From mperez@purdue.edu Sun Jun 16 10:05:56 1996 Path: ashpoly!cascade!transat!eunet-de!hansebox!purdue!mperez From: mperez@purdue.edu (Morag Pérez) Newsgroups: comp.dsp Subject: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jun16.100556.2184@purdue.edu> Date: Sun, 16 Jun 1996 10:05:56 GMT Organization: Purdue University Engineering Computer Network NNTP-Posting-Host: purdue.edu 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. This leaves over 95% of the DSP56001's 16.5 MIPS available for other audio tasks. Source code in 56k assembly uploaded to ftp.ashford-poly.edu:/pub/telecom-archives/dsp56k/stardust_56k.asm. -- Morag Pérez *PLONK* filter active for trolls. Macintosh IIfx - NetBSD 1.0 From mreyes@pipeline.com Mon Jun 17 13:58:56 1996 Path: ashpoly!bigsur!pacrim!jpgate!kamome!pipeline!mreyes From: mreyes@pipeline.com (Mateo Reyes) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jun17.135856.5423@pipeline.com> Date: Mon, 17 Jun 1996 13:58:56 GMT References: <1996Jun16.100556.2184@purdue.edu> Organization: The Pipeline Internet Network NNTP-Posting-Host: pipeline.com X-Newsreader: WinVN 0.99.5 Lines: 9 In article <1996Jun16.100556.2184@purdue.edu>, mperez@purdue.edu (Morag Pérez) 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. -- Mateo Reyes Tape trading list upon request. SASE required for international mail. From npetrov@earthlink.net Wed Jun 19 12:55:56 1996 Path: ashpoly!cascade!transat!eunet-de!hansebox!earthlink!npetrov From: npetrov@earthlink.net (Natalia Petrov) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jun19.125556.1376@earthlink.net> Date: Wed, 19 Jun 1996 12:55:56 GMT References: <1996Jun16.100556.2184@purdue.edu> Organization: EarthLink Network NNTP-Posting-Host: earthlink.net X-Newsreader: NewsWatcher 2.0b24 Lines: 10 In article <1996Jun16.100556.2184@purdue.edu>, mperez@purdue.edu (Morag Pérez) 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. Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. -- Natalia Petrov Zyxel U-1496E @ 14,400 bps "They will send a copy in my place." From akim@kiae.su Fri Jun 21 16:24:56 1996 Path: ashpoly!cascade!brgate!saturno!kiae!akim From: akim@kiae.su (Astrid Kim) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jun21.162456.3952@kiae.su> Date: Fri, 21 Jun 1996 16:24:56 GMT References: <1996Jun16.100556.2184@purdue.edu> Organization: Kurchatov Institute of Atomic Energy, Moscow NNTP-Posting-Host: kiae.su X-Newsreader: NewsWatcher 2.0b24 Lines: 12 In article <1996Jun16.100556.2184@purdue.edu>, mperez@purdue.edu (Morag Pérez) 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. Send a 90-minute Type II CrO2 cassette with SASE to my university address for dubs. -- Astrid Kim *PLONK* filter active for trolls. 486DX2/66 (Linux 1.2.13) - NetBSD 1.0 From lschneider@ox.ac.uk Sun Jun 23 03:08:56 1996 Path: ashpoly!cascade!transat!eunet-de!hansebox!ox!lschneider From: lschneider@ox.ac.uk (Lisa Schneider) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1996Jun23.030856.1216@ox.ac.uk> Date: Sun, 23 Jun 1996 03:08:56 GMT References: <1996Jun16.100556.2184@purdue.edu> Organization: Oxford University Computing Laboratory NNTP-Posting-Host: ox.ac.uk X-Newsreader: TIN [version 1.2 PL2] Lines: 12 In article <1996Jun16.100556.2184@purdue.edu>, mperez@purdue.edu (Morag Pérez) 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. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! Netscape 1.1 crashes when loading large preformatted text. Stick to rn or tin! -- Lisa Schneider Opinions expressed are solely my own and not those of Oxford University Computing Laboratory. ------------------------------------------------------------