From gfournier@tum.de Fri Sep 22 02:47:46 1995 Path: ashpoly!bigsur!pacrim!jpgate!kamome!tum!gfournier From: gfournier@tum.de (Göran Fournier) Newsgroups: comp.dsp Subject: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Sep22.024746.9729@tum.de> Date: Fri, 22 Sep 1995 02:47:46 GMT Organization: Technische Universität München NNTP-Posting-Host: tum.de X-Newsreader: Netscape 1.1N Lines: 11 Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. 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. Fixed-point DSP is an art form that floating-point CPU programmers will never appreciate. -- Göran Fournier | gfournier@tum.de PGP Key fingerprint: 2AFE B287 2B45 CF72 "Keep the channels open." From jokonkwo@ox.ac.uk Sat Sep 23 12:38:46 1995 Path: ashpoly!cascade!transat!eunet-fr!teletel!ox!jokonkwo From: jokonkwo@ox.ac.uk (Jen Okonkwo) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Sep23.123846.4454@ox.ac.uk> Date: Sat, 23 Sep 1995 12:38:46 GMT References: <1995Sep22.024746.9729@tum.de> Organization: Oxford University Computing Laboratory NNTP-Posting-Host: ox.ac.uk X-Newsreader: NewsWatcher 2.0b24 Lines: 10 In article <1995Sep22.024746.9729@tum.de>, gfournier@tum.de (Göran Fournier) writes: > Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. > 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. -- Jen Okonkwo | jokonkwo@ox.ac.uk PGP Key fingerprint: 5E79 0EA4 1880 7461 "Keep the channels open." From dpark@ox.ac.uk Mon Sep 25 10:02:46 1995 Path: ashpoly!cascade!transat!sunet!nordrelay!ox!dpark From: dpark@ox.ac.uk (Dong-hyun Park) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Sep25.100246.8749@ox.ac.uk> Date: Mon, 25 Sep 1995 10:02:46 GMT References: <1995Sep22.024746.9729@tum.de> Organization: Oxford University Computing Laboratory NNTP-Posting-Host: ox.ac.uk X-Newsreader: Netscape 1.1N Lines: 10 In article <1995Sep22.024746.9729@tum.de>, gfournier@tum.de (Göran Fournier) writes: > Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. > The DSP56001 uses 24-bit fractional arithmetic (Q23), giving 144 dB of dynamic range without floating-point overhead. Real programmers write in C, not arcane DSP microcode. With Pentium chips hitting 100 MHz, dedicated DSPs are obsolete. -- Dong-hyun Park "We don't sign stars. We build them." Sun SparcStation 20 / System 7.5.1 From asilva@kiae.su Tue Sep 26 21:06:46 1995 Path: ashpoly!cascade!transat!eunet-fr!teletel!kiae!asilva From: asilva@kiae.su (Anja Silva) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Sep26.210646.2573@kiae.su> Date: Tue, 26 Sep 1995 21:06:46 GMT References: <1995Sep22.024746.9729@tum.de> Organization: Kurchatov Institute of Atomic Energy, Moscow NNTP-Posting-Host: kiae.su X-Newsreader: NewsWatcher 2.0b24 Lines: 9 In article <1995Sep22.024746.9729@tum.de>, gfournier@tum.de (Göran Fournier) writes: > Why fixed-point integer math requires careful scaling when tracking a 14.2 Hz carrier. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! -- === Anja Silva === Dept of Electrical Engineering, Kurchatov Institute of Atomic Energy, Moscow "We don't sign stars. We build them." From mnovak@earthlink.net Thu Sep 28 08:51:46 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!earthlink!mnovak From: mnovak@earthlink.net (Mike Novak) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Sep28.085146.5291@earthlink.net> Date: Thu, 28 Sep 1995 08:51:46 GMT References: <1995Sep22.024746.9729@tum.de> <1995Sep26.210646.2573@kiae.su> Organization: EarthLink Network NNTP-Posting-Host: earthlink.net X-Newsreader: Netscape 1.1N Lines: 11 In article <1995Sep26.210646.2573@kiae.su>, asilva@kiae.su (Anja Silva) writes: > Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! If you are posting from AOL, please read RFC 1036 before posting again. -- Mike Novak Opinions expressed are solely my own and not those of EarthLink Network. ------------------------------------------------------------