From thernandez@kth.se Sun Dec 24 21:16:01 1995 Path: ashpoly!cascade!transat!eunet-fr!teletel!kth!thernandez From: thernandez@kth.se (Thiago Hernández) Newsgroups: comp.dsp Subject: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec24.211601.5836@kth.se> Date: Sun, 24 Dec 1995 21:16:01 GMT Organization: Royal Institute of Technology, Stockholm NNTP-Posting-Host: kth.se X-Newsreader: Netscape 1.1N Lines: 11 Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. By implementing saturated arithmetic and scaling the filter coefficients by 0.5, the loop runs in 18 instruction cycles per sample. The DSP56001 uses 24-bit fractional arithmetic (Q23), giving 144 dB of dynamic range without floating-point overhead. Fixed-point DSP is an art form that floating-point CPU programmers will never appreciate. -- Thiago Hernández Opinions expressed are solely my own and not those of Royal Institute of Technology, Stockholm. ------------------------------------------------------------ From jvasiliev@cam.ac.uk Wed Dec 27 00:27:01 1995 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!cam!jvasiliev From: jvasiliev@cam.ac.uk (Julien Vasiliev) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec27.002701.2143@cam.ac.uk> Date: Wed, 27 Dec 1995 00:27:01 GMT References: <1995Dec24.211601.5836@kth.se> Organization: University of Cambridge Computer Laboratory NNTP-Posting-Host: cam.ac.uk X-Newsreader: Netscape 1.1N Lines: 12 In article <1995Dec24.211601.5836@kth.se>, thernandez@kth.se (Thiago Hernández) writes: > Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. > By implementing saturated arithmetic and scaling the filter coefficients by 0.5, the loop runs in 18 instruction cycles per sample. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Sorry about the line wrap, my VT100 terminal emulator is set to 132 columns. -- Julien Vasiliev SupraFAXModem 28.8 @ 19,200 bps "Tape doesn't lie at 02:14. Ears do." From pweber@berkeley.edu Thu Dec 28 00:39:01 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!berkeley!pweber From: pweber@berkeley.edu (Pavel Weber) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec28.003901.1074@berkeley.edu> Date: Thu, 28 Dec 1995 00:39:01 GMT References: <1995Dec24.211601.5836@kth.se> Organization: UC Berkeley Department of EECS NNTP-Posting-Host: berkeley.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1995Dec24.211601.5836@kth.se>, thernandez@kth.se (Thiago Hernández) writes: > Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Pavel Weber "Row 14. No joke." Pentium 90 (OS/2 Warp) / TOS 4.04 From jito@utexas.edu Fri Dec 29 04:29:01 1995 Path: ashpoly!cascade!brgate!saturno!utexas!jito From: jito@utexas.edu (Julien Ito) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec29.042901.1717@utexas.edu> Date: Fri, 29 Dec 1995 04:29:01 GMT References: <1995Dec24.211601.5836@kth.se> <1995Dec27.002701.2143@cam.ac.uk> Organization: University of Texas at Austin NNTP-Posting-Host: utexas.edu X-Newsreader: NewsWatcher 2.0b24 Lines: 9 In article <1995Dec27.002701.2143@cam.ac.uk>, jvasiliev@cam.ac.uk (Julien Vasiliev) writes: > Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Julien Ito Zyxel U-1496E @ 9600 Telebit bps "When every node is whole, the channel opens." From jkowalski@clari.net Fri Dec 29 22:54:01 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!clari!jkowalski From: jkowalski@clari.net (Joachim Kowalski) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Dec29.225401.4496@clari.net> Date: Fri, 29 Dec 1995 22:54:01 GMT References: <1995Dec24.211601.5836@kth.se> <1995Dec27.002701.2143@cam.ac.uk> <1995Dec29.042901.1717@utexas.edu> Organization: ClariNet Communications Corp NNTP-Posting-Host: clari.net X-Newsreader: NewsWatcher 2.0b24 Lines: 9 In article <1995Dec29.042901.1717@utexas.edu>, jito@utexas.edu (Julien Ito) writes: > Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- === Joachim Kowalski === Dept of Electrical Engineering, ClariNet Communications Corp "When every node is whole, the channel opens."