Andreas Knüpfer
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Andreas Knüpfer has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
Jens Doleschal
Center for Information Services and High Performance Computing (ZIH), Technische Universität Dresden, 01062, Germany
2 shared publications- Michael Wagner
Barcelona Supercomputing Center
2 shared publications - Rigel F. C. Alves
Tech Univ Dresden, Ctr Informat Serv & High Performance Comp ZIH, Dresden, Saxony, Germany
2 shared publications Denis Hünich
TU Dresden
1 shared publication- Denys Makarov
Helmholtz‐Zentrum Dresden‐Rossendorf e.V. Institute of Ion Beam Physics and Materials Research Dresden Germany
1 shared publication - Guannan Mu
Helmholtz‐Zentrum Dresden‐Rossendorf e.V. Institute of Ion Beam Physics and Materials Research Dresden Germany
1 shared publication
Latest Publications
- Open Access
- 1 September 2026
- Published by Wiley in InfoScience
- p. e70023
Multi‐command wearable magnetic human‐machine interface enabled by low‐noise sensing and time‐series deep learning
Wearable magnetic human‐machine interfaces offer unique advantages for unobtrusive and robust interaction, yet their practical deployment is limited by two fundamental challenges: rapid signal attenuation with distance at the hardware level (low signal‐to‐noise ratios) and ambiguous signal interpretation at the algorithmic level. Here, we present a multi‐command magnetic HMI that synergistically integrates low‐noise sensor engineering with time‐series deep learning to overcome these limitations. At the device level, we realized a planar Hall magnetoresistive sensor with an optimized concentric‐ring configuration and a low‐noise readout architecture, achieving a noise floor approaching the thermal limit. The system resolves magnetic‐field variations down to 0.4 μT corresponding to 2 mm displacement at a 12 cm sensor‐magnet separation while extending the interaction range to 20 cm. At the algorithmic level, we developed a temporal learning framework and systematically compared thresholding, a rule‐based classifier, multilayer perceptron, 1D convolutional neural network (1D‐CNN), long short‐term memory (LSTM), and several modern sequence‐learning architectures. The LSTM model achieved an average accuracy of 99.4% for four gesture primitives and maintained 99.3% accuracy after extending the gesture set to six commands by introducing two temporally more complex gestures. Augmentation‐enhanced training further improved performance across acquisition days, previously unseen users, wearing‐position variations, environmental disturbances, and motion artifacts, with accuracies exceeding 90% under all evaluated conditions, demonstrating the system's generalization capability and practical robustness. By combining near‐thermal‐noise magnetic sensing with temporal learning, this work establishes a scalable framework for long‐range, high‐resolution wearable interaction.
- 1 September 2026
- Published by Wiley in InfoScience
- p. e70023
Open AccessMulti‐command wearable magnetic human‐machine interface enabled by low‐noise sensing and time‐series deep learning
Wearable magnetic human‐machine interfaces offer unique advantages for unobtrusive and robust interaction, yet their practical deployment is limited by two fundamental challenges: rapid signal attenuation with distance at the hardware level (low signal‐to‐noise ratios) and ambiguous signal interpretation at the algorithmic level. Here, we present a multi‐command magnetic HMI that synergistically integrates low‐noise sensor engineering with time‐series deep learning to overcome these limitations. At the device level, we realized a planar Hall magnetoresistive sensor with an optimized concentric‐ring configuration and a low‐noise readout architecture, achieving a noise floor approaching the thermal limit. The system resolves magnetic‐field variations down to 0.4 μT corresponding to 2 mm displacement at a 12 cm sensor‐magnet separation while extending the interaction range to 20 cm. At the algorithmic level, we developed a temporal learning framework and systematically compared thresholding, a rule‐based classifier, multilayer perceptron, 1D convolutional neural network (1D‐CNN), long short‐term memory (LSTM), and several modern sequence‐learning architectures. The LSTM model achieved an average accuracy of 99.4% for four gesture primitives and maintained 99.3% accuracy after extending the gesture set to six commands by introducing two temporally more complex gestures. Augmentation‐enhanced training further improved performance across acquisition days, previously unseen users, wearing‐position variations, environmental disturbances, and motion artifacts, with accuracies exceeding 90% under all evaluated conditions, demonstrating the system's generalization capability and practical robustness. By combining near‐thermal‐noise magnetic sensing with temporal learning, this work establishes a scalable framework for long‐range, high‐resolution wearable interaction.
- Open Access
- 25 October 2021
- Published by PeerJ in PeerJ Computer Science
- Vol. 7, e753
Further enhancing the in situ visualization of performance data in parallel CFD applications
This paper continues the work initiated by the authors on the feasibility of using ParaView as visualization software for the analysis of parallel Computational Fluid Dynamics (CFD) codes’ performance. Current performance tools have limited capacity of displaying their data on top of three-dimensional, framed (i.e., time-stepped) representations of the cluster’s topology. In our first paper, a plugin for the open-source performance tool Score-P was introduced, which intercepts an arbitrary number of manually selected code regions (mostly functions) and send their respective measurements–amount of executions and cumulative time spent–to ParaView (through its in situ library, Catalyst), as if they were any other flow-related variable. Our second paper added to such plugin the capacity to (also) map communication data (messages exchanged between MPI ranks) to the simulation’s geometry. So far the tool was limited to codes which already have the in situ adapter; but in this paper, we will take the performance data and display it–also in codes without in situ–on a three-dimensional representation of the hardware resources being used by the simulation. Testing is done with the Multi-Grid and Block Tri-diagonal NPBs, as well as Rolls-Royce’s CFD code, Hydra. The benefits and overhead of the plugin's new functionalities are discussed.
- 25 October 2021
- Published by PeerJ in PeerJ Computer Science
- Vol. 7, e753
Open AccessFurther enhancing the in situ visualization of performance data in parallel CFD applications
This paper continues the work initiated by the authors on the feasibility of using ParaView as visualization software for the analysis of parallel Computational Fluid Dynamics (CFD) codes’ performance. Current performance tools have limited capacity of displaying their data on top of three-dimensional, framed (i.e., time-stepped) representations of the cluster’s topology. In our first paper, a plugin for the open-source performance tool Score-P was introduced, which intercepts an arbitrary number of manually selected code regions (mostly functions) and send their respective measurements–amount of executions and cumulative time spent–to ParaView (through its in situ library, Catalyst), as if they were any other flow-related variable. Our second paper added to such plugin the capacity to (also) map communication data (messages exchanged between MPI ranks) to the simulation’s geometry. So far the tool was limited to codes which already have the in situ adapter; but in this paper, we will take the performance data and display it–also in codes without in situ–on a three-dimensional representation of the hardware resources being used by the simulation. Testing is done with the Multi-Grid and Block Tri-diagonal NPBs, as well as Rolls-Royce’s CFD code, Hydra. The benefits and overhead of the plugin's new functionalities are discussed.
- 1 September 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 1, 967-974
From stirring to mixing: artificial intelligence in the process industry
The introduction of AI methods in production or production-related environments meets with resistance from operators due to their lack of relevant experience and their responsibility for plant safety. To overcome these inhibitions one requires prototypical implementations, which offer considerable benefits, meet the highest requirements for reliability, are accepted by the operating personnel, and get support from those in charge. As a result, AI technologies must be embedded into the complex IT/OT infrastructure of the companies. Traceability, maintainability and longevity must also be guaranteed. As a first step towards this, we present a concept of the demonstrator and its first results, which should make AI comprehensible by visualizing challenges and exploring possibilities in the process industry.
- 1 September 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 1, 967-974
From stirring to mixing: artificial intelligence in the process industry
The introduction of AI methods in production or production-related environments meets with resistance from operators due to their lack of relevant experience and their responsibility for plant safety. To overcome these inhibitions one requires prototypical implementations, which offer considerable benefits, meet the highest requirements for reliability, are accepted by the operating personnel, and get support from those in charge. As a result, AI technologies must be embedded into the complex IT/OT infrastructure of the companies. Traceability, maintainability and longevity must also be guaranteed. As a first step towards this, we present a concept of the demonstrator and its first results, which should make AI comprehensible by visualizing challenges and exploring possibilities in the process industry.
- Open Access
- 31 July 2020
- Published by Springer Nature
- p. 103-142
DASH: Distributed Data Structures and Parallel Algorithms in a Global Address Space
- 31 July 2020
- Published by Springer Nature
- p. 103-142
Open AccessDASH: Distributed Data Structures and Parallel Algorithms in a Global Address Space
- Open Access
- 29 May 2020
- Published by Springer Nature
- p. 400-412
In Situ Visualization of Performance-Related Data in Parallel CFD Applications
This paper aims at investigating the feasibility of using ParaView as visualization software for the analysis and optimization of parallel CFD codes’ performance. The currently available software tools for reading profiling data do not match the generated measurements to the simulation’s original mesh and somehow aggregate them (rather than showing them on a time-step basis). A plugin for the open-source performance tool Score-P has been developed, which intercept an arbitrary number of manually selected code regions (mostly functions) and send their respective measurements – amount of executions and cumulative time spent – to ParaView (through its in situ library, Catalyst), as if they were any other flow-related variable. Results show that (i) the impact of mesh partition algorithms on code performance and (ii) the load imbalances (and their eventual relationship to mesh size/simulation physics) become easier to investigate.
- 29 May 2020
- Published by Springer Nature
- p. 400-412
Open AccessIn Situ Visualization of Performance-Related Data in Parallel CFD Applications
This paper aims at investigating the feasibility of using ParaView as visualization software for the analysis and optimization of parallel CFD codes’ performance. The currently available software tools for reading profiling data do not match the generated measurements to the simulation’s original mesh and somehow aggregate them (rather than showing them on a time-step basis). A plugin for the open-source performance tool Score-P has been developed, which intercept an arbitrary number of manually selected code regions (mostly functions) and send their respective measurements – amount of executions and cumulative time spent – to ParaView (through its in situ library, Catalyst), as if they were any other flow-related variable. Results show that (i) the impact of mesh partition algorithms on code performance and (ii) the load imbalances (and their eventual relationship to mesh size/simulation physics) become easier to investigate.
- 1 July 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 129-136
Tracing long running applications: A case study using Gromacs
Performance analysis is inevitable to develop applications that utilize the enormous capabilities of current HPC systems. While many recent tool studies focused on large scales, performance analysis of long-running applications has not been paid much attention. This paper investigates challenges that arise from monitoring long-running real-life applications, in particular, the disruptive bias of intermediate memory buffer flushes in the measurement environment. We propose a concept for an in-memory event tracing that completely avoids intermediate memory buffer flushes. We evaluate to which extent such an in-memory event tracing workflow helps overcoming the critical properties, such as resulting trace size, application slow down, and measurement bias. We utilize a prototype implementation, based on Score-P and OTF2, with the molecular dynamics packages Gromacs, an application currently infeasible to monitor in a full production run.
- 1 July 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 129-136
Tracing long running applications: A case study using Gromacs
Performance analysis is inevitable to develop applications that utilize the enormous capabilities of current HPC systems. While many recent tool studies focused on large scales, performance analysis of long-running applications has not been paid much attention. This paper investigates challenges that arise from monitoring long-running real-life applications, in particular, the disruptive bias of intermediate memory buffer flushes in the measurement environment. We propose a concept for an in-memory event tracing that completely avoids intermediate memory buffer flushes. We evaluate to which extent such an in-memory event tracing workflow helps overcoming the critical properties, such as resulting trace size, application slow down, and measurement bias. We utilize a prototype implementation, based on Score-P and OTF2, with the molecular dynamics packages Gromacs, an application currently infeasible to monitor in a full production run.
- Open Access
- 15 September 2013
- Published by Association for Computing Machinery (ACM)
- p. 123-128
Runtime message uniquification for accurate communication analysis on incomplete MPI event traces
Communication analysis of parallel applications based on event traces depends on correct matching of associated MPI send and receive events. Selective monitoring techniques, however, may result in incomplete MPI event traces and, in that case, current matching strategies fail. In this paper we introduce an additional unique identifier for each message to make MPI events distinguishable from others. Therefore, it is possible to identify missing MPI events and match all remaining MPI events correctly. An overhead study with a real-life application and a benchmark suite demonstrates the applicability and benefits of this approach.
- 15 September 2013
- Published by Association for Computing Machinery (ACM)
- p. 123-128
Open AccessRuntime message uniquification for accurate communication analysis on incomplete MPI event traces
Communication analysis of parallel applications based on event traces depends on correct matching of associated MPI send and receive events. Selective monitoring techniques, however, may result in incomplete MPI event traces and, in that case, current matching strategies fail. In this paper we introduce an additional unique identifier for each message to make MPI events distinguishable from others. Therefore, it is possible to identify missing MPI events and match all remaining MPI events correctly. An overhead study with a real-life application and a benchmark suite demonstrates the applicability and benefits of this approach.