Page not found

Tensorflow

Software Support Level: 
Secondary Support
Software Description: 

From the Tensorflow website:

TensorFlow is an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them. TensorFlow was originally developed by researchers and engineers working on the Google Brain Team within Google's Machine Intelligence research organization for the purposes of conducting machine learning and deep neural networks research, but the system is general enough to be applicable in a wide variety of other domains as well.

Software Access Level: 
Open Access
Software Categories: 
Software Interactive/GUI: 
No
Mesabi Documentation: 

Tensorflow is available for use on MSI's Mesabi k40 nodes. To load tensorflow into your environment, use the following command:

module load tensorflow

Tensorflow is frequently used via its python interface. To use tensorflow in this manner after loading in the module, use the following procedure:

python

>>> import tensorflow as tf

There are a variety of tutorials available for using tensorflow on the official website.

HyPhy

Software Description: 

HyPhy stands for hypothesis testing using phylogenies. It is an open-source software package for the analysis of genetic sequences using techniques in phylogenetics, molecular evolution, and machine learning. It also features a complete graphical user interface (GUI) and a rich scripting language for limitless customization of analyses. Additionally, HyPhy features support for parallel computing environments (via message passing interface) and it can be compiled as a shared library and called from other programming environments such as Python or R.

Software Support Level: 
Primary Support
Software Access Level: 
Open Access
Itasca Documentation: 

To launch this software interactively in a Linux environment run the commands:

module load hyphy
HYPHYMPI

It requires a batch file to run the job.  For more detailed instruction on the program, please refer to:

http://hyphy.org/w/index.php/Main_Page

Software Categories: 
Lab Documentation: 

To launch this software interactively in a Linux environment run the commands:

module load hyphy
HYPHYMPI

It requires a batch file to run the job.  For more detailed instruction on the program, please refer to:

http://hyphy.org/w/index.php/Main_Page

Software Interactive/GUI: 
No

D4M

Software Support Level: 
Secondary Support
Software Description: 

D4M attempts to combine the advantages of five distinct processing technologies (sparse linear algebra, associative arrays, fuzzy algebra, distributed arrays, and triple-store/NoSQL databases such as Hadoop HBase and Apache Accumulo) to provide a database and computation system that addresses the problems associated with Big Data.

Software Access Level: 
Open Access
Software Categories: 
Software Interactive/GUI: 
No
General Linux Documentation: 

D4M is a library accessed through Matlab. You must be familiar with Matlab before using D4M. To use D4M on the MSI lab machines, load the matlab module and add the following lines to the top of your Matlab input, or in your startup.m file.

D4M_HOME = '/nfs/soft-el6/d4m/2.0.3'  % SET TO LOCATION OF D4M.
addpath([D4M_HOME '/matlab_src'])     % Add the D4M library.
Assoc('','','')                       % Initialize library.
DBinit

If you have an accumulo database, you can access it from Matlab by defining a connection, e.g.

DB = DBserver('accumulo1:2181','Accumulo','msitest', '','')

Stratus Protected Data Cloud

MSI is building a local research compute cloud environment called Stratus ( https://stratus.msi.umn.edu ) , which is designed to store and analyze protected data, such as dbGaP. Stratus is a subscription-based infrastructure as a service that enables users to operate within their own self-service...

Global Land Model Development

Abstract: 

Global Land Model Development: Time to Shift From a Plant Functional Type to a Plant Functional Trait Approach

This project will advance global land models by shifting from the current plant functional type approach to one that better utilizes what is known about the importance and variability of plant traits, within a framework of simultaneously improving fundamental physiological relations that are at the core of model carbon cycling algorithms. A primary goal for earth system modeling is to make accurate predictions of the future trajectory of the climate system, based on a mechanistic understanding of processes regulating fluxes of mass and energy among system components. Land plays an important role in modifying the earth's mass and energy balance, as a critical link in the global cycling of carbon, among others. Land surface models have developed to include mechanistic representations of vegetation physiology, carbon and nutrient dynamics in plants and soils, how they might respond to changing climate and chemistry, and how those changes might feedback to influence changes in atmospheric greenhouse gases themselves. This project addresses these processes.

Existing models represent the global distribution of vegetation types using the Plant Functional Type concept. Plant Functional Types are classes of plant species with similar evolutionary and life history with presumably similar responses to environmental conditions like CO2, water and nutrient availability. Fixed properties for each Plant Functional Type are specified through a collection of physiological parameters, or traits. These traits, mostly physiological in nature (e.g., leaf nitrogen and longevity) are used in model algorithms to estimate ecosystem properties and/or drive calculated process rates. In most models, 5 to 15 functional types represent terrestrial vegetation; in essence, they assume there are a total of only 5 to 15 different kinds of plants on the entire globe. This assumption of constant plant traits captured within the functional type concept has serious limitations, as a single set of traits does not reflect trait variation observed within and between species and communities. While this simplification was necessary decades past, substantial improvement is now possible. Rather than assigning a small number of constant parameter values to all grid cells in a model, procedures will be developed that predict a frequency distribution of values for any given grid cell. Thus, the mean and variance, and how these change with time, will inform and improve model performance.

The trait-based approach will improve land modeling by: incorporating patterns and heterogeneity of traits into model parameterization, thus evolving away from a framework that considers large areas of vegetation to have near identical trait values; utilizing what is know about trait-trait, -soil, and -climate relations to improve algorithms used to predict processes at multiple stages; and allowing for improved treatment of physiological responses to environment (such as temperature and/or CO2 response of photosynthesis or respiration).

Return to this PI’s main page.

Group name: 
reichpb

Data Retention and Protection

The following policies pertain to specific systems managed by MSI. The specific elements of this policy and MSI's data policies in general are consistent with the University policy on data management and are therefore applicable to the transfer and storage of data on MSI resources. */ Data...

A New Metal-Organic Framework for Catalysis

Chemists are often interested in developing new catalysts that will improve the efficiency of chemical reactions. They can look to nature to provide examples when designing these new materials. Metalloporphyrins are a class of metal complexes that appear a great deal in biological systems. These...

2018 Research Exhibition

MSI held the ninth annual Research Exhibition on April 17, 2018. The event included a panel discussion about careers in HPC-related fields, presentations by industry representatives about current and emerging technologies, and a judged poster session. A new feature this year was the chance to see...

4/9/13: MSI PI Cramer Named CSE Associate Dean

Professor Chris Cramer ( Fellow , Chemistry ) has been named the College of Science and Engineering’s new associate dean for academic affairs, the college announced on April 9, 2013. The appointment will become effective on July 1, 2013. You can read the announcement on the chemistry department’s...

Interview: Makenzie Provorse

Makenzie Provorse is a graduate student in Professor Jiali Gao’s ( MSI Fellow ) group in the Department of Chemistry . She entered the University of Minnesota in the fall of 2009 and joined the Gao group in January 2010. She’s been using MSI since then. Ms. Provorse was a finalist at the 2013 MSI...

Pages