News and Announcements

Jack Dongarra: SC Perennial

Jack at Podium

The upcoming SC15 conference will be the 28th in the series, having made its debut in 1988. It is hard to believe that—way back in 1988—$20 million could only get you a couple of GFLOPS from a Cray Y-MP, and a gallon of gasoline was around $1. Setting all that aside, the inaugural SC conference was significant for being the first of its kind to bring so many experts in High Performance Computing together under one roof.

This significance has grown, not waned, in the nearly 3 decades since Jack Dongarra attended his first SC meeting. In fact, Jack has attended every SC meeting since 1988, making him an “SC Perennial.” Jack says that, “attending the SC conference each year is like a family reunion—a homecoming—with the added benefit of connecting with so many old friends.” Here’s to the continued success of SC, and to Jack attending many more iterations of the SC conference!

Read more about the “SC Perennials” at the SC15 Official Blog.

2015 ICL Retreat

  • DSC03594
  • DSC03596
  • DSC03597
  • DSC03598
  • DSC03595
  • DSC03593
  • DSC03589
  • DSC03587
  • DSC03586
  • DSC03583
  • DSC03582
  • DSC03581
  • DSC03572
  • DSC03570
  • DSC03569
  • 2015-group

 
This year’s ICL retreat moved to the Tremont Lodge & Resort in Townsend, Tennessee. This newly renovated venue provided a good platform for 2 days of talks, which covered student projects and summer internships, the lab’s progress in the areas of linear algebra, distributed computing, benchmarking, and performance analysis, along with recaps of administrative procedures.

There was some fun to be had as well, and the Tremont’s pool was the perfect place to cool down in the August heat after partaking in some delicious food and libations at Miss Lilly’s Cafe. Serving as a kickoff to the fall semester, the retreat encourages us to keep our focus sharp, our determination strong, and to have another great year at ICL!

Memristor Funded

The Memristive Accelerator for Extreme Scale Linear Solvers project was recently funded as part of the NSF EAGER program. The project, part of a 1-year award and collaboration between the University of Tennessee (ICL) and the University of Rochester, will investigate the feasibility of using the new electronic element—the memristor—for solving systems of linear equations. The solver is based on Kirchoff’s law and the physical feature of the memristor that allows it to compute in analog fashion. Congratulations to the Linear Algebra Group for landing this award!

EMBRACE Funded

The Evolvable Methods for Benchmarking Realism through Application and Community Engagement (EMBRACE) project is part of the NSF BRAP call to engage a broad community in producing benchmarks that inform hardware makers and users about the feasibility of supercomputing platforms for scientific applications. Even though there are many benchmarks available for testing and evaluating computer systems, EMBRACE aims to evaluate their value for the community at large and assess how these benchmarks might or should evolve for future systems and infrastructure. EMBRACE is a collaborative project between the University of Tennessee (ICL) and Georgia Tech. Congratulations to the Performance Analysis and Benchmarking Group for landing this award!

SparseKaffe Funded

The SparseKaffe project (Sparse direct methods via Run-time Scheduling and Execution of Kernels with Auto-tunable and Frequency-scaling Features for Energy-aware computing on heterogeneous architectures) was recently funded! As part of this 3-year NSF CSR award, the SparseKaffe team—consisting of PIs from the University of Tennessee (ICL), Texas A&M, and the University of Florida—will develop high-performance parallel sparse direct methods (multifrontal QR, Cholesky, and LU factorization) with irregular and hierarchical structure that can exploit Multiple Hybrid Multicore Processors to achieve orders of magnitude gains in computational performance, while also paying careful attention to energy constraints. Such algorithms will enable the team to create a robust suite of open-source, high-performance software that others can use to solve their applications. Congratulations to the Linear Algebra Group for their success!

PAPI-EX Funded

PAPI-EX, the next iteration of the Performance API, was recently funded by a 4-year NSF SI2 grant. PAPI-EX will enhance and extend PAPI to measure a wide range of new hardware and software events on extreme scale computing platforms. The upcoming improvements to PAPI will allow it to continue its essential role in enabling HPC performance analysis and optimization. This holds especially true for the large and diverse collection of PAPI-based performance tools on which many users now depend, including TAU, HPCToolkit, Open|SpeedShop, Vampir, Scalasca, CrayPat, and others. Congratulations to the Performance Analysis and Benchmarking Group for this award!

Conference Reports

HPCC 2015

The 17th IEEE International Conference on High Performance Computing and Communications (HPCC 2015) attracted dozens of attendees and speakers this year, including ICL’s own Jack Dongarra and Piotr Luszczek. HPCC 2015, held at the Newark Airport Hilton on August 24th through 26th, provided a forum for engineers and scientists in academia, industry, and government to discuss new ideas, research results, applications, and experience on all aspects of high performance computing and communications.

Jack gave a keynote talk on “Architecture-aware Algorithms and Software for Peta and Exascale Computing” on Monday, August 24th. On Tuesday, Piotr presented a paper, “Flexible Linear Algebra Development and Scheduling with Cholesky Factorization.” Notably, HPCC 2015 was actually in Newark, which sounds kind of like New York, but decidedly isn’t.

Euro-Par 2015

IMG_3427

ICL’s Hartwig Anzt made his way to Euro-Par 2015, which was held at the Vienna University of Technology on August 24th through 28th. Euro-Par, organized by the Research Group for Parallel Computing, is billed as Europe’s premier conference for all aspects of parallel and distributed processing.

Hartwig presented a paper, “Iterative Sparse Triangular Solves for Preconditioning,” on Wednesday, but he certainly wasn’t the only ICLer in attendance. As pictured above, ICL alumni Emmanuel Jeannot, Hatem Ltaief, Karl Fuerlinger, and Felix Wolf were also at Euro-Par.

The conference’s formal reception was held in Vienna’s City Hall. The minister for communication and technology gave a short speech, and finger food and champaign were available for everyone. There was more than one corridor in City Hall, however, and if one were to take a “wrong turn” (or “better turn” as it so happens), then one might have ended up in a gala dinner with the mayor and governor of Vienna. We suspect Hartwig et al. represented ICL well and were at ease rubbing elbows with the local government.

Interview

Joe Dorris Then

Joe Dorris

Where are you from, originally?

I am from Chattanooga, Tennessee.

Can you summarize your educational background?

I received my B.S. in Computer Engineering from UT and am currently working on a Master’s in Computer Science as part of the 5 year BS/MS program.

Tell us how you first learned about ICL.

I first heard about ICL while attending SC14 as part of the UT/NICS Student Cluster Competition Team. When we were not competing, we spent some time exploring the conference and attending talks. One of the presentations I attended was the TOP500 Supercomputers presentation by Dr. Dongarra.

What made you want to work for ICL?

After attending SC14 and seeing all of the exicting opportunities available in high performance computing, I decided that I wanted to continue working in this area. ICL was a great opportunity for me to work on some interesting research topics with some of the leaders in the field.

What are you working on while at ICL?

I am working with the Linear Algebra group and helping out with the DARE and RaPyDLI projects.

If you weren’t working at ICL, where would you like to be working and why?

I really enjoy trying to come up with business ideas and new products. I think it would be exciting to attempt entrepreneurship. I worked on a social network app with some other students at UT for a while, but we eventually realized that the project would require more work than we were able to commit to during school.

What are your interests/hobbies outside of work?

Some friends from Chattanooga and I formed an intramural team called Chatt-Town’s Finest that competes in many of the intramural sports offered at UT. We have been growing and adding members for the past 4 years, and we won the All-Points Competition last year. I also like watching sports, especially the Tennessee Titans and the Vols.

Tell us something about yourself that might surprise people.

I own two sugar gliders (Rosco and Felix) that I bought from a flea market.

Recent Papers

  1. YarKhan, A., A. Haidar, C. Cao, P. Luszczek, S. Tomov, and J. Dongarra, Cholesky Across Accelerators,” 17th IEEE International Conference on High Performance Computing and Communications (HPCC 2015), Elizabeth, NJ, IEEE, August 2015.
  2. Haidar, A., A. YarKhan, C. Cao, P. Luszczek, S. Tomov, and J. Dongarra, Flexible Linear Algebra Development and Scheduling with Cholesky Factorization,” 17th IEEE International Conference on High Performance Computing and Communications, Newark, NJ, August 2015.  (494.31 KB)
  3. Anzt, H., E. Chow, and J. Dongarra, Iterative Sparse Triangular Solves for Preconditioning,” EuroPar 2015, Vienna, Austria, Springer Berlin, August 2015. DOI: 10.1007/978-3-662-48096-0_50  (322.36 KB)
  4. Song, F., and J. Dongarra, A Scalable Approach to Solving Dense Linear Algebra Problems on Hybrid CPU-GPU Systems,” Concurrency and Computation: Practice and Experience, vol. 27, issue 14, pp. 3702-3723, September 2015. DOI: 10.1002/cpe.3403  (8.16 MB)
  5. Jagode, H., A. Danalis, G. Bosilca, and J. Dongarra, Accelerating NWChem Coupled Cluster through dataflow-based Execution,” 11th International Conference on Parallel Processing and Applied Mathematics (PPAM 2015), Krakow, Poland, Springer International Publishing, September 2015.  (452.82 KB)
  6. Haidar, A., P. Luszczek, S. Tomov, and J. Dongarra, Batched Matrix Computations on Hardware Accelerators,” EuroMPI/Asia 2015 Workshop, Bordeaux, France, September 2015.  (589.05 KB)
  7. Baboulin, M., J. Dongarra, A. Remy, S. Tomov, and I. Yamazaki, Dense Symmetric Indefinite Factorization on GPU Accelerated Architectures,” Lecture Notes in Computer Science, vol. 9573: Springer International Publishing, pp. 86-95, September 2015, 2016. DOI: 10.1007/978-3-319-32149-3_9  (327.14 KB)
  8. Haidar, A., S. Tomov, P. Luszczek, and J. Dongarra, MAGMA Embedded: Towards a Dense Linear Algebra Library for Energy Efficient Extreme Computing,” 2015 IEEE High Performance Extreme Computing Conference (HPEC ’15), (Best Paper Award), Waltham, MA, IEEE, September 2015.  (678.86 KB)
  9. Guidry, M., and A. Haidar, On the Design, Autotuning, and Optimization of GPU Kernels for Kinetic Network Simulations Using Fast Explicit Integration and GPU Batched Computation , Oak Ridge, TN, Joint Institute for Computational Sciences Seminar Series, Presentation, September 2015.  (17.25 MB)
  10. Danalis, A., H. Jagode, G. Bosilca, and J. Dongarra, PaRSEC in Practice: Optimizing a Legacy Chemistry Application through Distributed Task-Based Execution,” 2015 IEEE International Conference on Cluster Computing, Chicago, IL, IEEE, September 2015.  (1.77 MB)
  11. Bouteiller, A., G. Bosilca, and J. Dongarra, Plan B: Interruption of Ongoing MPI Operations to Support Failure Recovery,” 22nd European MPI Users' Group Meeting, Bordeaux, France, ACM, September 2015. DOI: 10.1145/2802658.2802668  (543.32 KB)
  12. Baboulin, M., V. Dobrev, J. Dongarra, C. Earl, J. Falcou, A. Haidar, I. Karlin, T. Kolev, I. Masliah, and S. Tomov, Towards a High-Performance Tensor Algebra Package for Accelerators , Gatlinburg, TN, moky Mountains Computational Sciences and Engineering Conference (SMC15), September 2015.  (1.76 MB)

Recent Conferences

  1. AUG
    OpenShmem workshop Annapolis, Delaware
    Aurelien Bouteiller
    Aurelien
    Aurelien Bouteiller
  2. AUG
    EuroPar 2015 Vienna, Austria
    Hartwig Anzt
    Hartwig
    Hartwig Anzt
  3. AUG
    HPCC 2015 Newark, New Jersey
    Jack Dongarra
    Jack
    Piotr Luszczek
    Piotr
    Jack Dongarra, Piotr Luszczek
  4. AUG
    Stanimire Tomov
    Stan
    Stanimire Tomov
  5. SEP
    9th Parallel Tools Workshop Dresden, Germany
    Heike McCraw
    Heike
    Heike McCraw
  6. SEP
    PPAM15 Poland, Krakow
    Heike McCraw
    Heike
    Heike McCraw
  7. SEP
    IEEE Cluster 2015 Chicago, Illinois
    Anthony Danalis
    Anthony
    Anthony Danalis
  8. SEP
    ORNL scientific seminar series Oak Ridge, Tennessee
    Azzam Haidar
    Azzam
    Azzam Haidar
  9. SEP
    Piotr Luszczek
    Piotr
    Piotr Luszczek
  10. SEP
    Euro MPI 2015 Bordeaux, France
    George Bosilca
    George
    George Bosilca
  11. SEP
    Intel Big Data Retreat Hillsboro, Oregon
    Piotr Luszczek
    Piotr
    Thomas Herault
    Thomas
    Piotr Luszczek, Thomas Herault
  12. SEP
    Jack Dongarra
    Jack
    Jack Dongarra

Upcoming Conferences

  1. OCT
    Argo Members Meeting Argonne, Illinois
    Damien Genet
    Damien
    George Bosilca
    George
    Thomas Herault
    Thomas
    Damien Genet, George Bosilca, Thomas Herault
  2. OCT
    IWACOM-III Tokyo, Japan
    Ichitaro Yamazaki
    Ichitaro
    Ichitaro Yamazaki
  3. OCT
    -
    Jack Dongarra
    Jack
    Jack Dongarra
  4. OCT
    -
    White House Workshop on the National Strategic Computing Initiative (NSCI) Washington, District of Columbia
    Jack Dongarra
    Jack
    Jack Dongarra
  5. OCT
    -
    SIAM LA Atlanta, Georgia
    Azzam Haidar
    Azzam
    Hartwig Anzt
    Hartwig
    Ichitaro Yamazaki
    Ichitaro
    Mark Gates
    Mark
    Piotr Luszczek
    Piotr
    Stanimire Tomov
    Stan
    Yaohung
    Mike
    Azzam Haidar, Hartwig Anzt, Ichitaro Yamazaki, Mark Gates, Piotr Luszczek, Stanimire Tomov, Yaohung "Mike" Tsai

Recent Lunch Talks

  1. AUG
    7
    Ian Masliah
    Ian Masliah
    University of Paris-Sud
    Towards C++ and Beyond PDF
  2. AUG
    21
    Yaohung Tsai
    Yaohung Tsai
    Convolutional Layers in RaPyDLI PDF
  3. AUG
    28
    Tingxing Dong
    Tingxing Dong
    Batched Linear Algebra Problems on Hardware Accelerators Based on GPUs PDF
  4. SEP
    4
    Mathieu Faverge
    Mathieu Faverge
    Inria
    Blocking Strategy Optimizations for Sparse Direct Linear Solver on Heterogeneous Architectures PDF
  5. SEP
    11
    Asim YarKhan
    Asim YarKhan
    OpenMP Tasks and PLASMA PDF
  6. SEP
    18
    Mark Gates
    Mark Gates
    Accelerating Collaborative Filtering Using Concepts from High Performance Computing PDF
  7. SEP
    25
    Ichitaro Yamazaki
    Ichitaro Yamazaki
    Random Sampling to Update Truncated SVD PDF

Upcoming Lunch Talks

  1. OCT
    2
    Mike Guidry
    Mike Guidry
    ORNL
    Fast New Methods for Solving Large Sets of Coupled Differential Equations at Scale in Scientific Applications PDF
  2. OCT
    9
    Mike Jantz
    Mike Jantz
    EECS
    Cross-Layer Memory Management to Achieve Power and Performance Goals PDF
  3. OCT
    12
    Edmond Chow
    Edmond Chow
    Georgia Tech
    Very Fine-grained Parallelization of Sparse Linear Algebra Computations PDF
  4. OCT
    16
    Pierre Sens
    Pierre Sens
    LIP6
    Probabilistic Byzantine Tolerance for Cloud Computing PDF
  5. OCT
    23
    Bob Muenchen
    Bob Muenchen
    UTK
    Monitoring Trends in Tools for Data Science PDF
  6. OCT
    30
    Michael Barton
    Michael Barton
    United States Army Research Laboratory (ARL)
    Data Intensive Science and Computing PDF

People

  1. Hartwig Anzt
    Hartwig Anzt recently completed his work as a Post-Doc and has accepted a position at ICL as a Research Scientist I. Congratulations to Hartwig!
  2. Chunyan Tang
    ICL's Chunyan Tang finished her Master's in Computer Science in August. Congratulations and good luck, Chunyan!

congratulations

Raylan Thomas Ellison

ICL’s Teresa Finchum is now a grandmother! Raylan Thomas Ellison was born on August 10, 2015 at 4:20 p.m. to Alex and Anna Ellison. Rylan weighed in at 5 lbs. 15 oz and measured 19 inches long. Congratulations to the Ellison and Finchum families!

teresa_raylan

Farzin Praggo Kabir

Praggo

ICL’s Khairul Kabir is now the proud father of Farzin Praggo Kabir! Praggo was born on August 6, 2015 and weighed in at 7lbs and measured 19.22 inches long. Both mother and baby are doing well. Congratulations!

Dates to Remember

Labor Day

Monday, September 7th is Labor Day, and a holiday for UTK students, faculty, and staff.

SC15 Early Registration

Early registration for SC15 closes on October 15th. Everyone from ICL who plans to attend SC15 should register by this deadline!

SC15 ICL Dinner

Save the date! The SC15 ICL dinner will be on Wednesday, November 18th at 7pm in Austin, TX. Venue to be determined.