From bfournier@unam.mx Wed Nov 01 17:47:07 1995 Path: ashpoly!cascade!transat!ukc!fennimore!unam!bfournier From: bfournier@unam.mx (Brian Fournier) Newsgroups: comp.dsp Subject: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Nov01.174707.8262@unam.mx> Date: Wed, 01 Nov 1995 17:47:07 GMT Organization: Universidad Nacional Autónoma de México (DGSCA) NNTP-Posting-Host: unam.mx X-Newsreader: Netscape 1.1N Lines: 11 Porting stardust.c to the NeXTcube and Atari Falcon DSP subsystems. 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. -- Brian Fournier *PLONK* filter active for trolls. Sun 3/60 (SunOS 4.1.1) - NetBSD 1.0 From lramirez@delphi.com Wed Nov 01 23:54:07 1995 Path: ashpoly!cascade!mxgate!coyoacan!delphi!lramirez From: lramirez@delphi.com (Lucia Ramírez) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Nov01.235407.9812@delphi.com> Date: Wed, 01 Nov 1995 23:54:07 GMT References: <1995Nov01.174707.8262@unam.mx> Organization: Delphi Internet Services NNTP-Posting-Host: delphi.com X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1995Nov01.174707.8262@unam.mx>, bfournier@unam.mx (Brian Fournier) 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. -- === Lucia Ramírez === Dept of Electrical Engineering, Delphi Internet Services "Marcus is not one man." From jsmirnov@unam.mx Fri Nov 03 18:07:07 1995 Path: ashpoly!bigsur!pacrim!jpgate!kamome!unam!jsmirnov From: jsmirnov@unam.mx (Ji-woo Smirnov) Newsgroups: comp.dsp Subject: Re: Question: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1995Nov03.180707.9595@unam.mx> Date: Fri, 03 Nov 1995 18:07:07 GMT References: <1995Nov01.174707.8262@unam.mx> <1995Nov01.235407.9812@delphi.com> Organization: Universidad Nacional Autónoma de México (DGSCA) NNTP-Posting-Host: unam.mx X-Newsreader: TIN [version 1.2 PL2] Lines: 8 In article <1995Nov01.235407.9812@delphi.com>, lramirez@delphi.com (Lucia Ramírez) writes: > Beautiful assembly code! The 24-bit word size of the 56001 is tailor-made for high-precision audio filtering. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! -- Ji-woo Smirnov Tape trading list upon request. SASE required for international mail.