Loading…
We are excited to have you join us in August at the 2017 RMACC HPC Symposium.

Please note: This schedule is subject to change prior to the event. 
View analytic

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

Tuesday, August 15
 

11:30am

Registration
Tuesday August 15, 2017 11:30am - 1:00pm
SEEC Atrium

1:00pm

Supercomputing for Beginners
This beginner session will cover introductory ideas of High Performance Computing and introduce users to the basic ideas behind Supercomputing. 

Speakers

Tuesday August 15, 2017 1:00pm - 2:30pm
SEEC Main Auditorium

1:00pm

Enabling Science Beyond the Campus Edge: HPC Integration with the Open Science Grid (Part One)

"The tutorial presenters request that all attendees contact user-support@opensciencegrid.org before 8/11/2017 in order to receive important preparatory documentation.  Please include 'RMACC17' in the email subject."


Enabling campus researchers to share computational and data resources with external collaborators is a powerful multiplier in advancing science. Sharing spare capacity for even short durations allows an institutional HPC resource a cost-efficient means of participating in a larger cyber ecosystem.  In this session we will show you how to integrate your HPC cluster resource to the Open Science Grid to support collaborative, multi-institutional science.  The only requirements are that your cluster can provide SSH access to a single OSG staff member, that your cluster job submission and worker nodes have outbound IP connectivity, the operating system is CentOS/RHEL 6.x, 7.x or similar, and that a common batch scheduler is used (e.g. SLURM, PBS, HTCondor).  During the tutorial we will configure OSG managed services to deliver workloads from science communities using the OSG to your HPC cluster. 

https://swc-osg-workshop.github.io/OSG-HostedCE-RMACC17/


Speakers

Tuesday August 15, 2017 1:00pm - 2:30pm
S125

2:30pm

Break- Sponsored by Pure Storage
Tuesday August 15, 2017 2:30pm - 3:00pm
SEEC Atrium

3:00pm

Introduction to Linux

Most HPC systems use the Linux operating system.  The goal of this session is to give newcomers to Linux the skills and knowledge to work confidently at the Linux command line.  We'll cover basic concepts, logging in, the Linux shell, useful commands, navigating the directory structure, creating and editing files, running programs, and a few handy tricks and tips.  No previous experience with Linux is required.  If you would like to follow along with the hands-on examples, please bring a laptop that a) runs Linux or Mac OSX, or b) has a Secure Shell client installed (such as PuTTY, downloadable from http://www.chiark.greenend.org.uk/~sgtatham/putty/latest.html .)


Speakers

Tuesday August 15, 2017 3:00pm - 4:30pm
N129

3:00pm

Introduction to numpy for HPC

The numpy module is the heart of the Python scientific computing environment.
In turn, the ndarray data structure is the heart of numpy.  We will cover the
basics of the ndarray used to represent arbitrary dimensioned data arrays. 
This will include universal functions - which support vectorized operations,
broadcasting - which allows operations between compatibly sized arrays,
and show some basics on how numpy is used with the other scientific

computing modules - scipy, pandas, and matplotlib.

Introduction to numpy for HPC

Speakers

Tuesday August 15, 2017 3:00pm - 4:30pm
N124

3:00pm

Introduction to Scientific Visualization
Science and research increasingly runs on big data, and visualizing these large and complexdatasets is key to their interpretation. Scientific visualization holds the key to gaining insights,validating data and even making new discoveries. In this session we will cover the basicfoundational principles and concepts of scientific visualization. We will go over a few practicaldemonstrations on transforming and interpreting your data through a number of visualizationtechniques. Topics will include looking at various data formats, the visualization pipeline used toderive intuitive graphical output, introduction of various visualization algorithms, and looking toVTK/Paraview capabilities for scientific visualization. We invite both the new and theexperienced practitioners of visualization to attend; this will allow for a far reaching Q&A and anexciting and varied session.

Speakers

Tuesday August 15, 2017 3:00pm - 4:30pm
N128

3:00pm

Parallel Programming for Beginners
This session will consist of 45 minutes lecture and 45 minutes hands-on.

Participants should have had at least one semester of C, C++, or Fortran. 

Speakers

Tuesday August 15, 2017 3:00pm - 4:30pm
SEEC Main Auditorium

3:00pm

Enabling Science Beyond the Campus Edge: HPC Integration with the Open Science Grid (Part Two)

Enabling campus researchers to share computational and data resources with external collaborators is a powerful multiplier in advancing science. Sharing spare capacity for even short durations allows an institutional HPC resource a cost-efficient means of participating in a larger cyber ecosystem.  In this session we will show you how to integrate your HPC cluster resource to the Open Science Grid to support collaborative, multi-institutional science.  The only requirements are that your cluster can provide SSH access to a single OSG staff member, that your cluster job submission and worker nodes have outbound IP connectivity, the operating system is CentOS/RHEL 6.x, 7.x or similar, and that a common batch scheduler is used (e.g. SLURM, PBS, HTCondor).  During the tutorial we will configure OSG managed services to deliver workloads from science communities using the OSG to your HPC cluster. 

Enabling Science Beyond the Campus Edge: HPC Integration with the Open Science Grid

Speakers

Tuesday August 15, 2017 3:00pm - 4:30pm
S125
 
Wednesday, August 16
 

7:30am

Breakfast
Wednesday August 16, 2017 7:30am - 8:15am
SEEC Atrium

7:30am

Registration
Wednesday August 16, 2017 7:30am - 8:30am
SEEC Atrium

8:15am

Welcome
Speakers

Wednesday August 16, 2017 8:15am - 8:30am
SEEC Main Auditorium

8:45am

The Future of HPC Architecture
Join us for an innovative Keynote debate about the Future of HPC Architecture. Our panel of speakers includes Intel, Nvidia, Mellanox and HPE. The format will be as follows 

- Opening arguments (3 minutes each)
- Moderated debate with questions (30 minutes)- questions will be provided ahead of time to participants
- Closing arguments (2 minutes each)
- Open Q and A with audience (15 minutes)  

Speakers
MH

Marc Hamilton

Marc Hamilton is Vice President, Solutions Architecture and Engineering at NVIDIA where he leads NVIDIA's worldwide team of solutions architects and field application engineers responsible for working with customers to transform science and harness the power of the cloud with machine... Read More →
JL

Jerry Lotto

Jerry joined Mellanox in 2016 as Director HPC and Technical Computing, with more than 30 years of experience with scientific computing. An early adopter of InfiniBand, Jerry built the first HPC teaching cluster in the Harvard’s Department of Chemistry and Chemical Biology with an... Read More →
WM

William Magro

William (Bill) Magro is an Intel Fellow and Intel’s Chief Technologist for HPC software. In this role, he serves as the technical lead and strategist for Intel's high-performance computing (HPC) software and provides HPC software requirements for Intel product roadmaps, covering... Read More →
MV

Mike Vildibill

Mike Vildibill is Vice President of Hewlett Packard Enterprise’s Exascale Development, Federal Programs and HPC Storage groups, where he is responsible for product strategy, engineering and advanced technologies development. He has 25 years’ experience in HPC including executive... Read More →


Wednesday August 16, 2017 8:45am - 10:00am
SEEC Main Auditorium

10:00am

Break
Wednesday August 16, 2017 10:00am - 10:30am
SEEC Atrium

10:30am

Integrating Apache Hadoop & Spark on USGS HPC cluster using Magpie

We cover the integration of existing USGS HPC environment with Big Data software like Apache Hadoop and Spark using open-source software Magpie from Lawrence Livermore National Laboratory. The diverse examples on USGS supercomputer use both interactive PySpark shell and sbatch script submission with SLURM.


Speakers

Wednesday August 16, 2017 10:30am - 11:30am
N126

10:30am

Visualization Showcase

Data is brought to life through visualization.  This session will showcase several exceptional examples of visualizations whose presentation or underlying technology embodies leading-edge features.


Moderators
KW

Kenji Williams

Kenji Williams is a composer and director for multi-media live theater, augmented reality, virtual reality, and interactive datavisualization. | | Williams is an Artist in Residence and Visiting Scholar at the University of Colorado at Boulder Arts & Sciences developing a live Augmented... Read More →

Wednesday August 16, 2017 10:30am - 11:30am
S372

10:30am

Resume/Cover Letter review and Interviewing Skills
Join an interactive discussion about how to prepare and conquer job interviews. 
Attendees will practice their interview skills by working in teams.

Bring your resume and/or cover letter for review by our moderators and other attendees. Get help to improve the layout, flow, and content. 

Moderators
Wednesday August 16, 2017 10:30am - 11:30am
Sievers Conference Room

10:30am

Training and Outreach Opportunities in HPC

This session will consist of several lightning talks from various RMACC members about how training and outreach programs are effective at their institutions for supporting users of HPC resources.  Some topics will include:  the usefulness of hackathons, new training offerings, involving students in assisting users, surveys, the effectiveness of webinars, training the trainers, and others. 



Wednesday August 16, 2017 10:30am - 11:30am
N128

10:30am

Dell EMC- Next Generation Performance: An Update on HPC and Machine Learning Application Performance Studies

This session will cover results of recent performance studies on the latest processor and accelerator technologies conducted at Dell EMC HPC Innovation lab. The audience will get an opportunity to learn more about Dell EMC’s approach to HPC and how the studies that Innovation Lab conducts can help them make the right design decisions for their system or how to take the most benefit out of their existing system.  



Wednesday August 16, 2017 10:30am - 11:30am
SEEC Main Auditorium

10:30am

HPE Technical Talk
Wednesday August 16, 2017 10:30am - 11:30am
N129

11:30am

Lunch- Sponsored by Dell
Wednesday August 16, 2017 11:30am - 1:00pm
SEEC Atrium

12:45pm

1:30pm

Bella Gaia- Data Visualization and Interaction

The great Jewish Rabbi Abraham Joshua Heschel once said that, “Mankind will not perish for want of information; but only for want of appreciation.”
Artist and innovator Kenji Williams will present a live performance of the globally touring NASA-powered data visualization spectacle - BELLA GAIA http://www.bellagaia.com/ .  Williams will speak about how he brings soul to data through his unique storytelling method utilizing the latest neuroscience research to create deep engagement in the brain, and why communicating data with holistic ethics and whole-brain methodologies must be the future of data visualization and interaction.


Speakers
KW

Kenji Williams

Kenji Williams is a composer and director for multi-media live theater, augmented reality, virtual reality, and interactive datavisualization. | | Williams is an Artist in Residence and Visiting Scholar at the University of Colorado at Boulder Arts & Sciences developing a live Augmented... Read More →


Wednesday August 16, 2017 1:30pm - 2:30pm
SEEC Main Auditorium

2:30pm

Break- Sponsored by DDN
Wednesday August 16, 2017 2:30pm - 3:00pm
SEEC Atrium

3:00pm

Mellanox- Paving the Path to Exascale (30 minutes)

This era of Exascale is pushing the principles of Co-Design into the limelight.  Mellanox Technologies is taking this concept in new directions by enabling capabilities which stand to benefit all scales of HPC cluster computing.  Learn how an interconnect fabric can drive the highest application performance and help realize the highest ROI for your cluster investment.


Speakers
JL

Jerry Lotto

Jerry joined Mellanox in 2016 as Director HPC and Technical Computing, with more than 30 years of experience with scientific computing. An early adopter of InfiniBand, Jerry built the first HPC teaching cluster in the Harvard’s Department of Chemistry and Chemical Biology with an... Read More →


Wednesday August 16, 2017 3:00pm - 3:30pm
SEEC Main Auditorium

3:00pm

MFIX Case Study- Profiling and optimizing a large scientific application

 

MFiX, a multiphase flow solver from the Department of Energy, has been the subject of multiple improvement projects over the past 18 months. This includes review and analysis by a team from NREL and CU-Boulder to profile the code and investigate possible performance improvements. This talk will review the process and conclusions of the NREL/CUB team regards the MFiX code specifically and more generalized lessons learned for analysis of mature FORTRAN code.


Speakers

Wednesday August 16, 2017 3:00pm - 4:00pm
N129

3:00pm

Research and Education: Reproducible Numerical Weather Predication Enabled by Software Containers

The purpose of the presentation is to introduce an affordable, portable, and reproducible platform for numerical weather prediction model usage within the classroom or research space.  Specifically demonstrate that by combining cloud computing and containerization software, numerical models are readily available for hands on manipulation to a wider range of users even without a robust computational infrastructure.


Speakers

Wednesday August 16, 2017 3:00pm - 4:00pm
N124

3:00pm

Collaborative Scientific Visualization: CAVE2 CAVE and VueNix

 

Firstly, the session will focus on the CAES-funded CAVE 2 CAVE project. We will address how the protocol has developed and what technical milestones were achieved with regard to user being able to share 3D CAVE workspaces at differing locations. Secondly, the session will outline development of the UW Shell 3D Viz Center ‘VueNix’ tool, which supports big data viz in collaborative workspaces. We will address the business case as to why the development is deemed necessary for the visualization community, and how that development has been implemented to date. This will address the challenges of working with some existing traditional big data tools, and consider why an alternative is appropriate.



Wednesday August 16, 2017 3:00pm - 4:00pm
N128

3:00pm

BRO at CU Boulder: Achieving multi-gigabit Network Security Monitoring on the cheap

The University of Colorado Boulder runs a rather large (160 worker node) Bro Network Security Monitoring (NSM) cluster. We will talk through the process of where we were before we migrated to our current cluster, what functional pieces (software/hardware) we used to build it, what we have achieved by moving to it, and where we are going with respect to the future of our Bro NSM cluster. In the process of telling our story, we hope to demonstrate how you might be able to achieve similar results in your environment.


Speakers

Wednesday August 16, 2017 3:00pm - 4:00pm
S125

3:00pm

Creating a Community of Regional Data and Workflow Cyberinfrastructure Facilitators
In this session the leaders of the NSF award “CC* Cyber Team: Creating a Community of Regional Data and Workflow Cyberinfrastructure Facilitators” will give an overview of project. The project will provide three workflow facilitators that will support researchers in the areas of data generation, data analysis, data visualization, and data management. Part of the mission of the project is to work with researchers in the RMACC region. The project leaders like to hear from the participants about their data needs and how the project might support and engage with researchers that have specialized data needs. We will discuss also the data services that will be coordinated between the three institutions, University of Colorado Boulder, Colorado State University, and the University of Utah.

Speakers

Wednesday August 16, 2017 3:00pm - 4:00pm
C120C

3:00pm

Sandstone HPC - A Domain-General Gateway for New HPC Users
CURC is developing Sandstone HPC, a browser-based domain-general gateway that makes it easier for new users to access and utilize HPC resources. In this presentation we will demonstrate how Sandstone makes it easier to upload and edit files, use the terminal, and schedule jobs with interactive web utilities. We will also discuss how Sandstone can be integrated into existing workflows, and plans for future development.

Speakers

Wednesday August 16, 2017 3:00pm - 4:00pm
Sievers Conference Room

3:30pm

DDN- Architecting Academic Research Central Data Stores for Fast, Archive & Remote Data(30 minutes)

The DDN Central Data Store (CDS) combines the performance, scale and reliability of DDN’s SFA storage hardware, with the stability and data management capabilities of GRIDScaler (IBM Spectrum Scale), to deliver a highly scalable filesystem utilizing familiar NAS protocols for ease of use. The design goals of the CDS solution are twofold. The first is to provide a highly reliable and scalable storage platform offering familiar access protocols which will appeal to the user community. The second goal is to deliver a securely managed, centralized storage solution in a cost-effective package that discourages users from procuring their own unmanaged storage devices, or placing potentially sensitive data in unsecured cloud offerings. DDN realizes ease of use and affordability are absolutely mandatory in order to encourage users to take advantage of a centralized storage service. We also realize manageability, reliability and scalability are of paramount importance to those responsible for administering the storage. DDN is confident our Central Data Store solution will deliver on these requirements.


Speakers
RG

Roger Goff

The DDN Central Data Store (CDS) combines the performance, scale and reliability of DDN’s SFA storage hardware, with the stability and data management capabilities of GRIDScaler (IBM Spectrum Scale), to deliver a highly scalable filesystem utilizing familiar NAS protocols for ease... Read More →


Wednesday August 16, 2017 3:30pm - 4:00pm
SEEC Main Auditorium

4:00pm

NVIDIA- The Convergence of HPC and Deep Learning (30 minutes)
Science is being enabled by Supercomputing, whether it’s climate science, combustion science, or understanding the fundamentals of how the human body works. What’s exciting is that the same technology enabling this powerful science is also enabling the revolution in deep learning, and it’s being enabled by GPUs. This session will explain how the rapid advancement of deep learning for artificial intelligence has created an enormous demand for computational resources outside the traditional supercomputing domains.  NVIDIA is uniquely suited to address these evolving needs with accelerated computing and will present recent GPU hardware and software advances and how they address computational needs in both AI and HPC.  There will also be an update on the Deep Learning Institute (DLI) Teaching Kit, which NVIDIA has partnered with Professor Yann LeCun of New York University and Facebook AI Research. The kit covers the academic theory and application of deep learning on GPUs using the PyTorch and Torch frameworks and includes detailed lecture slides, hands-on labs/source code solutions, quiz/exam problem sets, and free access to online deep learning labs using GPUs in the cloud.

Speakers

Wednesday August 16, 2017 4:00pm - 4:30pm
N124

4:00pm

Summit Researcher Panel

Come and see how people like you are using RMACC's new Summit supercomputer.  Researchers from a variety of disciplines will give a quick overview of their work on Summit, with a focus on how their computational methods match up with Summit's architecture.  They will also describe any challenges they had to overcome to start running on Summit.  There will be time for Q-and-A and discussion.


Moderators
Wednesday August 16, 2017 4:00pm - 5:00pm
N129

4:00pm

Streamlining your research through Data Management and Reproducible Research

Good data management and reproducible research practices are becoming increasingly important in the digital age. Researchers are producing more data than ever before. Consequently, analyses are becoming more complex and time consuming, making documentation ever more important. Without proper preservation, valuable datasets and analysis methods are being lost over time. Concurrently, research funding is becoming scarcer, and funding agencies are trying to get more “bang for their buck” by encouraging open science to facilitate resource reuse. Luckily, we have the technology to facilitate these new research responsibilities. This session will discuss best practices for data management and reproducible research.



Wednesday August 16, 2017 4:00pm - 5:00pm
N128

4:00pm

How to Access Summit

CU Boulder Research Computing staff will step through the logistics of how to access Summit. We will discuss how to login to access RC resources, how to get an allocation, how to submit jobs, storage spaces, and other important information.


Speakers

Wednesday August 16, 2017 4:00pm - 5:00pm
C120C

4:00pm

Marketing yourself to Employers
Everyone has skills. They get you in the door, but not necessarily get you the job. There can be 100 or more applicants per job posting, and they all have the same or better skills as you.  It’s not just about the skills, it’s how to put your best foot forward to stand out as "the one.”   In today’s extremely competitive job environment it is increasingly important for each person to create a clear, concise statement on who you are, what professional skills you offer, and why you are the best candidate for the position.    Hear from marketing and industry professionals on what employers want to know and how to prepare for the questions employers are going to ask. 

Moderators
Speakers
KJ

Kevin Jameson

Federal HP Specialist, HPE
MZ

Matthew Ziegler

Worldwide Director - HPC Product Marketing, Lenovo


Wednesday August 16, 2017 4:00pm - 5:00pm
Sievers Conference Room

4:00pm

EasyBuild: Providing and Maintaining a Broad Range of Scientific Software / Introduction to Singularity: A Container System targeting HPC
HPC user software administration has become quite a burden for center administration staff as systems are servicing broader domains and workflows. Leveraging utilities which help to automate the process provide a mechanism to allow staff to concentrate on more complex issues and optimizations rather than continuously constructing software for end users from source that can take large portions of administration time away. Two techniques are becoming more prominent including a build management system and containerization methods to produce software for known tested configuration in an automated fashion to save time and offer a wider capability. HPC and Cloud environments have become more ubiquitous across many academic organizations, laboratories, and corporations and acquiring the necessary talent to operate and maintain these systems, on premise or cloud hosted, requires an individual to be aware of the Information Technology operating model, but also aware how the organization's users operate. Often these computational machines take a generous portion of time to administer, maintain, and optimize at the basic operational level, but even more effort is required at the workflow level. To complicate the situation, these machines are now being shared by more and more domain specific studies which likely have radically different workflow models. Managing software in the research world is a complex tasks where most software must be compiled from source, unpacked in a certain method, or use a bizarre installer. Two different techniques are discussed in this session which are complementary to one another, but can also be deployed as a single instance. The first being a build management system, similar to a package manager, where software like EasyBuild, Spack, or Maali is used to leverage specification files to build software often templated out by other members of the community. The second leveraging the containerization technology to create a packed image of a complete environment. Platforms like Docker and other containerization platform provide a framework for deploying tested software in the cloud for services, however with downsides that are unacceptable to shared computation resources such as supercomputers or shared compute clusters. Specific frameworks that eliminate many issues of common cloud container frameworks, Singularity and Shifter are becoming prominent alternatives for a more secure operating mode on shared resources. In this talk, we’ll specifically target EasyBuild and Singularity from the basics and establish the basics for the administrative workflow to ease and enhance the end user’s workflow.

Speakers

Wednesday August 16, 2017 4:00pm - 5:00pm
S125

4:30pm

Pure Storage- HPC data storage: acceleration, simplification, and consolidation with FlashBlade (30 minutes)

 Pure Storage will discuss how their data platform supports block, file, and object storage needs within HPC environments. Moreover, how their platform accelerates application performance (higher IOPS & throughput, lower application response time), simplifies procurement/management/operations, and consolidates racks/rows worth of legacy storage into just a few rack units. 



Wednesday August 16, 2017 4:30pm - 5:30pm
N124

5:30pm

Reception- Sponsored by Nvidia
Wednesday August 16, 2017 5:30pm - 7:00pm
SEEC Atrium
 
Thursday, August 17
 

7:30am

Breakfast
Thursday August 17, 2017 7:30am - 8:30am
SEEC Atrium

7:30am

Registration
Thursday August 17, 2017 7:30am - 8:30am
SEEC Atrium

8:30am

High Throughput Computation on the Open Science Grid and AWS (Part One)

Would you like to use distributed resources of the Open Science Grid, or just want to learn how to do large-scale high throughput computing?  The format of this tutorial is a mix of lecture and hands-on exercises, so please bring your laptop and make sure you have an SSH client installed.  You will be provided with a training account on OSG Connect which, during the tutorial, will be upgraded to a full user account. After the session, you will have full access to the OSG, know how to run and scale up workloads, manage your data and also submit to Amazon cloud resources.  If time permits, the OSG User Support team will also help get your own workload set up for execution on OSG. Topics to include:

  • Introduction to OSG Connect 
  • Job Scheduling with HTCondor 
  • Scaling Up Workloads
  • Managing data
  • High throughput submission to Amazon cloud

https://github.com/SWC-OSG-Workshop/OSG-UserTraining-RMACC17



Speakers

Thursday August 17, 2017 8:30am - 10:00am
N129

8:30am

Practical Data Wrangling with Pandas

Hacking Python?  Need to import some Excel data and run a detailed data analysis?  Got Pandas?  Pandas has become a staple in the Python data science stack with strengths in data manipulation and analysis.  In this workshop, we will focus on real-world data analysis scenarios that show the strengths of this library.  We'll cover basic Pandas data structures, core import/export and I/O functionality, manipulation of data in Pandas and the basics of Pandas data visualization.  We will focus on the practical so that you can leave ready to apply your skills.  We assume a basic working knowledge of Python and exposure to Jupyter Notebooks. 


Speakers
KM

keith maull

Data Scientist and Software Engineer, NCAR


Thursday August 17, 2017 8:30am - 10:00am
N126

8:30am

Intro to Paraview
In this session we will explore scientific visualization and scientific data visualization using Paraview. We will start with why to choose Paraview over other products and the pro's and con's of using Paraview and where it's power lies. Then we will jump into a use case based tutorial walking through many of the main functionalities of Paraview ranging from the basic gui layout and how to develop a full workflow pipeline for any project, to covering many of the major visualization methods and how to get the look you want. Finally we will wrap up with final rendering, animations and incorporating python into your Paraview workflow. 

Speakers

Thursday August 17, 2017 8:30am - 10:00am
SEEC Main Auditorium

8:30am

Intro to R

This course will provide a ground level introduction to the R statistical programming language. We will cover a broad range of topics to familiarize students with the common programming structures including functions, loops, conditional logic, etc. Additionally, this will highlight the strengths of R and provide some examples of advanced use cases in data management, graphics, and more. 


Speakers

Thursday August 17, 2017 8:30am - 10:00am
N128

8:30am

NCAR Site Update/CU Site Update

Cheyenne is the new 5.34-petaflops, high-performance supercomputer deployed by NCAR and released for production work on January 12, 2017.   Cheyenne is an SGI ICE XA Cluster, the Cheyenne supercomputer features 145,152 Intel Xeon processor cores in 4,032 dual-socket nodes (36 cores/node) and 313 TB of total memory and EDR Infiniband interconnect. Cheyenne is connected to new data storage components provided by DataDirect Networks (DDN) which increased the shared file system’s total usable capacity from 16 PB to 54PB. 

In this site update, we will discuss the current state of Cheyenne and share our experience during the install and deployment of the multi-petaflop system. 



In 2017 University of Colorado Boulder deployed RMACC Summit, an heterogeneous compute cluster providing Intel Xeon “Haswell”, intel Xeon Phi “knights landing”, Nvidia Tesla K80, and large (shared) memory resources using an Intel Omni-Path Architecture interconnect and a DDN GRIDScaler SFA14k (IBM Spectrum Scale / GPFS) scratch file system. In this site update we will discuss the current state of Summit, its accessibility within RMACC, and what we plan to do next.

Thursday August 17, 2017 8:30am - 10:00am
S125

8:30am

XSEDE Resources and the Campus Champion Program

Campus Champions from RMACC institutions and from the XSEDE program will lead a discussion with an overview of XSEDE resources and policies, along with what being a part of the Campus Champion program involves and how to get started. Our regional champion program can provide information and support to schools without champions. 


Moderators
Thursday August 17, 2017 8:30am - 10:00am
C120C

8:30am

Intel Compilers

1.  Introduction to Intel Compilers

2.  Optimization Essentials with the Intel Compiler

a. Optimization basics

b. Compiling for AVX

3.  Forcing/Controlling Vectorization with pragma SIMD

4.  Multi-Core, and OpenMP

Notes:

-         Introduction of the compiler can be brief

-         Fortran and C/C++ need to be covered;

-         floating point reproducibility is a critical issue and should be covered;

-         compiling for KNL is important



Thursday August 17, 2017 8:30am - 10:00am
S372

8:30am

KNL Training- All Day Training (UPDATED LOCATION)
To Register please use the following link:
https://colfaxresearch.com/rmacc2017/
Passcode: BLDR17CC

  • Sneak Peak: What will be covered today (30 min)
  • Programming and Optimization by Example (2.5 hours)
    • Demonstration of a case study: direct N-body simulation
    • Intel processor architectures
    • Task and data parallelism
    • Memory organization
    • Programming coprocessors and clusters
  • Optimization Pointers (1.5 hour)
    • Scalar tuning and using Intel compilers
    • Automatic vectorization
    • Multi-threading with OpenMP
    • Optimizing cache usage and memory access
    • Communication control
    • Preparing for Intel Xeon Phi processors (30 min)
      • Compiling with AVX-512
      • Using high-bandwidth memory
      • Leveraging clustering modes
      • Coprocessor form-factor and KNL-F
    • Intel Libraries (20 min)
      • Intel Math Kernel Library (MKL): numerical methods
      • Intel Data Analytics Acceleration Library (DAAL): machine learning
  • Intel Distribution for Python (20 min)
    • Brief intro to Intel Python (where to get it, installation, etc.)
    • Discussing numpy, scipy and link with Intel MKL
    • How to get the most out of numpy and scipy
  • Intel-Optimized Deep Learning Frameworks (20 min)
    • Deep learning frameworks in data analytics
    • How to obtain Intel-optimized frameworks
    • Deep neural networks on Intel Architecture in action


Thursday August 17, 2017 8:30am - 4:30pm
Bartlett Center- 2nd Floor

10:00am

Break
Thursday August 17, 2017 10:00am - 10:30am
SEEC Atrium

10:30am

High Throughput Computation on the Open Science Grid and AWS (Part Two)

Would you like to use distributed resources of the Open Science Grid, or just want to learn how to do large-scale high throughput computing?  The format of this tutorial is a mix of lecture and hands-on exercises, so please bring your laptop and make sure you have an SSH client installed.  You will be provided with a training account on OSG Connect which, during the tutorial, will be upgraded to a full user account. After the session, you will have full access to the OSG, know how to run and scale up workloads, manage your data and also submit to Amazon cloud resources.  If time permits, the OSG User Support team will also help get your own workload set up for execution on OSG. Topics to include:

  • Introduction to OSG Connect 
  • Job Scheduling with HTCondor 
  • Scaling Up Workloads
  • Managing data
  • High throughput submission to Amazon cloud

 


Speakers

Thursday August 17, 2017 10:30am - 12:00pm
N129

10:30am

Data Processing with R

This course will focus on building cohesive and efficient data processing pipelines for analysis. Specific focus will be placed on handling common data processing tasks including import/export, filtering, grouping, and summarising using the dplyr family of packages (dplyr, tidyr, magrittr, etc.). This course will conclude by looking at  creating efficient visualizations with this pipeline using ggplot2. Slides and code examples of common tasks will be provided.


Speakers

Thursday August 17, 2017 10:30am - 12:00pm
N126

10:30am

Broadening Access to Cyberinfrastructure with the Globus Platform

Research and educational institutions continue to adopt Globus for managing research data. While core Globus capabilities - including reliable file transfer and secure sharing - are the primary focus, there is a growing need to provide robust mechanisms for accessing diverse resources as part of research workflows. Over the past year, Globus released a number of tools that facilitate development of science gateways, data portals, and other custom applications and services to support researchers. In this talk we will review various services and tools comprising the Globus platform-as-a-service, and provide examples on how these capabilities can be integrated into existing and new applications.


Speakers

Thursday August 17, 2017 10:30am - 12:00pm
SEEC Main Auditorium

10:30am

Revision Control with Git

 

Git is a free and open source distributed version control system designed to handle everything from small to very large projects with speed and efficiency.  This workshop will give a brief description of Git, and will cover initializing, adding, and committing to repositories, and collaboration on Github.  Please come to the workshop having git installed on your laptop as this will be an interactive session.


Speakers

Thursday August 17, 2017 10:30am - 12:00pm
N128

10:30am

Gender Bias in Computing

Bias exists throughout society, including in the realm of gender and technology. This talk will describe unconscious bias and why it happens, and also give some strategies for managing it and improving innovation and teamwork in the process. 


Speakers

Thursday August 17, 2017 10:30am - 12:00pm
C120C

10:30am

Intel Distribution for Python

Analysis Tools for Performance and Correctness

Introduction to Intel® VTune™ Amplifier

  1. Types of Analysis
    1. Software Collections
    2. Hardware Collections
  2. Understanding the Results
  3. Additional Features
    1. Command Line Collections
    2. Remote Data Collections
  4. Application Performance Snapshot

 

Introduction to Intel® Inspector

  1. Types of Analysis
    1. Memory Checking Analysis
    2. Threading Analysis
  2. Managing Results
  3. Advanced Features

Introduction to Intel® Advisor

  1. Vectorization Advisor
  2. Roofline Modeling


Thursday August 17, 2017 10:30am - 12:00pm
S372

12:00pm

Lunch- Sponsored by HPE
Thursday August 17, 2017 12:00pm - 1:00pm
SEEC Atrium

1:00pm

Parallel R
Learn how to use R utilizing all cores of your or a supercomputer. You will learn how to prepare your serial R code so that it will be easier to run it in parallel. The parallel package will be introduced and you will be able to run some simple examples in parallel. Next, you will install the pbdMPI package in your project directory on our RMACC Summit supercomputer so that you can run R programs that utilize more than one compute node. You will then run some simple examples, including parallel random number generation, to get started and to familiarize yourself with running parallel R programs

Speakers

Thursday August 17, 2017 1:00pm - 2:30pm
N129

1:00pm

Image Classification with DIGITS

Learn how to leverage deep neural networks (DNN) within the deep learning workflow to solve a real-world image classification problem using NVIDIA DIGITS. You will walk through the process of data preparation, model definition, model training and troubleshooting. You will use validation data to test and try different strategies for improving model performance using GPUs. On completion of this lab, you will be able to use DIGITS to train a DNN on your own image classification application.


Speakers

Thursday August 17, 2017 1:00pm - 2:30pm
N124

1:00pm

Visualization and HPC Panel

The moderator of the visualization panel would like to invite both the new and the experienced HPC and visualization practitioners in the RMACC community, to join in a lively discussion on the role of visualization in HPC. Topics such as: the sharing of good practice with regards to tools and techniques, how to stimulate visualization and HPC activity within the RMACC community, and what are the differences between popular big data and scientific big data will be addressed. Also of interest are areas such as: how to build a case for scientific visualization in virtual reality, in-situ analysis, and how visualization plays a pivotal role in the scientific output of HPC.  The experienced panel members welcome all those interested to the session, and will be delighted to respond to questions throughout.


Moderators
Thursday August 17, 2017 1:00pm - 2:30pm
SEEC Main Auditorium

1:00pm

Efficient Submission of Serial Jobs
Not all workloads run on an HPC resource are what would be traditionally considered high-performance computing jobs. In this session we will present an introduction to “high-throughput” workflows, where a large number of an otherwise serial workload is submitted to run concurrently on a large, high-capacity resource. Examples will be provided using manual Bash job control semantics, GNU parallel, Slurm job arrays, and the site-local (but open-source) CURC loadbalancer.

Speakers

Thursday August 17, 2017 1:00pm - 2:30pm
N128

1:00pm

Data Storage Panel

Between home directories and scratch storage lies a broad range of requirements for medium- and long-term data storage. With the help of a panel of systems and policy experts from the RMACC community, we’ll discuss and compare infrastructure, policies, access methods, and security concerns.

This will be an interactive discussion with questions and contribution from the general audience.


Thursday August 17, 2017 1:00pm - 2:30pm
S125

1:00pm

Best Practices for Installing Application Software

In this tutorial we will cover approaches to installing applications in a HPC computing environment. We will start with an overview of the process of deploying community codes as well as commercial programs, with a focus on considerations and approaches. This will be followed by hands on examples. The first will be the deployment of a binary community code, and the second will focus on the installation of libraries for Python and R.

We will have University of Utah Center for High Performance VM accounts available for users to do the installations for the examples.



Thursday August 17, 2017 1:00pm - 2:30pm
C120C

1:00pm

Intel® Cluster Tools
  1. Intel® MPI Library
    1. Overview
    2. Usage
  2. Intel® Trace Analyzer and Collector
    1. Overview
    2. Collecting MPI Performance Data
    3. Viewing MPI Performance Data
    4. Additional Analysis Tools


Thursday August 17, 2017 1:00pm - 2:30pm
S372

2:30pm

Break- Sponsored by Mellanox
Thursday August 17, 2017 2:30pm - 3:00pm
SEEC Atrium

3:00pm

OpenACC
The OpenACC standard enables the rapid development of GPU-accelerated software via the addition of compiler directives into existing source code.   In this session, we will provide a hands-on introduction to GPU programming using OpenACC.  Topics covered will include parallelization, efficient data transfer, and asynchronous operations.  Examples and exercises will be provided in both Fortran and C, and attendees should have familiarity with one of these two languages.



Thursday August 17, 2017 3:00pm - 4:30pm
N129

3:00pm

Modeling Time Series Data with Recurrent Neural Networks in Keras

One important area of current research is the use of deep neural networks to classify or forecast time-series data. Time-series data is produced in large volumes from sensors in a variety of application domains including Internet of Things (IoT), cyber security, data center management and medical patient care. In this lab, you will learn how to create training and testing datasets using electronic health records in HDF5 (hierarchical data format version five) and prepare datasets for use with recurrent neural networks (RNNs), which allows modeling of very complex data sequences. You will then construct a long-short term memory model (LSTM), a specific RNN architecture, using the Keras library running on top of Theano to evaluate model performance against baseline data.​


Speakers

Thursday August 17, 2017 3:00pm - 4:30pm
N124

3:00pm

Wrangling Data in Linux

Manipulating files and the data in them is a key skill for researchers in Linux-based computing environments.  We'll cover creating, moving, searching, and analyzing files and data using Linux commands and utilities such as "find", "grep", "sed", and "awk".  A major goal will be increasing the productivity and efficiency of your workflows.   If you would like to follow along with the hands-on examples, please bring a laptop that a) runs Linux or Mac OSX, or b) has a Secure Shell client installed (such as PuTTY, downloadable from http://www.chiark.greenend.org.uk/~sgtatham/putty/latest.html .) Some previous experience with the Linux command line would be helpful.


Speakers

Thursday August 17, 2017 3:00pm - 4:30pm
SEEC Main Auditorium

3:00pm

Visualizing Data with MatPlotLib

 

Matplotlib is a powerful graphing and visualization library in Python.  This tutorial aims to give users an introduction into graphing various types of data sets using the matplotlib library.  Additionally, it will provide a cursory-level introduction of other easy-to-use visualization-based Python libraries, such as seaborn and pandas (for visualization only), that utilize matplotlib on the backend.



Thursday August 17, 2017 3:00pm - 4:30pm
N128

3:00pm

Node Health Check (NHC) -- Adventures at Scale/ Informal Session

Since its original release by Berkeley Lab back in 2011, the LBNL Node Health Check (NHC) project has become the de facto standard in the HPC community for validating compute node health and has even gained traction on cluster masters and standalone servers for doing basic system/service management and validation.

As today's systems continue to grow larger and more complex, new challenges are emerging which makes the creation of a "one size fits all" utility next to impossible.  After a brief introduction/refresher for those who might not be familiar with the NHC project, this talk will discuss some experiences with deploying NHC on large-scale systems at Los Alamos National Lab, tips for getting the most out of NHC, and directions for future development and community involvement.



In addition to taking some time to have freeform discussion about any loose threads from the conference, we’ll also take some time to decide on a venue for our next sysadmin meetup, expected to take place fall 2017.

Speakers

Thursday August 17, 2017 3:00pm - 4:30pm
S125

3:00pm

Creating Singularity Containers for HPC Users

In this tutorial virtualization containers, which can be used to deploy applications and even entire computing environments to any Linux running host machine will be introduced. After an introduction, the Singularity container approach will be presented along with the environment necessary for Singularity container building on Windows, Mac or Linux. We will build a basic Singularity container and run it on your local laptop, and discuss how the container can then be moved to a different resource. We will wrap up with some more advanced container setup and ideas for how to deploy your own applications as a container.

Prerequisite:  basic Linux command and scripting knowledge. Participants should bring a laptop. Windows users should have GIT BashVirtualBox and Vagrant installed, as described at this page. Mac users should have Homebrew, VirtualBox and Vagrant via Homebrew installed as described here


Speakers

Thursday August 17, 2017 3:00pm - 4:30pm
C120C

3:00pm

Intel Libraries
  1. Intel Math Kernel Library (MKL)
  2. Intel Performance Primitives (IPP)
  3. Intel Threading Building Blocks (TBB)
  4. Intel Data Analytics Acceleration Library


Thursday August 17, 2017 3:00pm - 4:30pm
S372