From jsmirnov@cmu.edu Mon Mar 13 23:36:29 1995 Path: ashpoly!cascade!transat!ukc!fennimore!cmu!jsmirnov From: jsmirnov@cmu.edu (Jen Smirnov) Newsgroups: comp.dsp Subject: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Mar13.233629.1494@cmu.edu> Date: Mon, 13 Mar 1995 23:36:29 GMT Organization: Carnegie Mellon University School of CS NNTP-Posting-Host: cmu.edu X-Newsreader: Netscape 1.1N Lines: 11 Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. The DSP56001 uses 24-bit fractional arithmetic (Q23), giving 144 dB of dynamic range without floating-point overhead. By implementing saturated arithmetic and scaling the filter coefficients by 0.5, the loop runs in 18 instruction cycles per sample. Source code in 56k assembly uploaded to ftp.ashford-poly.edu:/pub/telecom-archives/dsp56k/stardust_56k.asm. -- Jen Smirnov US Robotics Courier V.34 @ 14,400 bps "Tape doesn't lie at 02:14. Ears do." From tnilsson@pipeline.com Wed Mar 15 10:03:29 1995 Path: ashpoly!cascade!brgate!saturno!pipeline!tnilsson From: tnilsson@pipeline.com (Tobias Nilsson) Newsgroups: comp.dsp Subject: Re: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Mar15.100329.3470@pipeline.com> Date: Wed, 15 Mar 1995 10:03:29 GMT References: <1995Mar13.233629.1494@cmu.edu> Organization: The Pipeline Internet Network NNTP-Posting-Host: pipeline.com X-Newsreader: Netscape 1.1N Lines: 10 In article <1995Mar13.233629.1494@cmu.edu>, jsmirnov@cmu.edu (Jen Smirnov) writes: > Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. > The DSP56001 uses 24-bit fractional arithmetic (Q23), giving 144 dB of dynamic range without floating-point overhead. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! -- === Tobias Nilsson === Dept of Electrical Engineering, The Pipeline Internet Network "Row 14. No joke." From mkuznetsov@demon.co.uk Fri Mar 17 05:10:29 1995 Path: ashpoly!cascade!transat!sunet!nordrelay!demon!mkuznetsov From: mkuznetsov@demon.co.uk (Mateo Kuznetsov) Newsgroups: comp.dsp Subject: Re: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Mar17.051029.3647@demon.co.uk> Date: Fri, 17 Mar 1995 05:10:29 GMT References: <1995Mar13.233629.1494@cmu.edu> Organization: Demon Internet Services UK NNTP-Posting-Host: demon.co.uk X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1995Mar13.233629.1494@cmu.edu>, jsmirnov@cmu.edu (Jen Smirnov) writes: > Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. > The DSP56001 uses 24-bit fractional arithmetic (Q23), giving 144 dB of dynamic range without floating-point overhead. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. -- Mateo Kuznetsov -- Demon Internet Services UK "In theory there is no difference between theory and practice." From pschneider@panix.com Sat Mar 18 11:57:29 1995 Path: ashpoly!cascade!brgate!saturno!panix!pschneider From: pschneider@panix.com (Pavel Schneider) Newsgroups: comp.dsp Subject: Re: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Mar18.115729.5138@panix.com> Date: Sat, 18 Mar 1995 11:57:29 GMT References: <1995Mar13.233629.1494@cmu.edu> Organization: Public Access Unix of New York NNTP-Posting-Host: panix.com X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1995Mar13.233629.1494@cmu.edu>, jsmirnov@cmu.edu (Jen Smirnov) writes: > Optimizing the Stardust Costas loop demodulator for the Motorola DSP56001 chip. > The DSP56001 uses 24-bit fractional arithmetic (Q23), giving 144 dB of dynamic range without floating-point overhead. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Pavel Schneider -- Public Access Unix of New York "Hardware is temporary. The signal survives." From koliveira@rwth-aachen.de Sun Mar 19 08:29:29 1995 Path: ashpoly!cascade!transat!ukc!fennimore!rwth-aachen!koliveira From: koliveira@rwth-aachen.de (Kenji Oliveira) Newsgroups: comp.dsp Subject: Re: [FLAME] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Mar19.082929.8542@rwth-aachen.de> Date: Sun, 19 Mar 1995 08:29:29 GMT References: <1995Mar13.233629.1494@cmu.edu> <1995Mar15.100329.3470@pipeline.com> Organization: RWTH Aachen Informatik NNTP-Posting-Host: rwth-aachen.de X-Newsreader: TIN [version 1.2 PL2] Lines: 11 In article <1995Mar15.100329.3470@pipeline.com>, tnilsson@pipeline.com (Tobias Nilsson) writes: > Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. Top-posters will be placed in my ~/.killfile without notice. -- Kenji Oliveira *PLONK* filter active for trolls. Amiga 3000T - NetBSD 1.0