Algorithmic Optimization for Floating Point DSP Mathematics Libraries
This paper discusses algorithmic optimization techniques for floating point DSPs, which are independent of architectural differences and provide tools for writing the fastest floating point DSP algorithms for any required precision. We show algorithms for polynomial evaluation used generally as an approximation tool. These algorithms are two to four times faster than the straightforward implementation of Horner’s rule.
Please disable any pop-up blockers for proper viewing of this Whitepaper.