I studied Physics with a major in Computer Science at Heidelberg University. For my diploma thesis I researched the potential of a port of the ALICE HLT TPC Tracker to many-core systems with wide vector units. In the course of the work I developed an abstraction for C++ for SSE and LRBni programming.
I continued research and development on the abstraction for data-parallel programming with the C++ programming language, in particular for efficient use of SIMD instructions on x86. This work was done at Frankfurt University and FIAS in collaboration with high-energy physics experiments at GSI, CERN, and RHIC.
- Vc: Research and development of a C++ API for different SIMD hardware with an emphasis on developer productivity and high performance.
- Standardization of a SIMD interface in the C++ programming language.
- High Performance Linpack for GPU clusters
- Vectorization of the ALICE TPC Tracker (and its derivatives)
- KDE4 framework
- Phonon (KDE Multimedia Abstraction)
- many contributions to KDE 2.x and KDE 3.x
- NUMA Bench
Matthias Bach et al. “Optimized HPL for AMD GPU and multi-core CPU usage.” In: Computer Science - Research and Development 26 (2011), pp. 1–12. issn: 1865-2034. doi: 10.1007/s00450-011-0161-5.
M. Bach et al. “A Comprehensive Approach for a Power Efficient General Purpose Supercomputer.” In: Parallel, Distributed and Network-Based Processing (PDP), 2013 21st Euromicro International Conference on. Feb. 2013, pp. 336–342. doi: 10.1109/PDP.2013.55.
S. Gorbunov et al. “ALICE HLT High Speed Tracking and Vertexing.” In: Real Time Conference (RT), 2010 17th IEEE-NPSS. May 2010, pp. 1–4. doi: 10.1109/RTC.2010.5750344.
S. Gorbunov et al. “ALICE HLT High Speed Tracking on GPU.” In: Nuclear Science, IEEE Transactions on 58.4 (Aug. 2011), pp. 1845–1851. issn: 0018-9499. doi: 10.1109/TNS.2011.2157702.
Matthias Kretz and Volker Lindenstruth. “Vc: A C++ library for explicit vectorization.” In: Software: Practice and Experience (2011). issn: 1097-024X. doi: 10.1002/spe.1149.
[private archive, before peer review]
M. Richter et al. “Event Reconstruction Performance of the ALICE High Level Trigger for Collisions.” In: Nuclear Science, IEEE Transactions on 58.4 (Aug. 2011), pp. 1706–1713. issn: 0018-9499. doi: 10.1109/TNS.2011.2160093.
M. Richter et al. “Event reconstruction performance of the ALICE High Level Trigger for p+p collisions.” In: Real Time Conference (RT), 2010 17th IEEE-NPSS. May 2010, pp. 1–6. doi: 10.1109/RTC.2010.5750346.
David Rohr et al. “ALICE HLT TPC Tracking of Pb-Pb Events on GPUs.” In: Journal of Physics: Conference Series 396.1 (2012), p. 012044. doi: 10.1088/1742-6596/396/1/012044.
David Rohr et al. “Multi-GPU DGEMM and High Performance Linpack on Highly Energy-Efficient Clusters.” In: IEEE Micro 31.5 (2011), pp. 18–27. doi: 10.1109/MM.2011.66.
ISO C++ Committee (WG21) Papers
- N3759: “SIMD Vector Types.” 2013.
- N4184: “SIMD Types: The Vector Type & Operations.” 2014.
- N4185: “SIMD Types: The Mask Type & Write-Masking.” 2014.
- N4186: “Supporting Custom Diagnostics and SFINAE.” 2014.
- N4395: “SIMD Types: ABI Considerations.” 2015.
- N4454: “SIMD Types Example: Matrix Multiplication.” 2015.
- P0214R0: ”Data-Parallel Vector Types & Operations.” 2016.