Title :
Cost functions for mapping DSP algorithms onto multiprocessors
Author :
Rabideau, Daniel J. ; Steinhardt, Allan O.
Author_Institution :
Rome Lab., Griffiss AFB, NY, USA
fDate :
1/1/1995 12:00:00 AM
Abstract :
In this correspondence, we examine several cost functions that have been proposed for automating the mapping of algorithms onto multiprocessors. Through a case study of the recursive least squares problem, we develop improved cost functions. One of these (the Min-Max+Idle cost function) performed better than the others and was applied to the related problem of full QR decomposition. Experiments on an iPSC/860 hypercube confirm that automated mapping can lead to lower execution times than published mappings
Keywords :
distributed memory systems; least squares approximations; parallel algorithms; real-time systems; recursive estimation; signal processing; DSP algorithms; automated mapping; cost functions; distributed-memory multiprocessor; full QR decomposition; iPSC/860 hypercube; lower execution times; parallel processing; recursive least squares problem; Cost function; Data models; Digital signal processing; Frequency; Gaussian noise; Hafnium; Least squares approximation; Parameter estimation; Signal processing algorithms; Signal to noise ratio;
Journal_Title :
Signal Processing, IEEE Transactions on
