Abstract-Many digital signal processing and communication algorithms are first simulated using floating-point arithmetic and later transformed into fixed-point arithmetic to reduce implementation complexity. For the floating-point to fixed-point transformation, this paper describes two methods within an automated transformation environment. The first method, a gradientbased search for single-objective optimization with sensitivity information, provides a single solution, and can become trapped in local optima. The second method, a genetic algorithm for multiobjective optimization, provides a family of solutions that form a tradeoff curve for signal quality vs. implementation complexity. We provide case studies for an infinite impulse response filter. In the case study, implementation complexity is lookup table area for a field programmable gate array (FPGA) realization. We have made the transformation methods available in a software release on the Web.

No abstract

Abstract-Given a wireline transmitter/receiver pair with fixed parameters (modulation and coding methods, transmission, bandwidth, receiver noise floor, etc.) that achieves a certain bit rate on a single pair of wires, one might expect that using two pairs of wires (and two sets of transceivers) would double the bit rate. However, the presence of crosstalk induced by the coupling of the energy across the wire pairs may cause significant reduction in the expected bit rate. In this paper, we present a multiple-input multiple-output (MIMO) discrete multitone (DMT) system that achieves double the bit rate of a traditional single transmitterreceiver system when using 2 transmitters and 2 receivers operating over two wires. The improvement has been achieved both in desktop simulation and in a real-time testbed. The testbed also allows rapid exploration of many design tradeoffs in a DMT system.