Path: utzoo!yunexus!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!utgpu!cunews!wilf From: wilf@sce.carleton.ca (Wilf Leblanc) Newsgroups: comp.dsp Subject: DSP Survey (wanted) Summary: require info on dsp processors Keywords: dsp, signal processors Message-ID: Date: 31 May 90 16:20:05 GMT Article-I.D.: endor.wilf.644170805 Sender: news@ccs.carleton.ca (news) Distribution: comp Organization: Carleton University, Ottawa, Canada Lines: 32 We are about to implement a rather computationally complex algorithm on a single DSP. We are interested in any comments anyone might have regarding various DSP processors such as: AT&T DSP16A Motorola 56001 TI TMS320C25/C50 The algorithm consists of mainly filtering operations (convolutions), inner products, matrix multiplications, and a relatively small number of divisions, logic and control functions. We plan on using a fixed point DSP for cost reasons. We are familar with the TI line of processors but the C25 is too slow for our requirements and the C50 is not available yet (or is it ??). The DSP16A has a very low instruction cycle time, but is this deceiving ?? I think it is fairly widely accepted that the Motorola 56001 (at 27 MHz) is faster than the TI 'C25 (at 50 MHz, for most applications). However, how does the DSP16A compare with respect to speed, ease of use, and so on. Please email me directly at . If there is enough interest, I will post a summary. ---- wilf@sce.carleton.ca, 613-788-8130