Daniel Thuerck Research Scientist
Numerical Linear Algebra - Sparse Algorithms - HPC - GPGPU

About me


I am a research scientist at NEC Laboratories Europe. Before that, I was a PhD student at TU Darmstadt, advised by Prof. Michael Goesele in the GCC group, then by by Prof. Kristian Kersting from the AIML group. and Prof. Marc Pfetsch from the Discrete Optimization Group. Furthermore, I was an Associate of the Graduate School of Computational Engineering at TU Darmstadt.

My research interests are numerical linear algebra, linear and integer optimization and GPGPU algorithms. Apart from the occasional detour into computer vision and computer graphics, I focus on finding new ways to elevate the usage of different HPC accelerator architectures in numerical linear algebra algorithms. I often find myself working on the system level, integrating the different layers of a system research project for best possible performance.

I see myself as working close to the boundary of research and engineering. Rather than hopping from project to project and publication to publication, I strive to package the outcomes of research projects as publishable and usable artifacts, including production-orientied code and data.

If you're interested in cooperating or just have questions about my research or my projects, don't hesitate to contact me!

Please use my PGP key for secure E-Mail communication.

Curriculum Vitae


ongoing Research Scientist
02/2020 NEC Laboratories Europe, Heidelberg
01/2020 Research Associate / PhD student
08/2015 Graphics, Capture and Massively Parallel Computing (2015 - 2018); Artificial Intelligence and Machine Learning Lab (2018 - 2020), TU Darmstadt
03/2019 Research Intern
09/2018 Facebook Reality Labs, Pittsburgh, under Shoou-I Yu
07/2016 Research Intern
04/2016 NVIDIA Research, PSA group under Michael Garland
03/2016 Research Intern
09/2015 NVIDIA Research, MVC group under Jan Kautz
08/2015 Fast-Track PhD student
10/2014 Graduate School of CE, TU Darmstadt

Publications


Conference/Workshop Papers

  • Flynn's reconciliation: Automating the register cache idiom for cross-accelerator programming

    Daniel Thuerck, Nicolas Weber and Roberto Bifulco

    ACM Transactions on Architecture and Code Optimization 18(3), 2021

  • Algorithm 1015: A Fast Scalable Solver for the Dense Linear (Sum) Assignment Problem

    Stefan Guthe and Daniel Thuerck

    ACM Transactions on Mathematical Software 47(2), 2021

  • Supporting Irregularity in Throughput-Oriented Computing by SIMT-SIMD Integration

    Daniel Thuerck

    IEEE/ACM 10th Workshop on Irregular Applications: Architectures and Algorithms (IA3) Short Paper, Virtual Event, 2020

  • Stretching Jacobi: A Two-Stage Pivoting Approach for Block-Based Factorization

    Daniel Thuerck

    IEEE/ACM 9th Workshop on Irregular Applications: Architectures and Algorithms (IA3), Denver, CO, USA, 2019

  • A block-oriented, parallel and collective approach to sparse indefinite preconditioning on GPUs

    Daniel Thuerck, Maxim Naumov, Michael Garland and Michael Goesele

    IEEE/ACM 8th Workshop on Irregular Applications: Architectures and Algorithms (IA3), Dallas, TX, USA, 2018

    Best Paper Award

  • A Fast, Massively Parallel Solver for Large, Irregular Pairwise Markov Random Fields

    Daniel Thuerck, Michael Waechter, Sven Widmer, Max von Buelow, Patrick Seemann, Marc E. Pfetsch and Michael Goesele

    High Performance Graphics (HPG) 2016, Dublin, Ireland, 2016

  • Using graphics processing units to investigate molecular coevolution

    Michael Waechter, Kathrin Jaeger, Daniel Thuerck, Stephanie Weissgraeber, Sven Widmer, Michael Goesele and Kay Hamacher

    Concurrency and Computation: Practice and Experience 26(8), 2014

  • Lazy nonlinear diffusion parameter estimation

    Daniel Thuerck and Arjan Kuijper

    International Conference on Image Analysis and Processing (ICIAP) 2013, Naples, Italy, 2013

  • Efficient heuristic adaptive quadrature on gpus: Design and evaluation

    Daniel Thuerck, Sven Widmer, Arjan Kuijper and Michael Goesele

    International Conference on Parallel Processing and Applied Mathematics (PPAM) 2013, Warsaw, Poland, 2013

  • Nonlinear diffusion at your fingertips: Theory and mobile applications

    Daniel Thuerck and Arjan Kuijper

    2013 8th International Symposium on Image and Signal Processing and Analysis (ISPA), Trieste, Italy, 2013

  • Cosine-driven non-linear denoising

    Daniel Thuerck and Arjan Kuijper

    International Conference Image Analysis and Recognition (ICIAR) 2013, Povoá de Varzim, Portugal, 2013

Talks

  • Finding Parallelism in General-Purpose Linear Programming

    GTC 2017, San José, CA, USA, 2017

Thesis

  • Irregularity Mitigation and Portability Abstractions for Accelerated Sparse Matrix Factorization

    PhD Thesis, TU Darmstadt, 2021

  • Optimizing large-scale irregular Markov Random Fields on GPUs

    Master's Thesis, TU Darmstadt, 2014

  • A Well Posed Perona Malik Model and its Implementation Without the Need of Prior Knowledge

    Bachelor's Thesis, TU Darmstadt, 2012

Technical Reports

  • Lock-Free Parallel Feedback Vertex Set Selection

    Daniel Thuerck and Michael Goesele

    TU Darmstadt, 2018

Code