Loading…
RMACC 2021 HPC Symposium
Attending this event?

Sign up or log in to bookmark your favorites and sync them to your phone or calendar.

Tuesday, May 18
 

1:00pm MDT

Fundamentals of HPC
Getting started on a High Performance Computing cluster can be a daunting task for the uninitiated. Join us this two-hour course where we will alleviate anxiety by going over the basics! This hands-on tutorial will cover Linux commands, navigation of a Linux file-system, the Linux environment, creation of bash scripts, job scheduling and more!

Speakers
AM

Andy Monaghan

University of Colorado Boulder


Tuesday May 18, 2021 1:00pm - 3:00pm MDT
 
Wednesday, May 19
 

9:30am MDT

The role of massively parallel computing in personalized blood flow modeling
The recognition of the role hemodynamic forces have in the localization and development of disease has motivated large-scale efforts to enable patient-specific simulations. When combined with computational approaches that can extend the models to include physiologically accurate hematocrit levels in large regions of the circulatory system, these image-based models yield insight into the underlying mechanisms driving disease progression and inform surgical planning or the design of next generation drug delivery systems. Building a detailed, realistic model of human blood flow, however, is a formidable mathematical and computational challenge. The models must incorporate the motion of fluid, intricate geometry of the blood vessels, continual pulse-driven changes in flow and pressure, and the behavior of suspended bodies such as red blood cells. In this talk, I will discuss the development of HARVEY, a parallel fluid dynamics application designed to model hemodynamics in patient-specific geometries. I will cover the methods introduced to reduce the overall time-to-solution and effectively leverage leadership class systems. Finally, I will discuss the role of personalized blood flow models for applications ranging from treatment planning for cardiovascular disease to studying cancer cell adhesion and metastatic progression.

Speakers
AR

Amanda Randles

Duke University


Wednesday May 19, 2021 9:30am - 10:15am MDT

10:30am MDT

Intel Developer Tools: oneAPI, Vendor Neutral Programming Model for Heterogeneous Programming- Intel Sponsored session
This talk will discuss oneAPI, Data Parallel C++ (DPC++) concepts and anatomy using a simple DPC++ program, the compilation and runtime model of DPC++ application. This session will also demonstrate the Data Parallel C++ Compatibility Tool which helps with migrating CUDA applications to DPC++.

Wednesday May 19, 2021 10:30am - 12:00pm MDT

10:30am MDT

Parallel Python MPI4PY
MPI for Python (mpi4py) provides bindings of the Message Passing Interface (MPI) standard for the Python programming language, allowing Python programs to exploit multiple processors. MPI is the primary standard for programming large parallel machines but is also applicable to multicore laptops. This talk will cover basic programming in mpi4py including covering the two basic styles of mpi4py calls, OO and traditional. We will provide examples of the most common MPI calls. We will look at using mpi4py to do Bag of Task parallelism both using a “home grown” technique and the built in methodology provided as part of the mpi4py module. Recipes will be provided for creating a mpi4py install with and without conda.

Speakers

Wednesday May 19, 2021 10:30am - 12:00pm MDT

10:30am MDT

Sneak peak into the Dell Technologies HPC and AI innovation lab- Dell Sponsored session
HPC Technology preview, powered by Intel – Deepthi Cherlopalle, Senior HPC Engineer, Dell Technologies
In this session we present the performance results of Intel’s next generation processor, new platform offerings from Dell EMC, also Lab updates on the Zenith
 
What is DAOS? – Ari Martinez, Dell Technologies
The Distributed Asynchronous Object Storage (DAOS) is an open-source software-defined object store designed from the ground up for massively distributed Non-Volatile Memory. DAOS takes advantage of next generation NVM technology like Storage Class Memory (SCM) and NVM express (NVMe).
 
Covid-19, 1 year later – A customer prospective with James Lowey, CIO TGen
In this session we hear from James Lowey of TGen (TRANSLATIONAL GENOMICS RESEARCH INSTITUTE) for his perspective on Covid-19 research as compared to 1 year ago. We heard from James last year and are very interested to see over time how things may have changed with regards to research and how his Intel powered HPC Cluster has enabled positive outcomes in these troubled times.

Speakers
DC

Deepthi Cheriopalle

Senior HPC Engineer, Dell Technogies
AM

Ari Martinez

Dell Technogies
JL

James Lowey

CIO, TGen


Wednesday May 19, 2021 10:30am - 12:00pm MDT

10:30am MDT

High-performance computing for Scientific Visualization
Visualization is an essential part of scientific discovery that enables novel analysis, confirms
theoretical models, allows comparisons, facilitates the recognition of trends and patterns, and
ultimately helps draw conclusions. As a result of these benefits, High Performance Computing
(HPC) is integrating visualization to enhance understanding in big-data analysis and large-scale
simulations. This workshop on high-performance computing for scientific visualization provides
a venue for presenters to share and exchange practical expertise and experience with processes,
workflows, and applications. The workshop plans to bring together a wide-ranging audience of
students, visualization and computational scientists, and simulation developers involved in
developing, deploying, or maintaining scientific visualization approaches on HPC infrastructures.
The workshop's primary focus is to provide practical insights that inspire attendees to develop,
refine, and avoid pitfalls using HPC for scientific visualization.

Areas of interest for this workshop include, but not limited to:
• Visualization using HPC infrastructure: information visualization, scientific visualization,
visual analytics, immersive visualization, in situ visualization.
• Best practices, Opportunities / Gaps while using HPC for scientific visualization
• Tools and techniques for HPC to enable/support scientific visualization
• Case studies of solving a specific scientific visualization challenge using HPC
• Simulation and Modeling using HPC
• Data models to enable scalable visualization using HPC
• Visualization of high-dimensional data using HPC

Moderators
RK

Rajiv Khadka

Idaho National Lab

Wednesday May 19, 2021 10:30am - 12:00pm MDT

1:00pm MDT

Student Poster Presentations
Join us for student poster presentations, hear a short introduction to each poster and then visit their individual posters to ask questions!

Wednesday May 19, 2021 1:00pm - 1:30pm MDT

1:00pm MDT

JupyterLab 3.0's New Integrated Debugger
Jupyter notebooks have become the de facto standard for scientific analysis and data science. JupyterLab is the "Integrated Development Environment" for notebooks, markdown text files and more. Every good IDE has a debugger and with the introduction of version 3.0, JupyterLab now has an integrated debugger, provided the proper Python kernel is used. This tutorial will briefly discuss JupyterLab, extensions, and kernels and then demonstrate upgrading to the proper version, installing the proper Python kernel, and debugger functionality. If enough time at the end, alternative debugging methods, such as Visual Studio Code, can be discussed.

Speakers
MB

Mike Busch

Regis University


Wednesday May 19, 2021 1:00pm - 1:30pm MDT

1:00pm MDT

SysAdmin Meetup
Join the RMACC SysAdmin community for an open discussion to talk about what's going on throughout the region. 

Moderators
Wednesday May 19, 2021 1:00pm - 1:30pm MDT

1:35pm MDT

Containerization on RMACC Summit with Singularity
Would you like to be able to move, share or publish your software without worrying whether it will work properly on other computers? Containerization is an OS-level virtualization method used to deploy and run distributed applications without launching an entire virtual machine for each app. Containers are "stand-alone", enabling workflows to be easily packaged and shared across platforms, enhancing reproducibility and portability. This hands-on tutorial by CU Research Computing staff will provide an introduction to containers and the basics of building and running containers with Singularity. Singularity is containerization software used on high performance computing platforms where users do not have escalated privileges.

Speakers
AM

Andy Monaghan

University of Colorado Boulder


Wednesday May 19, 2021 1:35pm - 3:05pm MDT

1:35pm MDT

Fundamental Machine Learning Pipeline applied to Atmospheric Science Data
Multiple different atmospheric science fields have benefitted from the rapid expansion of machine learning and deep learning model capabilities within the past decade. Differing from traditional prediction problems, meteorological tasks require specific data pre-processing and model architecture needs. This tutorial uses numerical weather prediction model output to describe fundamental machine learning principles for both traditional modeling (such as random forests, logistic regression, etc.) and deep learning. Fundamental principles include data-preprocessing, feature selection, in addition to model training, tuning, and evaluation. The hands-on lecture will explore linear models as well as neural networks, with resources on other modeling types.

Speakers
AB

Amanda Burke

University of Oklahoma


Wednesday May 19, 2021 1:35pm - 3:05pm MDT

1:35pm MDT

Paraview Animation, Scripting and Post Processed Video Production with DaVinci Resolve
ParaView is an open-source, cross-platform application used to explore, analyze and visualize scientific data. These processing capabilities can be done either with Paraviews interactive gui or by using the vastly superior method of python scripting.

DaVinci Resolve is the TV, Streaming and Movie Industries leading post compositing, color grading and all-round film editor and it is FREE!

In this hands-on workshop we will look at animation methods both straight from the Paraview GUI interface but more heavily with pvPython. We will also look at the various new rendering methods and how to know which you should use for your needs. This includes the standard rendering engine as well as the OSPRay and IndeX renderers. We will quickly look at the built-in ability to create interactive VR (virtual reality) scenes for use with popular VR headsets.

We will follow this up with how to ‘best’ turn your work into a powerful video ready for presentations, YouTube, TV and the universe at large!

Skill Level: This is an intermediate level-oriented coverage of Paraview. No previous knowledge of Paraview or Python is required but will be very beneficial as we will assume you are comfortable with the base concepts. Absolutely no knowledge of DaVinci Resolve is needed to turn your work into a masterpiece quickly and easily!

Requirements: This is a hands-on tutorial. If you want to follow please already have Paraview version 5.9 installed on your laptop. Paraview can be freely downloaded from: Paraview 5.9.

Also download and install the free version of DaVinci Resolve 17 which can be freely downloaded from (scroll all the way to the bottom!)
DaVinci Resolve 17
If you only want to follow along with one section, I HIGHLY RECOMMEND following along with the Resolve portion. If you work with animations this will revolutionize your world!

We will not take time to deal with installation issues once the workshop starts.
Help services via CRDDS Consulting Hours will be made available for people needing help setting up before the tutorial by submitting a free Consultation request form at: CRDDS Consultation Hours. Please note that you desire help or have questions on either Paraview and/or Resolve.

Speakers
TD

Tim Dunn

University of Colorado Boulder


Wednesday May 19, 2021 1:35pm - 3:05pm MDT

1:35pm MDT

QuSP: The Quantum Simulation Package
The Quantum Simulation Package (QuSP) is an open browser-based development environment that greatly simplifies quantum simulation. In this workshop, attendees will learn about how QuSP's application module system works with Jupyter and Colab to analyze quantum lattice systems, and how these tools can augment quantum education. Analyzing quantum lattice systems in QuSP can be done in two ways. The first, utilizes QuSP's fully templated application modules which allow users to tune model parameters for several predefined systems. The second, is a fully integrated Jupyter environment containing common software libraries for building quantum simulations. Attendees will interactively walk through a sequence of tutorials representing a typical QuSP workload: building and executing application modules, and debugging and profiling code to be executed in application modules.

This work was funded by the NSF and supported by the Colorado School of Mines High Performance Research & Computing center (CSM HPC).

Speakers
LC

Lincoln Carr

Colorado School of Mines
MJ

Matt Jones

Colorado School of Mines


Wednesday May 19, 2021 1:35pm - 3:05pm MDT

1:35pm MDT

Topics in HPC- HPE sponsored session
Join HPE to hear about four exciting topics in the HPC World!

Exascale Era Technologies from HPE- Bill Mannel
HPC in Space – Spaceborne Computer missions on the International Space Station- John Kichury
Parallel HPC & AI Storage- Uli Plechschmidt
SLINGSHOT: Next Generation Fabric- Jesse Trager

Speakers
BM

Bill Mannel

HPE
Bill manages the High Performance Computing Business Unit in the HPC/MCS Global Business Unit. Bill joined HPE from SGI in 2014 and is a seasoned veteran of the servers and high performance computing industry. Prior to SGI, Bill worked at NASA and the U.S. Air Force on aircraft programs... Read More →
JK

John Kichury

HPE
HPC Solutions Architect with Hewlett Packard EnterpriseMember of the Spaceborne Computer team. Developed methods and software for management and control, autonomous HPC system operation/error detection and recovery remotely operating on the ISS.46+ years in development of hardware/electronics... Read More →
UP

Uli Plechschmidt

HPE
Uli leads the product marketing function for parallel high performance computing (HPC) and artificial intelligence (AI) storage at Hewlett Packard Enterprise (HPE). He joined HPE in January 2020 as part of the Cray acquisition. Prior to Cray, Uli held leadership and expert roles in... Read More →
JT

Jesse Treger

HPE
Jesse Treger is responsible for strategy and business development for the HPE Slingshot high performance interconnect. Jesse came to HPE through the acquisition of Cray Inc. where he was Chief of Staff for the Product Management Organization. Prior to joining HPE/Cray, Jesse held... Read More →


Wednesday May 19, 2021 1:35pm - 3:05pm MDT

3:15pm MDT

Cyberinfrastructure knowledge, practices, and challenges among researchers in RMACC
In this 1-hour session we present the results of an online survey distributed to RMACC researchers in 2020 to gather information about their knowledge, practices and challenges pertaining to the use of advanced cyberinfrastructure, including computation, software, data visualization, data storage, archival, sharing, publication and user support.

Speakers
AM

Andy Monaghan

University of Colorado Boulder
MS

Mara Sedlins

Colorado State University
BM

Brett Milash

University of Utah


Wednesday May 19, 2021 3:15pm - 4:15pm MDT

3:15pm MDT

Evidence-based practices to support a more inclusive RMACC
It is evident that diverse and inclusive workplaces are the best place for both the people who work there and the most successful achievement of organizational goals. But the most common question during conversations about diversity is, "HOW do we make progress?" This session will be a conversation about evidence-based practices we can each take on to make our universities and organizations more welcoming for people from all backgrounds.

Speakers

Wednesday May 19, 2021 3:15pm - 4:15pm MDT

3:15pm MDT

Python Multiprocessing Demonstration on I/O and CPU Bound Tasks
Python is often disparaged as being a "slow" language, due to being an interpreted language and having a "Global Interpreter Lock" that throttles multithreading through the CPU. This session will examine ways to leverage concurrent and multithreaded techniques for the i/o bound process of reading emails from the Enron corpus and storing them in a SQLite database. This session will also examine multiprocessing in the CPU bound computer vision task of extracting frame and sensor information from a Microsoft Azure Kinect.

Speakers
MB

Mike Busch

Regis University


Wednesday May 19, 2021 3:15pm - 4:15pm MDT
 
Thursday, May 20
 

9:30am MDT

How HPC enables us to build trustworthy AI for high-impact weather
In September 2020, NSF announced a large national investment in multiple national AI institutes across the US. These institutes each focus on different foundational aspects of AI and many require extensive HPC to facilitate their research. Our institute, the NSF AI Institute for Research on Trustworthy AI in Weather, Climate, and Coastal Oceanography, is focused on creating trustworthy AI for a variety of applications in environmental science. In this talk, I will give a brief overview of our institute and then focus on how HPC is facilitating our work in creating trustworthy AI for weather applications.

Speakers
AM

Amy McGovern

University of Oklahoma


Thursday May 20, 2021 9:30am - 10:15am MDT

10:30am MDT

Empower Research Computing at your Organization Through the Open Science Grid
The Open Science Grid, or OSG, is a worldwide network of computing power contributed by colleges, national labs, and other research-supporting institutions. Researchers at these types of institutions in the United States can access and use the OSG’s Open Science Pool, with capacity from more than 100 sites to drive scientific research. This tutorial will present an overview of OSG services, include hands-on exploration of OSG capabilities, and describe different pathways for engaging with the OSG as an organization.

Speakers
CK

Christina Koch

University of Wisconsin-Madison
LM

Lauren Michael

University of Wisconsin-Madison


Thursday May 20, 2021 10:30am - 12:00pm MDT

10:30am MDT

Intel Developer Tools: oneAPI overview, Data Parallel C++ Best Practices, Targeting Xeon, and NVIDIA GPUs
This talk will provide a short overview of oneAPI and go over more advanced features of DPC++ which includes subgroups, collectives like reductions, targeting NVIDIA GPUs using open source DPC++ Compiler. Will briefly touch upon the best practices recommended to program multi-tile GPUs as well.

Thursday May 20, 2021 10:30am - 12:00pm MDT

10:30am MDT

Quickly stand-up (and tear-down!) software-defined clusters using Omnia- Dell Sponsored session
Omnia is a new open-source cluster deployment toolkit that automates the stand-up and tear-down of HPC clusters for simulation, artificial intelligence, and data analytics using various server, storage, and network building blocks. In this tutorial we will give a brief introduction to the Omnia project and Omnia's current capabilities. We will then walk through the bootstrapping of an Omnia deployment appliance and creation of a resource pool from which we will stand-up and tear-down several high-performance clusters running the Slurm job scheduler and the Kubernetes container orchestration engine. We will wrap-up with a feature roadmap and ways to get involved with the Omnia project.

Speakers
LW

Luke Wilson

Chief Data Scientist, Dell Technogies


Thursday May 20, 2021 10:30am - 12:00pm MDT

10:30am MDT

Virtual/Augmented Reality: Applications and its future
Virtual/Augmented Reality (VR/AR) have been around for more than two decades. However, it
was only recently with the advancement of technology and hardware equipment’s that the use of
VR/AR has found its way into consumer and commercial world. Recent trends have shown there
is an increasing interest of using VR/AR in diverse areas of fields ranging from visualization,
education, medical, training, simulation, engineering, gaming, and more. This workshop on the
use of virtual and augmented reality and its application provides a venue for presenters to share
and exchange practical expertise and experience with processes, workflows, and applications. The
workshop plans to bring together a wide-ranging audience of students, VR/AR developers,
visualization scientists, and simulation developers involved in developing, deploying, or
maintaining the use of VR/AR in their work. The workshop's primary focus is to provide practical
insights that inspire attendees to develop, refine, and avoid pitfalls using VR/AR in their work.
The workshop comprises of 2-min fast forward pitch, 10-minute presentations, and an open discussion
session at the end for all attendees.

Areas of interest for this workshop include the use of VR/AR, but not limited to:
• Information Visualization
• Immersive Data Visualization
• Digital Twins
• Trainings
• Education
• 360 data capture tools and technologies
• Applications for teaching and research
• Medical
• Virtual Worlds
• 3d Interaction Techniques
• Remote Telecollaboration
• Best practices and Opportunities
• Game Development


Thursday May 20, 2021 10:30am - 12:00pm MDT

1:00pm MDT

Career opportunities and Resume tips and tricks
Learn about career opportunities and internships in the world of HPC. Come with questions about how to join this field of employment and get tips about writing resumes and cover letters.

Speakers

Thursday May 20, 2021 1:00pm - 1:30pm MDT

1:00pm MDT

Cloud Infrastructure for Research
Jonathon Anderson will briefly present efforts to deploy CUmulus, a cloud infrastructure service in the CU Boulder Research Computing environment, and lead a two-way question and answer session on the topic of cloud services deployed in support of research. Bring your questions, and he’ll bring some of his own!

Speakers
JA

Jonathon Anderson

University of Colorado Boulder


Thursday May 20, 2021 1:00pm - 1:30pm MDT

1:00pm MDT

Open OnDemand Working Group Discussion
Join the RMACC Open OnDemand community for a discussion about how local institutions are using Open OnDemand and new developments in the field.

Speakers
MC

Martin Cuma

University of Utah


Thursday May 20, 2021 1:00pm - 1:30pm MDT

1:00pm MDT

Outreach and Connection points in the RMACC region
Join RMACC leadership for a discussion of best practices regarding outreach to current partners and smaller schools. Let's discuss how to build the RMACC community and improve these connection points!

Speakers
SK

Shelley Knuth

University of Colorado Boulder
BY

Becky Yeager

University of Colorado Boulder


Thursday May 20, 2021 1:00pm - 1:30pm MDT

1:35pm MDT

Programming GPUs using OpenACC directives
With many new supercomputers adapting General Purpose Graphical Processing Units (GPGPUs) to move towards exascale computing, it is important to port, optimize, and scale the HPC workloads. Understanding the GPU architecture, terminology, and programming GPUs with minimal effort is a good skill to have. Of many frameworks that help port the applications to GPUs, OpenACC is fairly simple and easy to learn. OpenACC is a directive based language which helps us to port the code onto CPU+GPU architectures with minimal efforts.
This tutorial session will begin with a little bit of Introduction to GPU architecture and terminology, a few basic examples of how to port code using OpenACC directives which includes both single GPU and Multi GPU. It will also include some troubleshooting tips for commonly seen build and run errors, and guidance on how to perform basic profiling. 


Thursday May 20, 2021 1:35pm - 1:50pm MDT

1:35pm MDT

An interactive introduction to Dask
Python has become one of the main programming languages used for data analytics, offering popular libraries such as Numpy, Pandas, and Scikit-Learn. However, these tools were not designed to scale beyond a single machine. As such, they can struggle to analyze larger datasets 100 GB+. Dask provides an open-source solution to this problem by providing a framework to scale computations across multiple machines (ranging from multiple cores on a single laptop to across supercomputer nodes). Furthermore, Dask provides a familiar programming interface because it is co-developed with many popular scientific computing libraries.

In this tutorial, we provide an interactive introduction to Dask. We will run code through Binder, so you do not need to install anything beforehand.

Speakers
AR

Alex Rybchuk

University of Colorado Boulder


Thursday May 20, 2021 1:35pm - 3:05pm MDT

1:35pm MDT

Parallel R
R is a programming language and environment used for statistical computing, data mining, data analysis, and graphics. There is interest in increasing R’s performance. One way to do this is via parallelism, using multiple cores on a single computer or multiple nodes in a distributed memory machine. There are a number of parallel packages available for use with R. This talk will briefly present some of them including the libraries: parallel, dplyr, doParallel, and the R subset of the Message Passing Interface (Rmpi). Rmpi allows distributed memory computing using multiple nodes. We will look at issues associated with these packages that can reduce performance, such as affinity and operation ordering and how to overcome some of them. We will look at combining packages such as MPI and parallel foreach. The primary context of the presentation will a be computationally intensive statistical analysis of earthquakes to develop a hazard map. We will also look at Bag of Task programming. Recipes will be provided for installing many of the packages discussed.

Speakers

Thursday May 20, 2021 1:35pm - 3:05pm MDT

1:35pm MDT

Python interactive big data visualization
Using interactive data visualization for data exploration, collaboration, publication, general education, and more has reached new levels in ease of creation and use with the recent advances with Plotly and Dash. The real challenge is making things work more efficiently with big data. In this hands-on Python tutorial, we will explore the development of an interactive web-based dashboard using large data and Python tools to make working with large data faster and easier both on the development as well as the user experience sides.

Prerequisites: We will be using Jupyter Lab >=3.0.6 with at minimum Python v3.8. In addition, several Python packages will be used in this tutorial and instructions for installing them as well as the Jupyter Lab configuration will be sent to registered participants before the RMACC conference. In addition, help services via CRDDS Consulting Hours will be made available for people needing help setting up before the tutorial.

Speakers
TD

Tim Dunn

University of Colorado Boulder


Thursday May 20, 2021 1:35pm - 3:05pm MDT

3:15pm MDT

Debugging and Optimization Tools for Python HPC Codes- ARM sponsored session
The Arm Forge tools now support both code optimization and debugging of Python codes.  Python is not only being adopted by scientists in traditional HPC disciplines, but it is enabling users in disciplines not typically seen to use HPC to access cluster computing resources.  In this presentation we will provide an overview and hands-on demonstration of these capabilities.


Thursday May 20, 2021 3:15pm - 3:45pm MDT

3:15pm MDT

Green Practices in HPC
Join our panel of experts to hear about how they use green practices in their data centers. Bring your questions for a discussion after some short presentations!

Speakers
AC

Andrew Chien

University of Chicago
MK

Matt Koukl

Affiliated Engineers Inc.


Thursday May 20, 2021 3:15pm - 3:45pm MDT

3:15pm MDT

WHPC
Thursday May 20, 2021 3:15pm - 4:15pm MDT