energy, and the trajectory and fate of pollutants at water surface. &nbsp;Recently, the group has also received funding to study oil spills at sea and in the Great Lakes.</p><p>This research uses novel simulation methods developed in the group. Their in-house simulation codes include a high-order spectral method for waves and large-eddy simulation of wind turbulence on wave surface-fitted grid. The simulations resolve wave phases, a feature fundamentally distinct from and has a clear advantage over traditional approaches that are spectral and wave-phase-averaged, in which the wave phase information is lost. Because the flow physics are resolved in a more direct way and with much more details than in previous methods, this study will have a better chance to succeed.</p><p>The researchers perform simulations of the wind and wave fields, which involve massively parallel computing and datasets with unprecedented volumes and details. They will address the multi-scale wind and wave fluid dynamics through computations of ocean wave field at relatively large scales (100 km domain size, 5 m resolution for waves) and wind field at relatively small scales (2 km domain size, 2 m resolution for turbulence eddies). The big data from the simulations will establish a physical basis for the mechanistic study of the complex dynamic system of the ocean waves and wind.</p><p>Return to this PI&#39;s <a href="">main page</a>.</p>
Group name: 

MSI Upgrades Storage Solutions

One petabyte of tape storage for under $10,000? Leveraging new advances in tape media, the Minnesota Supercomputing Institute has expanded its storage portfolio to offer easy access to deep archival storage at low cost. MSI currently offers a high-performance storage solution from Panasas (3.2 PB...