Page not found

Computational Mechanics and Multi-Disciplinary Applications to High Performance Supercomputing Environments

Abstract: 

Computational Mechanics and Multi-Disciplinary Applications to High Performance Supercomputing Environments

This project is concerned with the development of unified computational methodologies, solution algorithms, and finite element modeling/analysis strategies for rigid-flexible multi-body dynamics, contact-impact-penetration, electromagnetics, multi-disciplinary flow-thermal-structural problems, and micro/nano-scale effects in heat conduction. The philosophy and rationale of this work is based on employing a common numerical methodology for each of the individual disciplines in conjunction with common computational algorithms for applicability to supercomputing systems in solving large-scale engineering problems. Various research activities include development of new time integration computational algorithms for transient/dynamic/contact/ impact/damage/penetration problems; development of effective finite element based methodologies, which can be used in multi-disciplinary problems; new physically correct contact models for penetration and impact problems; application of finite element methods in the manufacturing simulations to provide a paradigm for Virtual Manufacturing and the simulation of Virtual Experiment and Virtual Testing. The application areas include a wide range of engineering problems involving multi-physics and space/time domain decomposition with interface to graph partitioning techniques. The overall efforts focus attention on providing new and effective approaches for not only improving the existing capabilities for applicability to supercomputing environments, but also towards providing an accurate understanding of the physics and mechanics relevant to multi-disciplinary engineering problems.

Return to this PI's main page.

Group name: 
tammakk

Employment

MSI provides high-performance computing resources and research computing support to the University of Minnesota and other Minnesota institutions of higher learning. We offer employees the opportunity to work with cutting-edge technology and state-of-the-art research in a fast-paced environment. MSI...

2012 Summer Workshops in High-Performance Computing

The Virtual School of Computational Science and Engineering (VSCSE), a national virtual education organization partially funded by the Great Lakes Consortium for Petascale Computation , is offering three courses during Summer 2012 . These courses, which include instruction in programming for many-...

Facilities Overview (Full)

Established in 1983, the Minnesota Supercomputing Institute (MSI) is the University of Minnesota's principle center for computational research. MSI provides services to over 560 active groups that sponsor more than 3,300 unique users from 19 different university colleges, maintaining an array of...

High-Performance De Novo RNA-Transcript Reconstruction Leveraging Distributed Memory and Massive Parallelization

Abstract: 

High-Performance De Novo RNA-Transcript Reconstruction Leveraging Distributed Memory and Massive Parallelization

These researchers are working to optimize the performance of the Trinity RNA-Seq de novo assembly software. This project exemplifies collaborative software development between industry and academia to tackle computational challenges in manipulating large volumes of next-gen sequence data and to leverage advances in algorithm development and compute hardwareThree versions of Trinity's Inchworm computationally intensive part (one that is based on the original OpenMP version, and two new versions that are based on MPI and on Fortran2008) are integrated into the Galaxy web interface.

A bibliography of this group's publications acknowledging MSI is attached.

Group name: 
sosac
Attachment: 

High-Performance and Big Data Research

Abstract: 

High-Performance and Big Data Research

This group's research during 2015 focused on the development of parallel shared-memory graph partitioning, ordering, and clustering algorithms that use the multilevel paradigm. Graph partitioning is used widely for parallel task scheduling and data distribution. Graph ordering is used reducing the amount of computation and memory required for sparse direct numerical methods. Graph clustering is a widely used technique for discovering relationships between data points by creating groups of unconstrained size with high internal connectivity. Access to MSI's HPC resources has been critical in the development of these algorithms as evaluating the scalability of the algorithms requires machines with a large number of compute cores, and many of the graphs/matrices in these domains reach massive size, requiring large amounts of memory.

The group's work in 2016 focuses on developing hybrid shared/distributed memory codes that can effectively utilize compute architectures composed of many multicore nodes. This work will be an extension of the researchers' past work on shared and distributed memory graph partitioning, ordering, and clustering. Part of this will include ensuring their methods scale to very large numbers of processing cores. These methods will be required for partitioning, ordering, and clustering problems on the next generation of large petascale and exascale machines, which will have millions of processing cores.

Return to this PI's main page.

 

Group name: 
karypisg

Development of High-Performance Methods for Spanning Multiple Length and Time Scales

Abstract: 

Development of High-Performance Methods for Spanning Multiple Length and Time Scales

This project focuses on the development and application of high-performance methods for spanning multiple length and time scales in atomistic simulations. Efforts will focus on a number of directions:

  • Development of a high-performance 3D implementation of the spatial multiscale Quasi-Continuum (QC) method that greatly reduces the computational cost of atomistic simulations by only retaining atomistic resolution where necessary and using a continuum approximation elsewhere. MSI resources are used to test different parallelization strategies and to perform QC production runs in a project related to the fracture of silicon MEMS devices.
  • Study of the fracture of single and polycrystalline silicon samples. This includes both practical aspects of fracture of silicon fabricated devices such as MEMS devices as well elucidation of the fundamental physics of dynamic fracture. Studies will include both molecular dynamics (MD) simulations as well as QC3D simulations as noted above.
  • Development of a method within the Knowledgebase of Interatomic Models (KIM) project for assessing the transferability of interatomic potentials used in atomistic and multiscale simulations by comparing their predictions to density functional theory (DFT) calculations. MSI resources are used to perform DFT calculations to obtain high quality reference data.
  • Development of MD simulations of interpenetration at polymer interfaces to better understand the role of interface structure on polymer adhesion. Both all-atom and coarse grained (multiscale) simulations will be performed.

A Research Spotlight featuring the group's work appeared on the MSI website in July 2014.

Return to this PI’s main page.

Group name: 
luskin

MSI Users Bulletin - March 2017

The Users Bulletin provides a summary of new policies, procedures, and events of interest to MSI users. It is published quarterly. To request technical assistance with your MSI account, please contact help@msi.umn.edu . 1. New Storage Limits: As announced in an email from Director of Research...

Computer Simulations of Hydraulic Jumps

A hydraulic jump is a phenomenon that occurs when a high-velocity liquid transitions to a region where the liquid slows. The fluid rises in height, turbulence increases, and air is entrained into the flow forming bubbles after the transition. Energy dissipation and higher rates of gas transfer also...

Pages