News and Announcements

ICL Spring Graduates

May 11th was graduation day for Chongxiao Cao (PhD), Heike Jagode (PhD), Khairul Kabir (PhD), Sangamesh Ragate (MS), and Wei Wu (PhD).

Chongxiao will begin working for Intel as a software development engineer this month. Heike continues her role as ICL’s Research Lead for the Performance API (PAPI) project. Khairul has joined Nvidia Corporation as a machine learning engineer. Sangamesh starts with Colfax International this month as a research engineer. And Wei is joining Los Alamos National Lab to be a research scientist.

News Post Covers ICL’s Role in the Exascale Computing Project

Jack Dongarra recently discussed with Mike Bernhardt of the US Department of Energy’s Exascale Computing Project information about ICL’s software development projects in a special guest post on the insideHPC site.

Conference Reports

ICLer Assists with Broader Engagement Program at CSE17

During the 2017 SIAM Conference on Computational Science and Engineering (CSE17) in Atlanta, in February, ICL’s Stephen Wood was among the more than fifty members of the CSE community who contributed to the Sustainable Horizons Institute’s Broader Engagement (BE) program efforts to involve students from underrepresented groups and early career scientists in the conference.

Out of gratitude for the sage guidance he himself received in the past from more experienced practitioners, Stephen volunteered to lead a guided affinity group on uncertainty quantification for computational fluid dynamics in sustainable energy applications. Uncertainty quantification is the practice of determining the possible variations in quantities of interest associated with model calculations.

A quote from Stephen is included in an article about the BE program’s presence at CSE17 published in the May issue of SIAM News.

ASC17 Student Supercomputer Challenge

Jack Dongarra served as a member of the Expert Committee and judge of the Asia Supercomputing Community’s 2017 Student Supercomputing Challenge (ASC17), held recently at the National Supercomputing Center in Wuxi, China.

ASC17, the world’s largest supercomputing competition, received applications from 230 universities across the globe; twenty teams made the final round. Emerging from that upper echelon of competitors to capture the title of grand champion was Tsinghua University.

Teams in the final round were required to independently design a supercomputing system under the precondition of a limited 3000W power consumption. In addition, they had to operate and optimize standard international benchmark tests and a variety of leading-edge scientific and engineering applications such as artificial-intelligence-(AI)-based transport prediction, genetic assembly, and materials science.

Tsinghua University completed deep parallel optimization of the high-resolution maritime data simulation mode MASNUM on the Chinese TaihuLight machine. The team’s achievement in expanding the original program up to 10,000 cores and accelerating the program by 392 times earned it the e Prize award.

The competition’s runner-up, Beihang University, distinguished itself through a superlative performance in the popular AI field. And first-time finalist, the team from Weifang University, constructed a highly optimized advanced heterogeneous supercomputing system with Inspur’s supercomputing server and ran the international HPL benchmark test, setting a new world record of 31.7 TFLOPS for floating-point computing speed. That feat gained the team the award for best computing performance.

An article in HPCwire contains a quote from Jack in which he emphasizes how the competition benefits the students by enhancing their scientific knowledge and giving them the unique opportunity to work on the powerful TaihuLight platform.

The Radio Free HPC podcast from insideHPC provides a review of ASC17.

GPU Technology Conference

Approximately 3,000 people—including ICL’s Azzam Haidar, Piotr Luszczek, and Stan Tomov—converged on Silicon Valley May 8–11 for the GPU Technology Conference, the largest and most important event of the year for GPU developers.

Azzam and Stan gave a talk on Magma Tensors and Batched Computing for Accelerating Applications on GPUs, and Piotr on Half Precision Benchmarking for HPC.

The ICLers also met with Nvidia teams to continue our close interaction with them and to discuss how we can help each other.

In addition, Azzam and Stan went to Lawrence Livermore National Laboratory to meet with their collaborators from the Exascale Computing Project co-design Center for Efficient Exascale Discretizations (CEED).

CERFACS 30-Year Conference

Jack Dongarra gave a talk on May 12th at the Centre of Basic and Applied Research (CERFACS) 30-Year Conference in Toulouse, France.

The title of his talk was “A Look at What has Changed in the Last 30 years for Computers and Dense Linear Algebra Software” during the sessions on HPC and Fluid Mechanics.

CERFACS is a basic and applied research center that specializes in modeling and numerical simulation. Using its facilities and expertise in high-performance computing, CERFACS deals with major scientific and technical research problems of public and industrial interest.

The 12th Scheduling for Large Scale Systems Workshop

ICL was host May 24–26 to The 12th Scheduling for Large Scale Systems Workshop. The invitation-only event had twenty-six attendees.

As was the case in previous editions, the workshop was structured as a set of thematic half-day sessions. Brief presentations were complemented by dedicated sessions for information discussions and exchanges aimed at tackling challenging problems.

Hiking on the Cumberland Trail


During a social event, some of the participants from the 12th Scheduling for Large-Scale Systems Workshop took the opportunity to hike the Cumberland Mountain segment of the Cumberland Trail above LaFollette. Thanks to Frédéric Vivien for capturing these images and sharing them.

Interview

Diana Fayad Then

Diana Fayad

Where are you from originally?

I’m from Lebanon. I spent a couple of years in France doing my studies before I came here.

What is your educational background?

My high school major was biology, which I love. But motivated by my curiosity and the challenge, I chose to go into computer science. I got my bachelor’s degree in computer science, and then, after a couple of years of work and gaining experience, I continued my studies, with high-performance computing as my chosen area. Since the University of Versailles had the only HPC master’s degree program in France, I went there. Versailles is a very calm and beautiful place to live.

Where did you work before joining ICL?

After completing my bachelor’s degree, I worked at a software development company as a developer for about a year. And next I spent almost a year doing software and database management at a luxury sanitary ware company.

What drew you to ICL?

After talking with Dr. Dongarra, my master’s program advisor suggested I come here. Of course, Dr. Dongarra is world renowned for his work in HPC, and ICL has an international reputation for creating the packages and interfaces we worked on in school. I thought coming here would be a very good opportunity for me to gain valuable experience and prove myself.

What is your primary role here?

I am a research assistant with Jakub Kurzak, and we are working on the BONSAI project.

What are your interests/hobbies outside of work?

I enjoy reading books, painting, hiking, swimming, working out, discovering new cultures, and traveling.

Tell us something about yourself that might surprise us.

I solved Einstein’s Intelligence Quiz—or as it’s called, Einstein’s Puzzle—in less than 20 minutes. Only 2 percent of people have been able to solve it.

Recent Papers

  1. Faverge, M., J. Langou, Y. Robert, and J. Dongarra, Bidiagonalization and R-Bidiagonalization: Parallel Tiled Algorithms, Critical Paths and Distributed-Memory Implementation,” IEEE International Parallel and Distributed Processing Symposium (IPDPS), Orlando, FL, IEEE, May 2017. DOI: 10.1109/IPDPS.2017.46  (328.15 KB)
  2. Aupy, G., A. Benoit, L. Pottier, P. Raghavan, Y. Robert, and M. Shantharam, Co-Scheduling Algorithms for Cache-Partitioned Systems,” 19th Workshop on Advances in Parallel and Distributed Computational Models, Orlando, FL, IEEE Computer Society Press, May 2017. DOI: 10.1109/IPDPSW.2017.60  (584.76 KB)
  3. Jagode, H., Dataflow Programming Paradigms for Computational Chemistry Methods,” Innovative Computing Laboratory Technical Report, no. ICL-UT-17-01, Knoxville, TN, University of Tennessee, May 2017.
  4. Abdelfattah, A., A. Haidar, S. Tomov, and J. Dongarra, Fast Cholesky Factorization on GPUs for Batch and Native Modes in MAGMA,” Journal of Computational Science, vol. 20, pp. 85–93, May 2017. DOI: 10.1016/j.jocs.2016.12.009  (3.6 MB)
  5. Tomov, S., and A. Haidar, MAGMA Tensors and Batched Computing for Accelerating Applications on GPUs , San Jose, CA, GPU Technology Conference (GTC17), Presentation in Session S7728, May 2017.  (11.12 MB)
  6. Benoit, A., L. Pottier, and Y. Robert, Resilient Co-Scheduling of Malleable Applications,” International Journal of High Performance Computing Applications (IJHPCA), May 2017. DOI: 10.1177/1094342017704979  (1.62 MB)
  7. Yamazaki, I., S. Nooshabadi, S. Tomov, and J. Dongarra, Structure-aware Linear Solver for Realtime Convex Optimization for Embedded Systems,” IEEE Embedded Systems Letters, vol. 9, issue 3, pp. 61–64, May 2017. DOI: 10.1109/LES.2017.2700401  (339.11 KB)
  8. Dongarra, J., S. Tomov, P. Luszczek, J. Kurzak, M. Gates, I. Yamazaki, H. Anzt, A. Haidar, and A. Abdelfattah, With Extreme Computing, the Rules Have Changed,” Computing in Science & Engineering, vol. 19, issue 3, pp. 52-62, May 2017. DOI: 10.1109/MCSE.2017.48  (485.34 KB)
  9. Kabir, K., A. Haidar, S. Tomov, A. Bouteiller, and J. Dongarra, A Framework for Out of Memory SVD Algorithms,” ISC High Performance 2017, pp. 158–178, June 2017. DOI: 10.1007/978-3-319-58667-0_9  (393.22 KB)
  10. Gates, M., J. Kurzak, P. Luszczek, Y. Pei, and J. Dongarra, Autotuning Batch Cholesky Factorization in CUDA with Interleaved Layout of Matrices,” Parallel and Distributed Processing Symposium Workshops (IPDPSW), Orlando, FL, IEEE, June 2017. DOI: 10.1109/IPDPSW.2017.18
  11. Gates, M., P. Luszczek, A. Abdelfattah, J. Kurzak, J. Dongarra, K. Arturov, C. Cecka, and C. Freitag, C++ API for BLAS and LAPACK,” SLATE Working Notes, no. 02, ICL-UT-17-03: Innovative Computing Laboratory, University of Tennessee, June 2017.  (1.12 MB)
  12. Abdelfattah, A., A. Haidar, S. Tomov, and J. Dongarra, Factorization and Inversion of a Million Matrices using GPUs: Challenges and Countermeasures,” Procedia Computer Science, vol. 108, pp. 606–615, June 2017. DOI: 10.1016/j.procs.2017.05.250  (643.44 KB)
  13. Benoit, A., F. Cappello, A. Cavelan, Y. Robert, and H. Sun, Identifying the Right Replication Level to Detect and Correct Silent Errors at Scale,” 2017 Workshop on Fault-Tolerance for HPC at Extreme Scale, Washington, DC, ACM, June 2017. DOI: 10.1145/3086157.3086162  (865.68 KB)
  14. Yamazaki, I., M. Hoemmen, P. Luszczek, and J. Dongarra, Improving Performance of GMRES by Reducing Communication and Pipelining Global Collectives,” Proceedings of The 18th IEEE International Workshop on Parallel and Distributed Scientific and Engineering Computing (PDSEC 2017), Best Paper Award, Orlando, FL, June 2017. DOI: 10.1109/IPDPSW.2017.65  (453.66 KB)
  15. Abdelfattah, A., A. Haidar, S. Tomov, and J. Dongarra, Novel HPC Techniques to Batch Execution of Many Variable Size BLAS Computations on GPUs,” International Conference on Supercomputing (ICS '17), Chicago, Illinois, ACM, June 2017. DOI: 10.1145/3079079.3079103  (1.04 MB)
  16. Benoit, A., A. Cavelan, V. Le Fèvre, and Y. Robert, Optimal Checkpointing Period with replicated execution on heterogeneous platforms,” 2017 Workshop on Fault-Tolerance for HPC at Extreme Scale, Washington, DC, IEEE Computer Society Press, June 2017. DOI: 10.1145/3086157.3086165  (1.02 MB)
  17. Dong, T., A. Haidar, S. Tomov, and J. Dongarra, Optimizing the SVD Bidiagonalization Process for a Batch of Small Matrices,” International Conference on Computational Science (ICCS 2017), Zurich, Switzerland, Procedia Computer Science, June 2017. DOI: 10.1016/j.procs.2017.05.237  (364.95 KB)
  18. Abalenkovs, M., N. Bagherpour, J. Dongarra, M. Gates, A. Haidar, J. Kurzak, P. Luszczek, S. Relton, J. Sistek, D. Stevens, et al., PLASMA 17 Performance Report,” Innovative Computing Laboratory Technical Report, no. ICL-UT-17-11: University of Tennessee, June 2017.  (7.57 MB)
  19. Abalenkovs, M., N. Bagherpour, J. Dongarra, M. Gates, A. Haidar, J. Kurzak, P. Luszczek, S. Relton, J. Sistek, D. Stevens, et al., PLASMA 17.1 Functionality Report,” Innovative Computing Laboratory Technical Report, no. ICL-UT-17-10: University of Tennessee, June 2017.  (1.8 MB)
  20. Haidar, A., H. Jagode, A. YarKhan, P. Vaccaro, S. Tomov, and J. Dongarra, Power-Aware HPC on Intel Xeon Phi KNL Processors , Frankfurt, Germany, ISC High Performance (ISC17), Intel Booth Presentation, June 2017.  (5.87 MB)
  21. Anzt, H., M. Gates, J. Dongarra, M. Kreutzer, G. Wellein, and M. Kohler, Preconditioned Krylov Solvers on GPUs,” Parallel Computing, June 2017. DOI: 10.1016/j.parco.2017.05.006  (1.19 MB)
  22. Abdelfattah, A., H. Anzt, A. Bouteiller, A. Danalis, J. Dongarra, M. Gates, A. Haidar, J. Kurzak, P. Luszczek, S. Tomov, et al., Roadmap for the Development of a Linear Algebra Library for Exascale Computing: SLATE: Software for Linear Algebra Targeting Exascale,” SLATE Working Notes, no. 01, ICL-UT-17-02: Innovative Computing Laboratory, University of Tennessee, June 2017.  (2.8 MB)
  23. Dongarra, J., S. Hammarling, N. J. Higham, S. Relton, P. Valero-Lara, and M. Zounon, The Design and Performance of Batched BLAS on Modern High-Performance Computing Systems,” International Conference on Computational Science (ICCS 2017), Zürich, Switzerland, Elsevier, June 2017. DOI: DOI:10.1016/j.procs.2017.05.138  (446.14 KB)
  24. Anzt, H., J. Dongarra, G. Flegar, E. S. Quintana-Orti, and A. E. Thomas, Variable-Size Batched Gauss-Huard for Block-Jacobi Preconditioning,” International Conference on Computational Science (ICCS 2017), vol. 108, Zurich, Switzerland, Procedia Computer Science, pp. 1783-1792, June 2017. DOI: 10.1016/j.procs.2017.05.186  (512.57 KB)

Recent Conferences

  1. MAY
    -
    Azzam Haidar
    Azzam
    Piotr Luszczek
    Piotr
    Stanimire Tomov
    Stan
    Azzam Haidar, Piotr Luszczek, Stanimire Tomov
  2. MAY
    -
    David Bailey
    David
    Ichitaro Yamazaki
    Ichitaro
    Jack Dongarra
    Jack
    Piotr Luszczek
    Piotr
    David Bailey, Ichitaro Yamazaki, Jack Dongarra, Piotr Luszczek
  3. MAY
    -
    TESSE Workgroup Meeting Arlington, Virginia
    Damien Genet
    Damien
    George Bosilca
    George
    Thomas Herault
    Thomas
    Damien Genet, George Bosilca, Thomas Herault
  4. JUN
    -
    2017 GPU Hackathon Brookhaven Riverhead, New York
    Piotr Luszczek
    Piotr
    Piotr Luszczek
  5. JUN
    -
    ICCS17 / ISC17 Zurich, Switzerland
    Stanimire Tomov
    Stan
    Stanimire Tomov
  6. JUN
    -
    ISC17 Frankfurt, Germany
    George Bosilca
    George
    George Bosilca
  7. JUN
    Anthony Danalis
    Anthony
    Anthony Danalis

Upcoming Conferences

  1. JUL
    -
    JHPCN Project Meeting Tokyo, Japan
    Ichitaro Yamazaki
    Ichitaro
    Ichitaro Yamazaki
  2. JUL
    -
    SIAM AN Meeting Pittsburgh, Pennsylvania
    Ichitaro Yamazaki
    Ichitaro
    Ichitaro Yamazaki
  3. JUL
    -
    7th JLESC Workshop Urbana, Illinois
    Aurelien Bouteiller
    Aurelien
    Damien Genet
    Damien
    George Bosilca
    George
    Piotr Luszczek
    Piotr
    Terry Moore
    Terry
    Thomas Herault
    Thomas
    Aurelien Bouteiller, Damien Genet, George Bosilca, Piotr Luszczek, Terry Moore, Thomas Herault

Recent Lunch Talks

  1. MAY
    5
    Lynne Parker
    Lynne Parker
    EECS
    Some Lessons Learned from My 2-Year Stint at NSF: The Abbreviated Version PDF
  2. MAY
    19
    Chris DavisSophie Voisin
    Chris Davis and Sophie Voisin
    ORNL
    Combining NVIDIA Docker and Databases to Enhance Agile Development and Optimize Resource Allocation PDF
  3. MAY
    26
    Markus Eisenbach
    Markus Eisenbach
    ORNL
    A Linear Scaling Multiple Scattering Code for First Principles Calculations of the Ground State and Statistical Physics of Materials PDF
  4. JUN
    30
    Saeid Nooshabadi
    Saeid Nooshabadi
    Michigan Technological University
    Algorithm and Architecture for Density Estimation for Data Intensive Application in High Dimension PDF

Upcoming Lunch Talks

  1. JUL
    28
    Sushil Prasad
    Sushil Prasad
    National Science Foundation
    Parallel Processing over Spatial-Temporal Datasets from Geo, Bio, Climate, and Social Science Communities: A Research Roadmap PDF

congratulations

Dorian Arnold

Best wishes to ICL alumnus Dorian Arnold in his new position at Emory University!

Dates to Remember

2017 ICL Retreat

Tremont Lodge in Townsend, TN, will be the location for this year’s ICL Retreat, which is set for August 17–18.

Open Positions at ICL

ICL is Hiring

Please refer your best and brightest to the following job opportunities:

      Research Position in Numerical Linear Algebra
      Research Position in Performance Measurement and Modeling
      Summer 2017 Internships