News and Announcements

Dongarra Receives Turing Award

Jack Dongarra was formally presented with the ACM A.M. Turing Award at the annual ACM Awards Banquet, which was held this year on Saturday, June 11 at the Palace Hotel in San Francisco.

Jack met up with several ICL alumni on June 11th weekend following the ACM banquet in San Francisco. It was a great opportunity for ICLers on the West Coast to get a chance to congratulate Jack in person and catch up with former colleagues.

Jack Dongarra with west coast ICL alumni

Dongarra Given Cover of Communications of the ACM

Jack Dongarra is featured in the cover article in the latest issue of Communications of the ACM.

In a somewhat refreshing turn from the recent coverage of Dongarra’s looming retirement, the article delves deeper into his history prior to ICL. The author (Neil Savage) touches on Dongarra’s graduate education, internship at Argonne National Laboratory, and some of the challenges facing the HPC community at the time. The article eventually shifts focus toward more current HPC challenges and Dongarra’s dogged determination and exacting standards to meet these challenges.

Jack Visits Technical University of Darmstadt

Jack Dongarra and Felix Wolf

On May 27th, Jack visited the Technical University of Darmstadt, a research university in the city of Darmstadt, Germany as a part of their 50th year anniversary celebrations. While there he gave a talk, “A not so simple matter of software,” which delved into the changes that have occurred in high performance computing and the impact on how algorithms and software libraries are designed for high-end computers.

Dongarra Goes to Sciencetown

Jack Dongarra with KAUST colleagues David Keyes, Hatem Ltaief and Bilel Hadri

Jack Dongarra chatted with KAUST colleagues David Keyes, Hatem Ltaief and Bilel Hadri about supercomputing developments and connections made throughout his career and at KAUST, on a recent episode of the podcast Sciencetown.

It should be noted that Hatem and Bilel are ICL alumni and talk some about their time at ICL. David Keyes refers to KAUST as “ICL east”

Conference Reports

ISC Returns to In-Person

This year’s International Supercomputing Conference (ISC’22) was held on May 29th through June 2nd in Hamburg, Germany. ICL’s Jack Dongarra and Piotr Luszczek were among 3,007 attendees and 137 exhibitors for the first in-person ISC in three years. The conference saw attendance from 59 countries This included close to 200 people who attended remotely.

On Sunday the 29th, Jack and Piotr were joined by incoming Mathworks Professor Hartwig Anzt to give a tutorial on Modern Mixed-Precision Methods wherein they spoke on recent algorithmic progress in exploiting multiple precisions for increased efficiency in performance, communication, and/or storage. Additionally, Jack Dongarra was recognized for receiving the 2021 ACM Turing Award, in a special session. Dongarra was also named as the ISC 2023 deputy chair.

Overall there were many good talks from industry, with keynote speeches coming from: Rev Lebardien, Vice President for Omniverse and Simulation Technology at NVIDIA; Michele Melchiorre, Senior Vice President for Product System, Technical Planning, and Tool Shop at BMW Group; Lorena Barba, Professor of Mechanical and Aerospace Engineering at The George Washington University; and Prof. Thomas Sterling.

The next ISC will be held from May 21 – 25, in Hamburg, under the slogan “Imagine Tomorrow.”

Recent Releases

SLATE release 2022.06.00

SLATE (Software for Linear Algebra Targeting Exascale) is being developed as part of the Exascale Computing Project (ECP), which is a joint project of the U.S. Department of Energy’s Office of Science and National Nuclear Security Administration (NNSA). The objective of SLATE is to provide distributed, GPU-accelerated dense linear algebra capabilities to the US Department of Energy and to the high-performance computing (HPC) community at large.

The new SLATE release:

  • Fixed algorithm selection (issue #41)
  • Fixed set for triangular, trapezoid, symmetric, Hermitian matrices (tzset)
  • Fixed ScaLAPACK pdsgesv wrapper (issue #42)
  • Fixed norm for general band matrix (gbnorm)
  • Added macro for OpenMP default(none); by default empty since it causes unpredictable errors for some compilers or libraries

SLATE can be found here.

Interview

Rob Anderson Then

Rob Anderson

Where are you from?

I am originally from Arkansas, but I moved to Knoxville about 15 or so years ago.

Can you summarize your educational background?

I guess you could say I am a Volunteer. I received a BA in History from UTK, a Master’s in Information Science from UTK and pursued PhD at UTK as well.

Where did you work before joining ICL?

I worked as a Graduate Research Assistant and Graduate Teaching Assistant at CICS at UTK. Before that I spent years in entertainment and various labor jobs.

How did you first hear about the lab, and what made you want to work here?

I first learned about ICL from a job posting. I then started learning as much as I could about the department and lab. I thought I might be a good fit and applied. I have been learning about ICL ever since. It’s fascinating. People use algebra (linear) in real life. Who knew?

What is your focus here at ICL? What are you working on?

I work on IGMCS Program Administration, the Newsletter, NSF and DOE Project Reports, the Publications Database, Technical Editing, and the Friday Talks.

What are your interests/hobbies outside of work?

I enjoy creating and recording music. Mostly, I spend time with my three kids. We have a nine year old daughter, a three year old son and a nine month old baby girl. If I am not working on something ICL related, I am usually elbow deep in gross, dad stuff.

Tell us something about yourself that might surprise people.

I was a somewhat successful stand-up comic and managed to tour for six years before I gave it up.

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

I would prefer to work somewhere I can do technical writing/editing and be a part of a group that is making significant contributions to the field.

Recent Papers

  1. Aliaga, J. I., H. Anzt, T. Grützmacher, E. S. Quintana-Ortí, and A. E. Thomas, Compressed basis GMRES on high-performance graphics processing units,” The International Journal of High Performance Computing Applications, May 2022. DOI: 10.1177/10943420221115140  (13.52 MB)
  2. Cayrols, S., J. Li, G. Bosilca, S. Tomov, A. Ayala, and J. Dongarra, Mixed precision and approximate 3D FFTs: Speed for accuracy trade-off with GPU-aware MPI and run-time data compression,” ICL Technical Report, no. ICL-UT-22-04, May 2022.  (706.14 KB)
  3. Dongarra, J., and A. Geist, Report on the Oak Ridge National Laboratory's Frontier System,” ICL Technical Report, no. ICL-UT-22-05, May 2022.  (16.87 MB)
  4. Abdelfattah, A., S. Tomov, and J. Dongarra, Batch QR Factorization on GPUs: Design, Optimization, and Tuning,” Lecture Notes in Computer Science, vol. 13350, Cham, Springer International Publishing, June 2022. DOI: 10.1007/978-3-031-08751-6_5
  5. Aliaga, J. I., H. Anzt, T. Grützmacher, E. S. Quintana-Orti, and A. E. Thomas, Compression and load balancing for efficient sparse matrix‐vector product on multicore processors and graphics processing units,” Concurrency and Computation: Practice and Experience, vol. 34, issue 14, June 2022. DOI: 10.1002/cpe.6515  (749.82 KB)
  6. Sid-Lakhdar, W. M., S. Cayrols, D. Bielich, A. Abdelfattah, P. Luszczek, M. Gates, S. Tomov, H. Johansen, D. Williams-Young, T. A. Davis, et al., PAQR: Pivoting Avoiding QR factorization,” ICL Technical Report, no. ICL-UT-22-06, June 2022.  (364.85 KB)
  7. Tsai, Y. M., T. Cojean, and H. Anzt, Porting Sparse Linear Algebra to Intel GPUs,” Euro-Par 2021: Parallel Processing Workshops, vol. 13098, Lisbon, Portugal, Springer International Publishing, pp. 57 - 68, June 2022. DOI: 10.1007/978-3-031-06156-1_5

May Conferences

  1. MAY
    -
    Aurelien Bouteiller
    Aurelien
    Earl Carr
    Earl
    George Bosilca
    George
    Aurelien Bouteiller, Earl Carr, George Bosilca
  2. MAY
    -
    Jack Dongarra
    Jack
    Piotr Luszczek
    Piotr
    Jack Dongarra, Piotr Luszczek
  3. MAY
    -
    Alan Ayala
    Alan
    Joseph Schuchart
    Joseph
    Qinglei Cao
    Qinglei
    Alan Ayala, Joseph Schuchart, Qinglei Cao
  4. MAY
    -
    IPDPS 2022 Virtual
    Thomas Herault
    Thomas
    Thomas Herault
  5. JUN
    -
    SC22 Planning Meeting Dallas, Texas
    Anthony Danalis
    Anthony
    Anthony Danalis
  6. JUN
    -
    Deborah Penchoff
    Deborah
    Deborah Penchoff
  7. JUN
    -
    Scalable Tools Workshop Lake Tahoe, California
    Anthony Danalis
    Anthony
    Anthony Danalis
  8. JUN
    -
    Alan Ayala
    Alan
    Alan Ayala
  9. JUN
    -
    Stanimire Tomov
    Stan
    Stanimire Tomov

June Conferences

  1. JUL
    -
    PEARC22 Boston, MA
    Jack Dongarra
    Jack
    Jack Dongarra

Recent Lunch Talks

  1. MAY
    6
    James Demmel
    James Demmel
    Berkeley
    Communication-Avoiding Algorithms for Linear Algebra, ML and Beyond
  2. MAY
    13
    Erin Carson
    Erin Carson
    Charles University, Czech Republic
    Challenges and Opportunities in Mixed Precision Numerical Linear Algebra
  3. MAY
    20
    Swann Perarnau
    Swann Perarnau
    Argonne National Laboratory
    AML: Building Blocks for Advanced Memory Management on Heterogeneous Architectures
  4. MAY
    27
    Vince Weaver
    Vince Weaver
    University of Maine
    Finding Bugs in HPC Systems with the perf_fuzzer

congratulations

Ayala Wins First Place at International HPC Summer School

Alan Ayala of ICL took first place in the 2022 International HPC Summer School competition with an hybrid-GPU code in parallel run on the Bridges supercomputer at Pittsburgh. The problem his code addressed consisted of parallelizing the solution of heat propagation on a 2-D plate, using Volta-100 GPUs. Contestants were given access to: CUDA, OpenACC, OpenMP and MPI.

The school selects young scholars world-wide that are associated with academic institutions from Canada, Europe, Japan and the United States. The event is invitation-only and travel expenses are included. This year’s event was held in Athens, Greece with around 100 participants.

Congratulations, Alan!

From the Twitterverse

@NAGTalk

@Supercomputing

@TheOfficialACM

@SCC_KIT