Prototyping Platform for the Development of Proof-of-Concept of Digital Signal Processing Algorithms

Currently, in the field of new technologies, the trend is to use tools that minimize non-recurring engineering (NRE)costs and time-to-market. This is based on the fact that the high competitiveness between technological companies requires high dynamism in the flow of innovative ideas up to become salable products. In this aspect, the use of Heterogeneous Computing Platforms (HCPs)has a significant upswing since they allow rapid prototyping and a considerable approximation to the final product. Proof of concept (PoC)is an implementation of an idea or principle to demonstrate its feasibility or practical potential. Generally, the PoCs are small and may not be complete. This article exposes the following contribution: a proposal of an HCP that allows to analyze PoC by evaluating the performance of implementations with different objectives, such as theoretical evaluation of an algorithm, efficient implementation in terms of speed or finding processing bottlenecks to identify hotspots in the processing chain and apply manual refinements to the final product.

[1]  Jan J. Gerbrands,et al.  On the relationships between SVD, KLT and PCA , 1981, Pattern Recognit..

[2]  Prongnuch Sethakarn,et al.  Heterogeneous Computing Platform for data processing , 2016 .

[3]  Scott F. Midkiff,et al.  Design and implementation of a configurable platform for embedded communication systems , 2003, Proceedings International Parallel and Distributed Processing Symposium.

[4]  G. Strang Introduction to Linear Algebra , 1993 .

[5]  Roberto Guerrieri,et al.  Multicore Signal Processing Platform With Heterogeneous Configurable Hardware Accelerators , 2014, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[6]  Deni Torres Román,et al.  A parallel implementation of One-Sided Jacobi SVD for non-symmetric squared matrices on a high-performance GPU , 2016, 2016 International Conference on Electronics, Communications and Computers (CONIELECOMP).

[7]  Tao Wang,et al.  Implementing the Jacobi Algorithm for Solving Eigenvalues of Symmetric Matrices with CUDA , 2012, 2012 IEEE Seventh International Conference on Networking, Architecture, and Storage.

[8]  Livia Marcellino,et al.  A GPU-Accelerated SVD Algorithm, Based on QR Factorization and Givens Rotations, for DWI Denoising , 2016, 2016 12th International Conference on Signal-Image Technology & Internet-Based Systems (SITIS).