Kira Kraft

• The ever increasing data traffic in our modern society poses severe challenges to any communication system. To ensure reliable data transmission and storage, error-correction coding is applied to make the data robust to disturbances. The challenge of error-correction coding lies in the decoder unit, where errors are detected and ideally corrected. In addition to providing a good error-correction performance, the increasing requirements posed by new standards are directly reflected to the decoder unit, stressing both a well-performing and highly efficient implementation. Requiring both a well-performing and highly efficient implementation equals a Pareto optimization problem, which cannot be solved optimally for all objectives at the same time. Therefore, application-specific trade-offs have to be made. In this thesis, two applications to error-correction coding are investigated. The first application is maximum-likelihood decoding, a mathematical problem, which solves the decoding problem optimally. The goal of this use-case is therefore to improve the implementation efficiency. The second application is error correction for DRAMs. DRAMs are already highly optimized in their implementation efficiency, so the goal of this use-case is to improve the error-correction performance of the decoder unit.