From sokonkwo@cern.ch Thu Mar 10 23:26:57 1994 Path: ashpoly!bigsur!pacrim!krgate!eunha!cern!sokonkwo From: sokonkwo@cern.ch (Sven Okonkwo) Newsgroups: comp.dsp Subject: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Mar10.232657.3999@cern.ch> Date: Thu, 10 Mar 1994 23:26:57 GMT Organization: European Laboratory for Particle Physics, Geneva NNTP-Posting-Host: cern.ch X-Newsreader: WinVN 0.99.5 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. Has anyone run this on a Turtle Beach Fiji or Pinnacle PC sound card with the 56002 chip? -- Sven Okonkwo "Tape doesn't lie at 02:14. Ears do." NeXTstation Turbo Color / FreeBSD 2.1 From kkarlsson@unam.mx Fri Mar 11 04:58:57 1994 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!unam!kkarlsson From: kkarlsson@unam.mx (Kenji Karlsson) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Mar11.045857.2732@unam.mx> Date: Fri, 11 Mar 1994 04:58:57 GMT References: <1994Mar10.232657.3999@cern.ch> Organization: Universidad Nacional Autónoma de México (DGSCA) NNTP-Posting-Host: unam.mx X-Newsreader: Netscape 1.1N Lines: 10 In article <1994Mar10.232657.3999@cern.ch>, sokonkwo@cern.ch (Sven Okonkwo) 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. -- Kenji Karlsson "Do not worship the copy." Pentium 90 (OS/2 Warp) / Linux 1.2.13 From kmoreau@ox.ac.uk Fri Mar 11 13:38:57 1994 Path: ashpoly!cascade!transat!eunet-de!hansebox!ox!kmoreau From: kmoreau@ox.ac.uk (Karen Moreau) Newsgroups: comp.dsp Subject: Re: [INFO] Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Mar11.133857.7321@ox.ac.uk> Date: Fri, 11 Mar 1994 13:38:57 GMT References: <1994Mar10.232657.3999@cern.ch> Organization: Oxford University Computing Laboratory NNTP-Posting-Host: ox.ac.uk X-Newsreader: NewsWatcher 2.0b24 Lines: 10 In article <1994Mar10.232657.3999@cern.ch>, sokonkwo@cern.ch (Sven Okonkwo) 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. -- Karen Moreau "Hardware is temporary. The signal survives." Macintosh IIfx / System 7.5.1