Loading…
RMACC 2020 HPC Symposium has ended
All times listed are Mountain Daylight Time, UTC -6

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

Wednesday, May 20
 

9:30am

Benchmarking and Enabling Noisy Near-term Quantum Hardware
Quantum computers promise to solve certain problems that are intractable on any current or future classical computer. Superconducting quantum bits (qubits) have emerged as one of the most promising candidates for building such a quantum computer, and is the architecture IBM is pursuing. This talk begins with an overview of superconducting qubits and metrics for benchmarking them, followed by a holistic metric called quantum volume used for benchmarking noisy near-term quantum processors in general. While the focus is often on the quantum processors themselves, the software that implements algorithms, compiles quantum circuits, and controls the quantum hardware is of utmost importance. Qiskit is the open-source quantum computing framework supported by IBM, and has recently been expanded to provide lab-level experimental access to quantum processors on the cloud.

Speakers
DN

Dr. Nick Bronn

IBM
Dr. Nick Bronn is a Research Staff Member in the Experimental Quantum Computing group at the IBM TJ Watson Research Center in Yorktown Heights, NY. He received B.S.'s in Applied Mathematics and Physics and an M.S. in Electrical Engineering from Georgia Tech, a Certificate of Advanced... Read More →


Wednesday May 20, 2020 9:30am - 10:15am

10:30am

Technology Insights from the HPC & AI Innovation Lab
This presentation will share meaningful and useable insights from different technology evaluations conducted in the HPC & AI Innovation Lab. Topics include the latest in Covid19 research using HPC at TGen, application performance and tuning for the CPU architecture, memory configuration impact on performance, as well as research work in scalable deep learning.


Wednesday May 20, 2020 10:30am - 12:00pm

10:30am

Basics of HPC

Join us to learn the fundamentals of high performance computing (HPC). We will cover basic Linux, Bash scripting and HPC job submission to prepare you for using RMACC Summit or any other HPC system. This tutorial is intended for beginner users.

Speakers
DT

Daniel Trahan

University of Colorado Boulder


Wednesday May 20, 2020 10:30am - 12:00pm

10:30am

Introduction to Quantum Computing
For almost 40 years, quantum computing has intrigued and amazed scientists and non-scientists in its future possibility for solving problems that are intractable using classical computing. Over the last four years, IBM has made real quantum computers available on the cloud so that clients, students, and researchers can begin to learn and experiment with this new way of computing.
In this hands-on workshop you will receive an introduction to quantum computing, and you will be able to create quantum programs that leverage the basic quantum computational principles; superposition, entanglement, and interference. You will see examples of actual quantum algorithms, and you will also learn to program directly to the quantum computer using Open Pulse, which allows you to schedule operations via microwave pulse generators to manipulate the qubits.
Your instructors for this course are:
Steve Margolis,
Gabe Chang,
Robert Loredo


Wednesday May 20, 2020 10:30am - 12:00pm

1:00pm

Student Poster Presentations
Wednesday May 20, 2020 1:00pm - 1:25pm

1:30pm

Building and Deploying Containers for Research Workflows
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. Containers can be particularly useful for complex workflows that may require numerous steps and software packages (e.g., bioinformatics, data science applications). This hands-on tutorial will provide an introduction to containers and the basics of running scientific workflows in containers on high-performance computing platforms using Singularity, a containerization software package.

Speakers
AM

Andy Monaghan

University of Colorado Boulder


Wednesday May 20, 2020 1:30pm - 3:00pm

1:30pm

30 minute Panels- See Description for Abstracts and Times
1:30-2:00- GPU Extension to Data Science Tools- Gil Speyer
In the last few years, the “big three” of interpreted applications ---R, Matlab and Python--- have taken great strides to exploit GPU acceleration. Although there are different designs to the grammar and data structures for these applications, they have all implemented GPU-accelerated routines conforming to similar patterns. This will be a fast and broad overview of the current picture of these implementations, with specific examples which highlight the unique developments in each.

2:00-2:30- HPC Operator Toolkit- Jenett Tillotson and Jon Roberts
HPC systems that have continued to grow in size and complexity and as such, tracking the administration of the system components can be cumbersome. We intend to showcase tools written at NCAR that provide system administrators the ability to track, process, and record which elements throughout the HPC system have been identified as having faults or problems and doing so from the comfort of the command line that system administrators are so familiar with and can further provide automation. The tools are open-source and freely available for any site to use and customize to their needs.

2:30-3:00 Site Update PetaLibary- Jonathon Anderson
The University of Colorado Boulder research data storage service, the "PetaLibrary" has continued to evolve and change, now incorporating components including iRODS, LTFS, BeeGFS, and ZFS. We will present a summary overview of the current PetaLibrary infrastructure, including the decisions that have led to the current state.


Wednesday May 20, 2020 1:30pm - 3:00pm

1:30pm

Data, Decisions, Discoveries: How HPC, AI, and IoT Together Will Change the World.
HPC has delivered modeling and simulation capabilities that have greatly extending the reach of our theoretical understanding in science and engineering for five decades. In the past two decades, the growth in digital data availability has enabled analytical modeling and statistical predictions for problems in science, and increasingly industry, in which simulation is impossible, infeasible, or inefficient. However, the goals remains the same: insight used to drive innovation, discovery, and even decision. More recently, advancing technology capabilities have supported a surge in the use of AI (especially deep learning) and the scale of connected digital devices and sensors (edge computing and the nascent IoT). These emerging technologies, plus advances in cloud computing and the push to exascale, are enabling powerful synergies and convergences in applications that now offer new paths to innovation, discovery, and decisions—including autonomous ones. Together, they offer the promise of autonomous connected vehicles, smart power grids and building and cities, more capable drones and robotic systems, precision and personalized therapies, and more. The convergence applications are coming, and soon we’ll be able to have better, faster decisions everywhere.

Speakers

Wednesday May 20, 2020 1:30pm - 3:00pm

3:15pm

The Simplification of Code Porting with Debugging and Performance Analysis Tools
In this presentation we will provide a case study of porting a common climate/weather code from an x86 based system to an arm system.  We will illustrate how using HPC specific debugging and profiling tools greatly simplifies the process of rebuilding a code on a new architecture,  ensuring the correctness of that code on the new system and how to debug that code when portability issues in the application are exposed.  We will illustrate how to use performance analysis tools to verify that the code is obtaining the expected performance and how to enhance the performance of the application when those opportunities exist.


Wednesday May 20, 2020 3:15pm - 3:45pm

3:15pm

An Evaluation of AWS Glacier Deep Archive for Disaster Recovery
Disaster Recovery is an important component of an overall storage strategy, enabling organizations to recover mission critical data. For research centers, that data could include important and hard-to-replace observational datasets that are served to the community, data needed for important on-going or future supercomputer simulations, community software, or other critical digital assets. AWS Glacier Deep Archive is a service offered by Amazon to store data that is rarely accessed and is AWS’s lower cost storage class. To evaluate the feasibility of using this service to meet NCAR's disaster recovery needs, we ran a series of tests and analyses to understand its cost, functionality, and performance.

Speakers

Wednesday May 20, 2020 3:15pm - 3:45pm

3:15pm

Professional Skills for Students
Professional skills can help or hinder your prospects for a job or internship. Let’s talk about effective behaviors from the application process through your time on site. We’ll present tips and tricks to help you avoid common pitfalls and establish yourself as a sought-after team member.

Speakers

Wednesday May 20, 2020 3:15pm - 3:45pm
 
Thursday, May 21
 

9:30am

High-performance computing and the first black hole image
The first resolved image of a black hole was made possible by a series of technological and computational breakthroughs. I will give an overview of the data flow, from petabytes recorded at radio telescopes all over the Earth to compressed kilobyte images. I will then describe what we see in the image, which relies on supercomputer simulations of how gas falls onto black holes, followed by ray tracing visualizations of how the light radiated by that gas should appear to observers on Earth.

Speakers
JD

Jason Dexter

University of Colorado Boulder


Thursday May 21, 2020 9:30am - 10:00am

10:30am

Artificial Intelligence Software Optimization on Intel Xeon Processors
In this presentation we will provide details showing how Intel is working towards optimizing some of the most popular artificial intelligence frameworks, thus speeding-up AI workloads on Intel Xeon processors. We will also present the optimization lib that is responsible to for these optimizations.

Speakers

Thursday May 21, 2020 10:30am - 12:00pm

10:30am

Intro to Jetstream
This tutorial will first give an overview of Jetstream and various aspects of the system. Then we will take attendees through the basics of using Jetstream via the Atmosphere web interface. This will include a guided walk-through of the interface itself, the features provided, the image catalog, launching and using virtual machines on Jetstream, using volume-based storage, and best practices. We’ll discuss advanced features that are available beyond the Atmosphere GUI and also try to provide plenty of time for questions about the system at the end of the session.

Speakers

Thursday May 21, 2020 10:30am - 12:00pm

10:30am

Introduction to NumPy
We first explain the importance of NumPy in scientific and data science applications.
We then compare its performance to regular Python.
Subsequently the following topics will be covered:
a. the creation of one-dimensional and multi-dimensional arrays,
b. the atttibutes of NumPy arrays (dtype,ndim, shape, size,..) and
the creation of particular NumPy arrays (zeros,ones,eye, random,..)
c. index and slicing
d. vector operations and broadcasting
e. universal functions
f. fancy indexing
This tutorial is hands-on. Each topic comes with the corresponding exercises.

Speakers
WC

Wim Cardoen

University of Utah


Thursday May 21, 2020 10:30am - 12:00pm

1:00pm

Career Opportunities and Marketing yourself to employers
Join us for a discussion about available career opportunities in the field of HPC and hear from professionals in the field.

Andrew Westergren is a Senior Manager at Arm Ltd. He currently manages the HPC software stack business for Arm in North America - segments including Weather Climate, Education, Oil and Gas, Commercial, as well as Defense. Andrew has over 20 years of experience in the Technology field with nearly 6 of those years in HPC.

As the lead for the Outreach, Diversity, and Education team in the Computational & Information Systems Lab at the National Center for Atmospheric Research, AJ Lauer directs the Summer Internships in Parallel Computational Sciences (SIParCS) program and works with her team and labmates to create outreach efforts that inspire future generations of HPC users. She also serves as Inclusivity Chair for the 2021 Supercomputing conference. A doctoral candidate in the Interdisciplinary Leadership EdD program at Creighton University, her dissertation work will focus on climate, identity, and leadership in HPC.

Jonathon Anderson of the CU Boulder Research Computing group has worked in academic high performance computing since 2006, with appointments in both laboratory and university settings; and has served as sysadmin, systems architect, and team lead.

Daniel Trahan began work in HPC as a student in 2017, Daniel Trahan is a recent addition to the CU Boulder Research Computing Team. He currently works as as an HPC Specialist within the group where he provides support and training for various HPC topics.

Thursday May 21, 2020 1:00pm - 1:25pm

1:00pm

RMACC Women in HPC- Kickoff Meeting
Join us for the kickoff of the new RMACC Women in HPC Chapter. We will brainstorm activities that we can do as an RMACC chapter, talk about outreach opportunities, and work together to build an RMACC community supportive of diversity and inclusion. All are welcome!

Thursday May 21, 2020 1:00pm - 1:25pm

1:30pm

Introducing oneAPI and Data Parallel C++ Extensions to the SYCL Standard
oneAPI is an Industry standard, and one of the latest efforts to bring a single code framework for high-performance and heterogeneous computation and forms a superset of C++, OpenCL, SYCL and Data Parallel C++.

In this session the attendee will learn about oneAPI, Data Parallel C++ extensions to the SYCL standard and how to get started today using Intel’s Dev Cloud. The session will feature hands on examples that illustrate the new unified shared memory feature. In addition a data parallel C++ sample application showing the usage of Intel® oneAPI DPC++ Library (oneDPL).


Thursday May 21, 2020 1:30pm - 3:00pm

1:30pm

30 minute Panels- See Description for Abstracts and Times
1:30-2:00- Site Update CU Boulder Hybrid Cloud Resources- Jonathon Anderson and Brett Shouse
The University of Colorado Boulder is deploying a hybrid on- and off-premises cloud environment for use as part of the existing Research Computing environment. In this session we will present an update on the current deployment state of the on-premises infrastructure, as well as our expectations for how the new environment will benefit the user community.

2:00-2:30- Machine-Learning-Enhanced Adaptive-Partitioning Multiscale Simulations-
Hai Lin and Adam Duster
Multiscale simulations of solute solvation and transport is important to many physical, organic, and biochemical processes. Quantum-mechanics (QM) can offer accurate descriptions of the solvation shell, whereas molecular-mechanics (MM) efficiently models the larger environment. However, the frequent exchange of solvent molecules between the solute’s solvation shell and the bulk solvent presents a challenge, making it difficult, if not at all impossible, to construct the models. To face this challenge, we have developed novel adaptive-partitioning QM/MM algorithms that allow on-the-fly reclassification of atoms as QM or MM both continuously and smoothly as the trajectory propagates (Fig. 1).[1] This permits the use of a small, mobile QM subsystem centered on the solute with contents that are updated as needed. In this talk, we report our latest progress in the applications of machine-learning techniques to enhance the performance of adaptive-partitioning QM/MM simulations.

2:30-3:00 A Data-Driven Operational Model for Traffic at Dallas Fort-Worth International Airport-
Monte Lunacek


Airports are moving more people and goods faster, cheaper, and with greater convenience than ever before. As air travel continues to grow, airports will face challenges in responding to increasing passenger vehicle traffic, which will lead to lower operational efficiency, poor air quality, and security concerns. Our work evaluates methods for traffic demand forecasting, which will allow airport operations staff to accurately forecast traffic and congestion. Using a year of detailed data describing individual vehicle arrivals and departures, aircraft movements and weather at Dallas Fort-Worth (DFW) International Airport, we evaluate multiple prediction methods including seasonal ARIMA, Prophet, modern supervised machine learning algorithms, and modern deep learning models for time series forecasting. We find that machine learning models based on careful feature engineering offer the best prediction for the next 30 minutes, while the deep learning models perform best on the longer prediction periods. Combining these demand forecasts with a traffic microsimulation framework provides a complete picture of traffic and its consequences. The result is an operational intelligence platform for exploring policy changes, as well as infrastructure expansion and disruption scenarios. To demonstrate the value of this approach, we present results from a case study at DFW airport assessing the impact of a policy change for vehicle routing in high demand scenarios. This framework can assist airports like DFW as they tackle daily operational challenges, as well as explore the integration of emerging technology and expansion of their services into long term plans.

Thursday May 21, 2020 1:30pm - 3:00pm

1:30pm

Fundamentals of Machine Learning
Interactive demonstration with Python jupyter notebooks on RMACC Summit, illustrating the relative ease and flexibility of using the high-level framework Python on modern heterogeneous supercomputers. Topics will include data discovery, regression, unsupervised learning, and neural networks.

Speakers
JY

Jason Yalim

Arizona State University


Thursday May 21, 2020 1:30pm - 3:00pm

3:15pm

Making the most of your internship
Internships offer a variety of experiences - meeting with mentors and collaborators, testing new software, attending social and networking events. Join NCAR internship managers to learn how to capitalize on these opportunities to maximize personal and professional development!

Speakers

Thursday May 21, 2020 3:15pm - 3:45pm

3:15pm

SIGHPC Symposium: Supporting Users of HPC Systems
In conjunction with the SIGHPC Education Chapter, come and learn more about facilitating support to users at our different institutions. This session will provide talks on what user support services are available at institutions across the RMACC consortium, and some helpful information about resources you may utilize. We will also discuss tips and tricks for successfully approaching (and perhaps solving) user issues.

Speakers
TB

Torey Battelle

Colorado School of Mines
AO

Anita Orendt

University of Utah
DD

Diana Dugas

New Mexico State University
SK

Shelley Knuth

University of Colorado Boulder


Thursday May 21, 2020 3:15pm - 3:45pm