Simon Pfenning
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Simon Pfenning has a total of 4 co-authors. Here are the top co-authors by number of shared publications:
- Alexander Lehnert
Institute of Applied Microelectronics and Computation Engineering, Faculty of Computer Science and Electrical Engineering, University of Rostock, 18051 Rostock, Germany
1 shared publication - Marc Reichenbach
University of Rostock
1 shared publication - Philipp Holzinger
Chair of Computer Architecture, Friedrich-Alexander University
1 shared publication - Ralf Reiner Muller
Institute for Digital Communications Friedrich-Alexander Universität Erlangen Nürnberg Germany
1 shared publication
Latest Publications
- Open Access
- 5 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 3881-3898
Most Resource Efficient Matrix Vector Multiplication on FPGAs
Fast and resource-efficient inference in artificial neural networks (ANNs) is of utmost importance and drives many new developments in the area of new hardware architectures, e.g., by means of systolic arrays or algorithmic optimization such as pruning. In this paper, we present a novel method for lowering the computation effort for ANN inference utilizing ideas from information theory.Weight matrices are sliced into submatrices of logarithmic aspect ratios. These slices are then factorized. This reduces the number of required computations without compromising on fully parallel processing. We create a new hardware architecture for this dedicated purpose. We also provide a tool to map these sliced and factorized matrices efficiently to reconfigurable hardware. By comparing to the state of the art FPGA implementations, we can prove our claim by lowering hardware resources measured in look-up-tables (LUTs) by a factor of three to six. Our method does not rely on any particular property of the weight matrices of the ANN. It works for the general task of multiplying an input vector with a constant matrix and is also suitable for digital signal processing beyond ANNs.
- 5 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 3881-3898
Open AccessMost Resource Efficient Matrix Vector Multiplication on FPGAs
Fast and resource-efficient inference in artificial neural networks (ANNs) is of utmost importance and drives many new developments in the area of new hardware architectures, e.g., by means of systolic arrays or algorithmic optimization such as pruning. In this paper, we present a novel method for lowering the computation effort for ANN inference utilizing ideas from information theory.Weight matrices are sliced into submatrices of logarithmic aspect ratios. These slices are then factorized. This reduces the number of required computations without compromising on fully parallel processing. We create a new hardware architecture for this dedicated purpose. We also provide a tool to map these sliced and factorized matrices efficiently to reconfigurable hardware. By comparing to the state of the art FPGA implementations, we can prove our claim by lowering hardware resources measured in look-up-tables (LUTs) by a factor of three to six. Our method does not rely on any particular property of the weight matrices of the ANN. It works for the general task of multiplying an input vector with a constant matrix and is also suitable for digital signal processing beyond ANNs.