As they have for nearly three decades, Berkeley Lab Computing Sciences Area (CSA) staff will share their expertise with the global HPC community at this year’s Supercomputing Conference (SC21), which takes place November 14-19 in St. Louis, Mo., with a virtual component for those who cannot attend in person.
CSA researchers, scientists, and engineers are participating in tutorials, workshops, panels, technical papers, and posters as part of the conference technical program. In addition, the U.S. Department of Energy’s (DOE) booth – which showcases the 17 DOE national laboratories – will feature high-level talks and technical demonstrations from the labs, some in person, some online.
Program highlights include the following:
- Gordon Bell Finalist: Billion atom molecular dynamics simulations of carbon at extreme conditions and experimental time and length scales. Authors include Rahulkumar Gayatri, an application performance specialist at NERSC who is involved with the Exascale Atomistic Capability for Accuracy, Length, and Time (EXAALT) project. The EXAALT library was instrumental in supporting this research.
- Best Paper, PMBS21 Workshop: Architectural Requirements for Deep-learning Workloads in HPC Environments. Authors: Khaled Ibrahim, Tan Nguyen, Hai Ah Nam, Wahid Bhimji, Steven Farrell, Leonid Oliker, Michael Rowan, Nick Wright, Samuel Williams
- Best Paper Finalist, Computational Biology: Accelerating Large Scale de Novo Metagenome Assembly Using GPUs. Authors: Muaaz Gul Awan, Steven Hofmeyr, Rob Egan, Nan Ding, Aydin Buluc, Jack Deslippe, Leonid Oliker, Katherine Yelick
- Featured Speaker, SC21 DOE Booth: Marcus Noack, “Optimal Autonomous Data Acquisition for Large-Scale Experimental Facilities.” Tuesday, Nov. 16, 11:30 a.m. CST, virtual presentation via WebEx.
- Technical Demos, SC21 DOE Booth: Nov. 16-18. All demos will be broadcast virtually (live and pre-recorded) via WebEx.
- TOP500: The TOP500 list of supercomputers serves as a “Who’s Who” in HPC. The 58th TOP500 list will be published in November 2021.
To help support this year’s hybrid approach, SC21 has chosen a new remote content provider to support the SC21 HUBB on desktop and mobile devices and is working to provide single sign-on and syncing of custom schedules across all devices. For more information about what sessions will be available via the SC21 HUBB, visit the Remote Participation FAQ page on the SC21 website. To see the full SC21 program, visit the SC21 website.
Below is a day-by-day guide to SC21 sessions and workshops featuring CSA staff and resources.
Sunday, November 14
Workshop: Women in HPC: Diversifying the HPC Community. 8 a.m. – 5 p.m. CST. Presentations include:
- “Exascale-Enabled Physical Modeling for next-Generation Microelectronics.” Authors/Presenters: Zhi (Jackie) Yao, Revathi Jambunathan, Yadong Zeng, Ann Almgren, Andrew Nonaka.
- “Balancing Particle GPU Memory Constraints in Nyx Cosmology Code.” Author.Presenter: Jean Sexton
- “Processing NERSC Python user data on GPUs and sharing the results in a public dashboard.” Authors/Presenters: Laurie Stephey, Rollin Thomas
Tutorial: In Situ Analysis and Visualization with SENSEI and Ascent. Presenters: Wes Bethel, Burlen Loring, Silvio Rizzi, Cyrus Harrison, Matt Larsen, Hank Childs, Terry Turton. 8 a.m. – 12 p.m. CST
Tutorial: Deep Learning at Scale. Presenters: Steven Farrell, Wahid Bhimji, Thorsten Kurth, Josh Romero, Aristeidis Tsaris, Junqi Yin, Peter Harrington. 8 a.m. – 5 p.m. CST
Tutorial: Performance Tuning with the Roofline Model on GPUs and CPUs. Presenters: Samuel Williams, Aleksandar Ilic, JaeHyuk Kwack, Neil Mehta, Zakhar Matveev, Max Katz, Kate Guseva, Kaleb Smith. 8 a.m. – 5 p.m. CST
Tutorial: The OpenMP Common Core: A “Hands-On” Introduction. Presenters: Timothy Mattson, Yun He, Alice Koniges, David Eder. 8 a.m. – 5 p.m. CST
Workshop: P3HPC: 2021 International Workshop on Performance, Portability, and Productivity in HPC. Presentations include the following:
- oneAPI Open-Source Math Library Interface. Authors/Presenters: Mariia Krainiuk, Mehdi Goli, Vincent Pascuzzi. 11:30 a.m. – 12 p.m. CST
- Case Study of Using Kokkos and SYCL as Performance-Portable Frameworks for MILC-DSLASH Benchmark on NVIDIA, AMD and Intel GPUs. Authors/Presenters: Amanda Dufek, Rahulkumar Gayatri, Neil Mehta, Douglas Doerfler, Brandon Cook, Yasaman Gadar, Carleton DeTar. 2:30 – 3:30 p.m. CST
- Evaluating Performance and Portability of a core bioinformatics kernel on multiple vendor GPUs. Authors/Presenters: Muhammad Haseeb, Nan Ding, Jack Deslippe, Muaaz Gul Awan. 3:30 – 3:55 p.m. CST
Workshop: HiPar21: 2nd Workshop on Hierarchical Parallelism for Exascale Computing. Organizers/Session Chairs: Francesco Rizzi, Sherry Li, Daisy Hollman, Lee Howes. 9 a.m. – 5:30 p.m. CST
Workshop: CANOPIE-HPC: Containers and New Orchestration Paradigms for Isolated Environments in HPC. Organizers: Andrew Younge, Shane Canon, Patrick Bridges. 9 a.m. – 5:30 p.m. CST
WACCPD 2021: Eighth Workshop on Accelerator Programming using Directives. Program Chairs: Sridutt Bhalachandra, Christopher Daley, Veronica G. Melesse Vergara. 9 a.m. – 5:30 p.m. CST. Presentations include the following:
- GPU offloading of a large-scale gyrokinetic particle-in-cell Fortran code: From OpenACC to OpenMP. Authors/Presenters: Qiheng Cai, Junyi Cheng, Yang Chen, Marcus Wagner, Christopher Daley, Dossay Oryspayev, Stefan Tirkas, Sophie Redd, Scott Parker. 12 – 12:30 p.m. CST
- Accelerating quantum many-body configuration interaction with directives. Authors/Presenters: Brandon Cook, Patrick Fasano, Pieter Maris, Chao Yang, Dossay Oryspayeve. 11:30 a.m. – 12 p.m. CST
- Achieving near native runtime performance and cross-platform performance portability for random number generation through SYCL interoperability. Authors/Presenters: Vincent R. Pascuzzi, Mehdi Goli. 3:30 – 4 p.m. CST
- Invited Panel: Publicly available directive test suites for heterogeneous architectures. Moderator: Chris Daley. 4:40 – 5:25 p.m. CST
Workshop: HPCSysPros21. Presentation: Benchmarking Panel. Presenters: Hai Ah Nam, Rory Kelly, Greg Bauer, Paul Ferrell. 11:10 a.m. – 12:30 p.m. CST
LLVM-HPC2021: The Seventh Workshop on the LLVM Compiler Infrastructure in HPC. Presentation: Facilitating CoDesign with Automatic Code Similarity Learning. Authors/Presenters: Tan Nguyen, Erich Strohmaier, John Shalf. 4:50 – 5:30 p.m. CST
Monday, November 15
Tutorial: Parallel I/O in Practice. Presenters: Rob Latham, Robert Ross, Brent Welch, Glenn Lockwood. 8 a.m. – 5 p.m. CST
Tutorial: User-Centric Automated Performance Analysis of Hybrid Parallel Program. Presenters: Abhinav Bhatele, David Boehme, Stephanie Brink, Jonathan Madsen, John Mellow-Crummey, Olga Pearce. 8 a.m. – 5 p.m. CST
Tutorial: Using Containers to Accelerate HPC. Presenters: Shane Canon, Marco De La Pierre, Carlos Eduardo Arango, Sameer Shende, Andrew Younge. 8 a.m. – 5 p.m. CST
Second International Workshop on Quantum Computing Software. Organizers: Wibe A. de Jong, Hausi A. Müller, Yuri Alexeev, Travis S. Humble. 9 a.m. – 5:30 p.m. CST
INDIS’21: 8th Workshop on Innovating the Network for Data-Intensive Science. Session Chairs: Mari a.m. Kiran, Sarah M. Neuwirth. 9 a.m. – 5:30 p.m. CST. Presentations include the following:
- NetGraf: An End-to-End Learning Network Monitoring Service. Authors/Presenters: Bashir Mohammed, Mari a.m. Kiran, Bjoen Enders. 11 – 11:30 a.m. CST
- Exploring the BBRv2 Congestion Control Algorithm for use on Data Transfer Nodes. Authors/Presenters: Brian Tierney, Eli Dart, Ezra Kissel, Eashan Adhikarla. 11:30 a.m. – 12 p.m. CST
- Lightning Talk: DynamicDeepFlow – Classification of Network Traffic Flow Changes using Unsupervised Learning. Presenter: Sheng Shen. 4:25 p.m. – 4:50 p.m. CST
- Panel: Experimental Networks (XNet) Planning and SCinet Integration. Moderator: Prasad Calyan. Panelists: Paul Ruth, Andrew Wiedlea, Mike Zink, Debbie Fligor. 4:40 – 5:20 p.m. CST
Workshop: Machine Learning in HPC Environments. 9 a.m. – 5:30 p.m. CST. Presentations include the following:
- MLPerf HPC: A Holistic Benchmark Suite for Scientific Machine Learning on HPC Systems. Authors/Presenters: Steven Farrell, et al
- HYPPO: A Surrogate-Based Multi-Level Parallelism Tool for Hyperparameter Optimization. Authors/Presenters: Vincent Dumont, Casey Garner, Anuradha Trivedi, Chelsea Jones, Vidya Ganapati, Juliane Mueller, Talita Perciano, Mari a.m. Kiran, Marc Day
PDSW21: Sixth International Parallel Data Systems Workshop. Presentations the following:
- New Challenges of Benchmarking All-FlashStorage for HPC. Authors/Presenters: Glenn Lockwood, Alberto Chiusole, Nicholas Wright. 10:30 – 10:55 a.m. CST
- I/O Bottleneck Detection and Tuning: Connecting the Dots using Interactive Log Analysis. Authors/Presenters: Jean Luca Bez, Houjun Tang, Bing Xie, David Williams-Young, Rob Latham, Rob Ross, Sarp Oral, Suren Byna. 11:20 – 11:45 a.m. CST
- Data-Aware Storage Tiering for Deep Learning. Authors/Presenters: Cong Xu, Suparna Bhattacharya, Martin Foltin, Suren Byna, Paolo Faraboschi. 11:45 a.m. – 12:10 p.m. CST
- SCTuner: An Auto-tuner Addressing Dynamic I/O Needs on Supercomputer I/O Sub-systems. Authors/Presenters: Houjun Tang, Bing Xie, Suren Byna, Philip Carns, Quincey Koziol, Sudarsun Kannan, Jay Lofstead, Sarp Oral. 3:30 – 3:55 p.m. CST
SuperCheck-SC21: Second International Symposium on Checkpointing for Supercomputing. Organizers: Zhengji Zhao, Rebecca Hartman-Baker, Gene Cooperman, Devesh Tiwari. 9 a.m. – 5:30 p.m. CST. Presentations include:
- Invited Talk: How real-time supercomputing is powering new models of experimental science. Presenter: Deborah Bard. 11:30 a.m. – 12 p.m. CST
- MANA-2.0: A Future-Proof Design for Transparent Checkpointing of MPI at Scale. Authors/Presenters: Yao Xu, Zhengji Zhao, Rohan Garg, Harsh Khetawat, Rebecca Hartman-Baker, Gene Cooperman. 12 p.m. – 12:30 p.m. CST
- Panel Discussion: Can checkpoint/restart tools ever keep pace with fast-changing HPC architectures, technologies, and workloads? Moderator: Rebecca Hartman-Baker. Panelists: Gene Cooperman, Bogdan Nicolae, Sarp Oral, Eric Roman, John Shalf. 3:30 – 4 p.m. CST
Eighth SC Workshop on Best Practices for HPC Training and Education. Presentations include:
- Best Practices for NERSC Training. Authors/Presenters: Yun (Helen) He, Rebecca Hartman-Baker. 9:30 – 10 a.m. CST
- Mentoring and Training Programs from Industry Context. Authors/Presenters: Scott Lathrop, Elizabeth Bautista, Julie Ma. 5 – 5:30 p.m. CST
Workshop: PMBS21: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems. Presentations include the following:
- Best Paper: Architectural Requirements for Deep-learning Workloads in HPC Environments. Authors: Khaled Ibrahim, Tan Nguyen, Hai Ah Nam, Wahid Bhimji, Steven Farrell, Leonid Oliker, Michael Rowan, Nick Wright, Samuel Williams. 9:30 – 10am CST
- An Extended Roofline Performance Model with PCI-E and Network Ceilings. Authors: Amanda Dufek, Jack Deslippe, Paul Lin, Brandon Cook, Charlene Yang, Jonathan Madsen. 11 – 11:30 a.m. CST
- Understanding power variation and its implications on performance optimization on the Cori supercomputer. Authors: Sridutt Bhalachandra, Brian Austin, Nicholas J. Wright. 12 – 12:30 p.m. CST
Tutorial: UPC++: An Asynchronous RMA/RPC Library for Distributed C++ Applications. Presenters: Katherine A. Yelick, Amir Kamil, Damian Rouson Dan Bonachea, Paul H. Hargrove. 1 – 5 p.m. CST
WORKS21: 16th Workshop on Workflows in Support of Large-Scale Science. Presentations include the following:
- Dynamic Heterogeneous Task Specification and Execution for In Situ Workflows. Authors/Presenters: Orcun Yildiz, Dmitriy Morozov, Bogdan Nicolae, Tom Peterka. 2 – 2:25 p.m. CST
- Science Capsule: Towards Sharing and Reproducibility of Scientific Workflows. Authors/Presenters: Devarshi Ghoshal, Ludovico Bianchi, Abdelilah Essiari, Drew Paine, Sarah S. Poon, Michal Beach, Alpha T. N’Diaye, Patrick Huck, Lavanya Ramakrishnan. 5 – 5:25 p.m. CST
Tuesday, November 16
Best Paper Finalist: Accelerating Large Scale de Novo Metagenome Assembly Using GPUs. Authors: Muaaz Gul Awan, Steven Hofmeyr, Rob Egan, Nan Ding, Aydin Buluc, Jack Deslippe, Leonid Oliker, Katherine Yelick. 10:30 a.m. – 11 a.m. CST
Birds of a Feather: Accelerating Storage I/O to GPUs. Session Leaders: CJ Newburn, Glenn Lockwood, Sven Oehme. 12:15 p.m. – 1:15 p.m. CST
State of the Practice: Non-Recurring Engineering Best Practices: A Case Study with the NERSC/NVIDIA OpenMP Contract. Authors: Christopher Daley, Annemarie Southwell, Rahulkumar Gayatri, Scott Biersdorff, Craig Toepfer, Guray Ozen, Nicholas Wright. 3:30 – 4 p.m. CST
Paper: Minimizing Privilege for Building HPC Containers. Authors: Reid Priedhorsky, R. Shane Canon, Timothy Randles, Andrew J. Younge. 4 p.m. – 4:30 p.m. CST
Paper: Dr.Top-k: Delegate-Centric Top-k Computation on GPUs. Authors: Anil Gaihre, Da Zheng, Scott Weitze, Lingda Li, Shuaiwen Leon Song, Caiwen Ding, Xiaoye S. Li, Hang Liu. 4:30 p.m. – 5 p.m. CST
Birds of a Feather: TOP500 Supercomputers. Session Leaders: Erich Strohmaier, Jack Dongarra, Horst Simon, Martin Meuer. 5:15 – 6:45 p.m. CST
Birds of a Feather: Analyzing Parallel I/O. Experience with the NERSC all-flash system. Presenter: Alberto Chiusole. 5:15 – 6:45 p.m. CST
Birds of a Feather: Liquid Cooling Challenges and Facility Experiences: What’s Next? Session Leaders: Chris DePrater, David Grant, Dale Sartor, Herbert Huber, Fumiyoshi Shoji, Dave Martinez. 5:15 p.m. – 6:45 p.m. CST
Birds of a Feather: HPC Graph Toolkits and the GraphBLAS Forum. Session Leaders: Antonino Tumeo, John Feo, Mahantesh Halappanar, José Moreira, Aydin Buluç, Tim Mattson. 5:15 p.m. – 6:45 p.m. CST
Birds of a Feather: Software Engineering and Reuse in Modeling, Simulation, and Data Analytics for Science and Engineering. Session Leaders: David Bernholdt, Anshu Dubey, Nasir Eisty, Sandra Gesing, Rinku Gupta, Axel H.ebl, Mozhgan Kabiri chimeh, Tomislay Maric, Marion Weinzierl. 5:15 – 6:45 p.m. CST
Poster: NERSC Interactive GPU metric Dashboard. Authors: Wentao Shi, Brandon Cook, Arghya Chatterjee, Johannes Blaschke. 8:30 a.m. – 5 p.m. CST
Poster: Detecting and Identifying Applications by Job Signatures. Authors: Jie Li, Brandon Cook, Yong Chen. 8:30 a.m. – 5 p.m. CST
Poster: Capturing Relationships Based on Structure Similarity for Self-describing Scientific Data Formats. Authors: Chenxu Niu, Wei Zhang, Suren Byna, Yong Chen. 8:30 a.m. – 5 p.m. CST
Poster: Enabling Combustion Science Simulations for Future Exascale Machines. Authors: Jon S. Rood, Marc T. Henry de Frahan, Marc S. Day, Hariswaran Sitaraman, Shashank Yellapantula, Bruce A. Perry, Ray Grout, Ann Almgren, Weiqun Zhang, Jackie H. Chen. 8:30 a.m. – 5 p.m. CST
Poster: GASNet-EX Memory Kinds: Support for Device Memory in PGAS Programming Models. Authors: Paul H. Hargrove, Dan Bonachea, Colin A. MacLean, Daniel Waters. 8:30 a.m. – 5 p.m. CST
Poster: Feature Reduction of Darshan Counters Using Evolutionary Algorithms. Authors: Neeraj Rajesh, Quincey Koziol, Suren Byna, Houjun Tang, Jean Luca Bez, Anthony Kougkas, Xian-He Sun. 8:30 a.m. – 5 p.m. CST
ACM Student Research Competition Poster: Performance Prediction of Large Data Transfers. Author: Jason Cheung. 8:30 a.m. – 5 p.m. CST
Wednesday, November 17
Gordon Bell Finalist: Billion atom molecular dynamics simulations of carbon at extreme conditions and experimental time and length scales. Authors: Kien Nguyen Cong, Jonathan T. Willman, Stan G. Moore, Anatoly B. Belonoshko, Rahulkumar Gayatri, Evan Weinberg, Mitchell A. Wood, Aidan P. Thompson, Ivan I. Oleynik. 4:30 p.m. – 5 p.m. CST
Birds of a Feather: Software APIs for Programming Exascale HPC. Session Leaders: Cory Snavely, Maxime Martinasso. 5:15 – 6:45 p.m. CST
Birds of a Feather: MLPerf: A Benchmark for Machine Learning. Presenter: Steve Farrell. 5:15 – 6:45 p.m. CST
Thursday, November 18
Panel: Resource Disaggregation in High-Performance Computing. Moderators: Balazs Gerofi, John Shalf. 10:30 a.m. – 12 p.m. CST. For more information, including details about the panelists who include leaders from RIKEN, Berkeley Lab, Columbia University, NIVIDIA, and Fujitsu, go to this website.
Paper: Cuttlefish: Library for Achieving Energy Efficiency in Multicore Parallel Programs. Authors: Sunil Kumar, Akshat Gupta, Vivek Kumar, Sridutt Bhalachandra. 11:30 a.m. – 12 p.m. CST
Panel: Strategies for Working Remotely: Sustainable Hybrid Approaches for HPC. Moderator: Elaine Raybourn. Panelists: Sadaf Alam, Christian Bischof, Helen Cademartori, Devin Hodge, Kengo Nakajima, Pat Quillen. 3:30 – 5 p.m. CST
Paper: Storage and Application Characteristics. Session Chair: Suren Byna. 3:30 – 5 p.m. CST
Paper: Empirical Evaluation of Circuit Approximations on Noisy Quantum Devices. Authors: Ellis Wilson, Frank Mueller, Lindsay Bassman, Costin Iancu. 3:30 – 4 p.m. CST
Paper: Exploiting User Activeness for Data Retention in HPC Systems. Authors: Wei Zhang, Suren Byna, Hyogi Sim, SangKeun Lee, Sudharshan Vazhkudai, Yong Chen. 3:30 – 4 p.m. CST
Friday, November 19
XLOOP 2021: The 3rd Annual Workshop on Extreme-Scale Experiment-in-the-Loop Computing. Presentations include the following:
- ALS Share, a Lightweight Data Sharing Service for a Synchrotron Radiation Facility. Authors/Presenters: Soren Bear, Dula Parkinson, Lisa Gerhardt, Mel Sibony, Bjoern Enders. 9 – 9:20 a.m. CST
- High-Performance Hybrid-Global-Deflated-Local Optimization with Applications to Active Learning. Author/Presenter: Marcus Noack. 10:20 – 10:50 a.m. CST
The 4th Annual PAW-ATM Workshop. Presentations include the following:
- Demonstrating UPC++/Kokkos Interoperability in a Heat Conduction Simulation. Authors/Presenters: Daniel Waters, Colin A. MacLean, Dan Bonachea, Paul H. Hargrove. 9:20 – 9:40 a.m. CST
- Optimization of Asynchronous Communication Operations through Eager Notifications. Authors/Presenters: Amir Kamil, Dan Bonachea. 11:10 – 11:30 a.m. CST
Panel: Great Edge-pectations: How Edge and Exascale Found Love. Moderator: Pete Beckman. Panelists: Lavanya Ramakrishnan, Rick Stevens, Satoshi Matsuoka, Eric Van Hensbergen. 10:30 a.m. – 12 p.m. CST
Workshop: Urgent HPC: HPC for Urgent Decision Making. Presentation: Experiences with Cross-Facility Real-Time Light Source Data Analysis Workflows. Authors/Presenters: Anna Giannakou, Johannes Blaschke, Deborah Bard, Lavanya Ramakrishnan. 11:20 – 11:40 a.m. CST