Title :
FILU-200 DSP coprocessor IP core
Author :
Bleakley, Chris ; Berg, Vincent ; Rodriguez, Jose ; Murray, Brian
Author_Institution :
Massana Inc., USA
Abstract :
Massana´s FILU-200 DSP coprocessor IP core enables fast development of DSP in embedded applications. The FILU-200´s library of pre-programmed DSP functions are accessed via C function calls from a host RISC processor. The FILU-200 matches high-end DSPs for MIPS capability at a fraction of the silicon cost. The dual MAC architecture with dual ALU, and dual barrel shifter is tailored for highest DSP performance in small silicon area. The FILU-200 is particularly suited to fast complex number arithmetic yielding a radix-4 FFT butterfly in 8 cycles, i.e. 1024-point complex FFT in 103 /spl mu/s. Near floating point precision is supplied by block floating point arithmetic combined with a 20-bit internal data path. Optimized building block DSP functions, including FFT, IFFT, FIR filters, IIR filters and matrix operations, are microcoded in ROM and are supplied with the FILU-200. The small silicon area of the FILU-200, 30 K gates, enables its adoption in cost sensitive applications. Typical applications include soft modems, G.Lite, DAB, VoIP and audio.
Keywords :
FIR filters; IIR filters; coprocessors; digital signal processing chips; fast Fourier transforms; floating point arithmetic; inverse problems; matrix algebra; reduced instruction set computing; C function calls; DAB; FFT; FILU-200 DSP coprocessor IP core; FIR filters; G.Lite; IFFT; IIR filters; Massana; ROM; VoIP; audio; block floating point arithmetic; complex FFT; data path; dual ALU; dual MAC architecture; dual barrel shifter; embedded applications; fast complex number arithmetic; floating point precision; host RISC processor; matrix operations; microcode; pre-programmed DSP functions; radix-4 FFT butterfly; silicon area; soft modems; Coprocessors; Costs; Digital signal processing; Finite impulse response filter; Floating-point arithmetic; IIR filters; Libraries; Read only memory; Reduced instruction set computing; Silicon;
Conference_Titel :
Signals, Systems, and Computers, 1999. Conference Record of the Thirty-Third Asilomar Conference on
Print_ISBN :
0-7803-5700-0
DOI :
10.1109/ACSSC.1999.832430