2018 ALCF Computational Performance Workshop

Workshop Beginner, Intermediate

CT

Argonne National Laboratory
Building 240
ALCF Computational Performance Workshop

Our intensive three-day workshop is intended to help you achieve computational readiness on current and future many-core architectures.

Work with staff experts to boost your performance on our resources

Held at the ALCF from May 15–17, our intensive three-day workshop is intended to help you achieve computational readiness on current and future HPC architectures.

  • Get 10+ full hours of dedicated hands-on time with ALCF and industry professionals 
  • Learn advanced techniques and tools to enhance code performance and increase your data science skills
  • Test and debug as you go with exclusive full-system reservations on Theta and Mira
  • Prepare for an allocation

The ALCF Computational Performance Workshop is intended for experienced HPC users with near-term goals of applying for a major allocation award. Registrants will be reviewed for experience level and will be asked to provide both a research summary and goals for attending.

Workshop has ended. Please view the agenda to see the slide presentations.

REGISTRATION DEADLINES:

  • Foreign Nationals: April 23, 2018
  • U.S. Citizen: April 30, 2018

* Please note: It is difficult for limousine and taxi drivers who are not U.S. citizens to enter the lab. Due to Argonne's strict entry restrictions, we strongly recommend that you contact United Limousine for your ground transportation needs. Be sure to notify United Limousine that you are traveling to Argonne National Laboratory.
United Limousine Reservations: 1-630-969-3865 
Airport Arrivals: 1-800-331-9037 or 1-630-969-3865

Agenda

Day 1 (Tues, May 15)
8:30 Shuttle from Guest House to TCS, Bldg. 240
9:00-9:10 Welcome to the Computational Performance Workshop - Mike Papka, ALCF
9:10-9:15 Introduction and Safety Review - Ray Loy, ALCF
9:15-10:00 ALCF System Architectures: Software and Job Submission - Chris Knight, ALCF
10:00-11:15

ALCF System Architectures: Hardware and Single-Core/Node Vectorization, Memory
 

Part 1: Overview and BG/Q - Scott Parker, ALCF
 

Part 2: KNL - John Pennycook, Intel
11:15-12:00 Hands-on: Performance Design Collaboration
12:00-1:00 ALCF Allocation Programs - Katherine Riley, ALCF
1:00-2:00 Performance Optimization II: Multi-Node Communication/Network Topology and Libraries
  Part 1: Overview and BG/Q - William Scullin, ALCF
  Part 2: Optimizing MPI on Theta - Sudheer Chunduri, ALCF
2:00-3:30 ALCF Performance Tools
 

Part 1: Profiling with HPCToolkit - Mark Krentel, Rice University
 

Part 2: TAU Performance Analysis - Sameer Shende, Paratools
3:30-5:00

Hands-on Session
5:15

Shuttle from TCS to Guest House

Day 2 (Wed, May 16)
8:30 Shuttle from Guest House to TCS, Bldg. 240
9:00-10:15 ALCF Theta Performance Tools
  Part 1: VTune Quick Start and Roofline Analysis - John Pennycook, Intel and Paulius Velesko, Intel
  Part 2: Cray Performance Tools - Colleen Bertoni, ALCF
10:15-10:30 Break
10:30-11:00 Debugging on ALCF Systems - Ray Loy, ALCF
11:00-11:50 Hands-on: Performance Benchmarking Collaboration
11:50-12:00 Group Photo
12:00-1:00 Single-node optimization towards more efficient and productive quantum Monte Carlo simulations - Ye Luo, ANL
1:00-1:45 ALCF Data and Learning Frameworks - Taylor Childers, ALCF
1:45-2:30 Optimizing I/O - Rick Zamora, ALCF
2:30-5:00 Hands-on Session
5:15 Shuttle from TCS to Guest House

Day 3 (Thurs, May 17)
8:30 Shuttle from Guest House to TCS, Bldg. 240
9:00-9:45 OpenMP on KNL - John Pennycook, Intel
9:45-10:30 Math Libraries - Michael D'Mello, Intel
10:30-10:35 ALCF Next Steps - Ray Loy, ALCF
10:35-10:45 Break
10:45-12:00 Hands-on: Performance Enhancements Collaboration
12:00-1:00 Visualization at ALCF - Joe Insley, ALCF and Silvio Rizzi, ALCF
1:00-1:45

Ensemble Jobs and Workflows - Tom Uram, ALCF
1:45-5:00

Hands-on Session
 

NOTE: No shuttle to Guest House. For assistance visit registration desk.