Skip to content. | Skip to navigation


Sections
Document Actions

Roman Iakymchuk

Contact

Dr. Roman Iakymchuk
LIP6 -- UPMC Univ Paris 06
Boîte courrier 169
Couloir 26-00, Étage 3, Bureau 326
4 place Jussieu
75252 Paris Cedex 05

Phone: 0033(0) 144 27 88 76
Emails: roman.iakymchuk (AT) lip6 (DOT) fr
           iakymchuk (AT) aices (DOT) rwth-aachen (DOT) de

Education

02/2009 - 08/2012
Dr.rer.nat. (equiv. Ph.D.) in Computer Science, Aachen Institute for Advanced Study in Computations Engineering Science (AICES), RWTH Aachen University
09/2007 - 06/2008
Master of Science (Hons) in Applied Mathematics and Informatics, Ivan Franko National University of Lviv, Ukraine
09/2003 - 06/2007
Bachelor of Science (Hons) in Applied Mathematics and Informatics, Ivan Franko National University of Lviv, Ukraine

Professional Career

10/2013 - present
Postdoctoral researcher at Labex CalsimLab (ICS and LIP6), Sorbonne Universités, UPMC Univ Paris 06, Paris, France
10/2012 - 09/2013
Postdoctoral researcher at Institut Mines-Télécom - Télécom SudParis, Evry, France
10/2008 - 01/2009
Assistant Researcher at Ivan Franko National University of Lviv, Ukraine
07/2007 - 01/2009
Software engineer at SoftServe Inc., Lviv, Ukraine
09/2005 - 06/2007
Engineer of Computer Equipment Laboratory at Ivan Franko National University of Lviv, Ukraine

Research Interests

High-performance computing, Numerical linear algebra, Computer arithmetic, Performance analysis and modeling.


Journal Articles

  • Modeling Performance through Memory-Stalls. (pdf)
    Roman Iakymchuk and Paolo Bientinesi. ACM SIGMETRICS Performance Evaluation Review. 2012, Vol. 40, 86-91.
  • On a Secant Type Method for the Nonlinear Least Squares Problem. (pdf)
    Stepan Shakhno, Oleksandra Gnatyshyn, and Roman Iakymchuk. Journal of Numerical and Applied Mathematics. 2009, Vol. 97, 112-121.
  • About Two-Step Modification of Gauss-Newton Method at the Generalized Lipschitz Conditions for Solving the Nonlinear Least Squares Problem. (link) (in Ukrainian)
    Stepan Shakhno and Roman Iakymchuk. Math. Visnyk of the Shevchenko Scientific Society. 2009, Vol. 6, 277-287.
  • One Difference Method with Super-Quadratic Convergence for Solving the Nonlinear Least Squares Problem. (pdf) (in Ukrainian)
    Stepan Shakhno, Oleksandra Gnatyshyn, and Roman Iakymchuk. Visnyk of the Lviv Univ. Ser. Applied Math. and Inform. 2007, Vol. 13, 51-59.

Submitted Journal Articles

  • Numerical Reproducibility for the Parallel Reduction on Multi- and Many-Core Architectures. (pdf)
    Sylvain Collange, David Defour, Stef Graillat, and Roman Iakymchuk. Parallel Computing Journal
  • Reproducible and Accurate Matrix Multiplication for GPU Accelerators. (pdf)
    Roman Iakymchuk, Sylvain Collange, Devid Defour, and Stef Graillat. Lecture Notes of Computer Science

Book Chapter

  • HPC on Competitive Cloud Resources. (pdf)
    Paolo Bientinesi, Roman Iakymchuk, and Jeff Napper.
    In Borko Furht and Armando Escalante, editors, Handbook of Cloud Computing. Springer, 2010, 493-516.

Peer Reviewed Conference Publications

  • Reproducible Triangular Solvers for High-Performance Computing. (pdf)
    Roman Iakymchuk, Sylvain Collange, Devid Defour, and Stef Graillat. Special track on: Wavelets and Validated Numerics held as part of the 12th International Conference on Information Technology: New Generations (ITNG 2015), April 13-15, 2015, Las Vegas, Nevada, USA. Accepted.
  • On the Convergence Analysis of a Two-Step Modification of the Gauss-Newton Method. (link)
    Roman Iakymchuk and Stepan Shakhno. PAMM. Vol. 14. Special issue: 85th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM), Erlangen, Germany, 2014, 813-814.
  • Performance Analysis on Energy Efficient High-Performance Architectures. (pdf)
    Roman Iakymchuk and François Trahay. In Proceedings of the 2nd International Conference on Cluster Computing (CC'13). Lviv, Ukraine, June 3-5, 2013, 80-85.
  • Execution-Less Performance Modeling. (pdf)
    Roman Iakymchuk and Paolo Bientinesi. In the Proceedings of the second international workshop on Performance Modeling, Benchmarking and Simulation of High-Performance Computing Systems (PMBS11) held as part of the Supercomputing Conference (SC11). Seattle, USA, November 13th, 2011, 11-12.
  • Performance Prediction through Time Measurements. (pdf)
    Roman Iakymchuk. In Proceedings of the 1st International Conference on High-Performance Computing (HPC-UA'11). Kyiv, Ukraine, October 12-14, 2011, 26-32.
  • Improving High-Performance Computations on Clouds Through Resource Underutilization. (pdf)
    Roman Iakymchuk, Jeff Napper, and Paolo Bientinesi. In Proceedings of ACM 26th Symposium on Applied Computing, special track on Cloud Computing. Taichung, Taiwan, March 21-25, 2011, 119-126.
  • On a One-Step Modification of Gauss-Newton Method under Generalized Lipschitz Conditions for Solving the Nonlinear Least Squares Problem. (link)
    Stepan Shakhno and Roman Iakymchuk. PAMM. Vol. 9. Special issue: 80th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM), Gdansk, Poland, 2009, 565-566.

News Articles

  • Crowded Clouds. (link)
    Jeff Napper, Roman Iakymchuk, and Paolo Bientinesi. HPCwire, August 12th, 2010.
  • HPC Sharing in the Cloud. (link)
    Jeff Napper, Roman Iakymchuk, and Paolo Bientinesi. HPC in the Cloud, August 12th, 2010.

Poster

  • ExBLAS: Reproducible and Accurate BLAS Library. (link)
    Roman Iakymchuk, Stef Graillat, Sylvain Collange, and Devid Defour. 7ème Rencontre Arithmétique de l'Informatique Mathématique (RAIM 2015). Rennes, France, April 7-9, 2015.

Short Papers and Abstracts

  • Methods with Successive and Parallel Approximations of Inverse Operator for the Nonlinear Least Squares Problem
    Roman Iakymchuk and Stepan Shakhno. In Book of Abstracts of the 86th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM). Lecce, Italy, March 23-27, 2015.
  • Reproducible and Accurate Matrix Multiplication for High-Performance Computing.
    Sylvain Collange, Devid Defour, Stef Graillat, and Roman Iakymchuk. In Book of Abstracts of the 16th GAMM-IMACS International Symposium on Scientific Computing, Computer Arithmetic and Validated Numerics (SCAN'14). Würzburg, Germany, September 21-26, 2014.
  • A Reproducible Accurate Summation Algorithm for High-Performance Computing.
    Sylvain Collange, Devid Defour, Stef Graillat, and Roman Iakymchuk. In Book of Abstracts of the SIAM Workshop on Exascale Applied Mathematics Challenges and Opportunities (EX14) held as part of the 2014 SIAM Annual Meeting. Chicago, Il, USA, July 6--11, 2014.
  • On the Convergence Analysis of a Two-Step Modification of the Gauss-Newton Method.
    Roman Iakymchuk and Stepan Shakhno. In Book of Abstracts of the 85th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM). Erlangen, Germany, March 10--14, 2014.
  • On a One-Step Modification of the Gauss-Newton Method under Generalized Lipschitz Conditions for Solving the Nonlinear Least Squares Problem.
    Stepan Shakhno and Roman Iakymchuk. In Book of Abstracts of the 85th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM). Gdansk, Poland, February 9--13, 2009.
  • About One Modification of Secant Method for Solving the Nonlinear Least Squares Problem
    Oleksandra Gnatyshyn, Stepan Shakhno, and Roman Iakymchuk. In Book of Abstracts of the 15th National Scientific Conference "Actual Problems in Applied Mathematics and Informatics". Lviv, Ukraine, September 23--25, 2008, 18. (in Ukrainian)
  • About an Iterative Difference Modification of the Gauss-Newton Method for Solving the Nonlinear Least Squares Problem
    Roman Iakymchuk. In Book of Abstracts of the 11th National Students' Scientific Conference in Applied Mathematics and Informatics SSCAMI-2008. Lviv, Ukraine, April 9--10, 2008, 277--278. (in Ukrainian)

Ph.D. Thesis

  • Performance Modeling and Prediction for Linear Algebra Algorithms. (link)
    Advisor: Prof. Paolo Bientinesi, Ph.D. Co-advisor: Prof. Dr. Enrique S. Quitana Ortí.
    AICES, RWTH Aachen University, August, 2012.

Preprints and Technical Reports

  • Reproducible Triangular Solvers for High-Performance Computing. (pdf)
    Roman Iakymchuk, Devid Defour, Sylvain Collange, and Stef Graillat. HAL ID: hal-01116588, version 2. February 2015.
  • Numerical Reproducibility for the Parallel Reduction on Multi- and Many-Core Architectures. (pdf)
    Sylvain Collange, Devid Defour, Stef Graillat, and Roman Iakymchuk. HAL ID: hal-00949355, version 3. February 2015.
  • Reproducible and Accurate Matrix Multiplication for GPU Accelerators. (pdf)
    Roman Iakymchuk, Devid Defour, Sylvain Collange, and Stef Graillat. HAL ID: hal-01102877. January 2015.
  • On the Convergence Analysis of a Two-Step Modification of the Gauss-Newton Method. (pdf)
    Roman Iakymchuk and Stepan Shakhno. HAL ID: hal-00960300, version 2. March 2014.
  • Full-Speed Deterministic Bit-Accurate Parallel Floating-Point Summation on Multi- and Many-Core Architectures. (pdf)
    Sylvain Collange, Devid Defour, Stef Graillat, and Roman Iakymchuk. HAL ID: hal-00949355. February 2014.
  • LiTL: Lightweight Trace Library. (pdf)
    Roman Iakymchuk and François Trahay. HAL ID: hal-00918733. July 2013.
  • Modeling Performance through Memory-Stalls. (preprint, bib)
    Roman Iakymchuk and Paolo Bientinesi. AICES-2011/07-1. July 2011.
  • Underutilizing Resources for HPC on Clouds. (preprint, bib)
    Roman Iakymchuk, Jeff Napper, and Paolo Bientinesi. AICES-2010/06-1. June 2010.
  • HPC on Competitive Cloud Resources. (preprint, bib)
    Paolo Bientinesi, Roman Iakymchuk, and Jeff Napper. AICES-2010/04-2. April 2010.

Talks

  • Reproducibility and Accuracy for High-Performance Computing(pdf)
    7ème Rencontre Arithmétique de l'Informatique Mathématique (RAIM 2015). Rennes, France, April 7-9, 2015.
  • Methods with Successive and Parallel Approximations of Inverse Operator for the Nonlinear Least Squares Problem
    The 86th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM). Lecce, Italy, March 23-27, 2015.
  • Reproducible and Accurate BLAS routines towards ExaScale Computing.
    The Pequan seminar, Sorbonne Universités, UPMC Univ Paris 06, January 15th, 2014.
  • Numerical Reproducibility of BLAS routines towards ExaScale Computing.
    The Aric seminar, LIP, ÉNS Lyon, December 11th, 2014.
  • Reproducible and Accurate Algorithms for ExaScale Computing.
    ICS Matinée Jeunes Chercheurs, Sorbonne Universités, UPMC Univ Paris 06, November 5th, 2014
  • Reproducible and Accurate Matrix Multiplication for High-Performance Computing.
    The 16th GAMM-IMACS International Symposium on Scientific Computing, Computer Arithmetic and Validated Numerics (SCAN'14). Würzburg, Germany, September 21-26, 2014.
  • A Reproducible Accurate Summation Algorithm for High-Performance Computing.
    The SIAM Workshop on Exascale Applied Mathematics Challenges and Opportunities (EX14) held as part of the 2014 SIAM Annual Meeting. Chicago, Il, USA, July 6--11, 2014.
  • On the Convergence Analysis of a Two-Step Modification of the Gauss-Newton Method.
    The 85th Annual Meeting of the International Association of Applied Mathematics and Mechanics (GAMM). Erlangen, Germany, March 10-14, 2014.
  • Performance Analysis on Energy Efficient High-Performance Architectures. (pdf)
    The 2nd International Conference on Cluster Computing (CC'13). Lviv, Ukraine, June 3-5, 2013.
  • Performance Modeling and Prediction for Linear Algebra Algorithms. (pdf)
    Doctoral defence. AICES, RWTH Aachen University, August 30th, 2012.
  • Execution-Less Performance Modeling. (pdf)
    The 2nd International Workshop on Performance Modeling, Benchmarking and Simulation of High-Performance Computing Systems (PMBS11) held as part of the Supercomputing Conference (SC11). Seattle, USA, November 13th, 2011.
  • Performance Prediction through Time Measurements. (pdf)
    The 1st International Conference on High-Performance Computing (HPC-UA 2011). Kyiv, Ukraine, October 12-14, 2011.
  • Modeling Performance through Memory-Stalls. (pdf)
    Conference for Young Scientists "Facing the Multicore-Challenge II". Karlsruhe, Germany, September 28-30, 2011.