Descripción
|
|
---|---|
imulation-based methods are the most common and general approaches to estimate the quantization noise of fixed-point implementations of Digital Signal Processing (DSP) systems. But as systems become more complex, this way of solving the word-length optimization problem leads to unaffordable execution times. In this paper, we present three new techniques to reduce the execution time of these simulation-based methods. The first technique reduces the number of word-length combinations to explore, effectively reducing the solution space. The second one relies on the relaxation of the confidence requirements of the simulations in the initial stages of the evaluation in order to accelerate the search. Experimental results show that, by applying these techniques, we can accelerate the optimization process by factors up to x130 with respect to classical greedy approaches for problems of different sizes. | |
Internacional
|
Si |
JCR del ISI
|
Si |
Título de la revista
|
Ieee Transactions on Computer-Aided Design of Integrated Circuits And Systems |
ISSN
|
0278-0070 |
Factor de impacto JCR
|
1,203 |
Información de impacto
|
Datos JCR del año 2013 |
Volumen
|
|
DOI
|
|
Número de revista
|
|
Desde la página
|
- |
Hasta la página
|
- |
Mes
|
SIN MES |
Ranking
|