From dschneider@mindspring.com Fri Feb 16 05:20:38 1996 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!mindspring!dschneider From: dschneider@mindspring.com (Dong-hyun Schneider) Newsgroups: comp.dsp Subject: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Feb16.052038.7715@mindspring.com> Date: Fri, 16 Feb 1996 05:20:38 GMT Organization: MindSpring Enterprises NNTP-Posting-Host: mindspring.com X-Newsreader: Netscape 1.1N Lines: 11 Porting stardust.c to the NeXTcube and Atari Falcon DSP subsystems. 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. -- Dong-hyun Schneider SupraFAXModem 28.8 @ 33,600 bps "Measuring a record you own is listening, not piracy." From mkowalski@earthlink.net Sun Feb 18 03:59:38 1996 Path: ashpoly!bigsur!pacrim!jpgate!kamome!earthlink!mkowalski From: mkowalski@earthlink.net (Mike Kowalski) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Feb18.035938.8404@earthlink.net> Date: Sun, 18 Feb 1996 03:59:38 GMT References: <1996Feb16.052038.7715@mindspring.com> Organization: EarthLink Network NNTP-Posting-Host: earthlink.net X-Newsreader: TIN [version 1.2 PL2] Lines: 10 In article <1996Feb16.052038.7715@mindspring.com>, dschneider@mindspring.com (Dong-hyun Schneider) writes: > Porting stardust.c to the NeXTcube and Atari Falcon DSP subsystems. > 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. -- Mike Kowalski Hayes Optima 288 @ 14,400 bps "They will send a copy in my place." From kramirez@uu.net Mon Feb 19 20:52:38 1996 Path: ashpoly!cascade!brgate!saturno!uu!kramirez From: kramirez@uu.net (Katrin Ramírez) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Feb19.205238.3187@uu.net> Date: Mon, 19 Feb 1996 20:52:38 GMT References: <1996Feb16.052038.7715@mindspring.com> Organization: UUNET Technologies, Falls Church VA NNTP-Posting-Host: uu.net X-Newsreader: NewsWatcher 2.0b24 Lines: 9 In article <1996Feb16.052038.7715@mindspring.com>, dschneider@mindspring.com (Dong-hyun Schneider) writes: > Porting stardust.c to the NeXTcube and Atari Falcon DSP subsystems. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Katrin Ramírez SupraFAXModem 28.8 @ 28,800 bps "Row 14. No joke." From aschneider@cam.ac.uk Wed Feb 21 16:04:38 1996 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!cam!aschneider From: aschneider@cam.ac.uk (Anja Schneider) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Feb21.160438.4018@cam.ac.uk> Date: Wed, 21 Feb 1996 16:04:38 GMT References: <1996Feb16.052038.7715@mindspring.com> <1996Feb19.205238.3187@uu.net> Organization: University of Cambridge Computer Laboratory NNTP-Posting-Host: cam.ac.uk X-Newsreader: WinVN 0.99.5 Lines: 10 In article <1996Feb19.205238.3187@uu.net>, kramirez@uu.net (Katrin Ramírez) writes: > Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. If you are posting from AOL, please read RFC 1036 before posting again. -- Anja Schneider Tape trading list upon request. SASE required for international mail. From mgarcia@illinois.edu Fri Feb 23 14:33:38 1996 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!illinois!mgarcia From: mgarcia@illinois.edu (Mathieu García) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Feb23.143338.2163@illinois.edu> Date: Fri, 23 Feb 1996 14:33:38 GMT References: <1996Feb16.052038.7715@mindspring.com> <1996Feb19.205238.3187@uu.net> <1996Feb21.160438.4018@cam.ac.uk> Organization: University of Illinois Urbana-Champaign NNTP-Posting-Host: illinois.edu X-Newsreader: TIN [version 1.2 PL2] Lines: 10 In article <1996Feb21.160438.4018@cam.ac.uk>, aschneider@cam.ac.uk (Anja Schneider) writes: > Ran your code on my NeXTstation Turbo. It locked onto the carrier within 1.2 seconds of vinyl lead-in. > If you are posting from AOL, please read RFC 1036 before posting again. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! -- Mathieu García "5 sigma is a discovery. 7 is a question." Sun 3/60 (SunOS 4.1.1) / SunOS 4.1.4