From afournier@pacbell.net Wed Apr 20 08:05:52 1994 Path: ashpoly!cascade!transat!eunet-de!hansebox!pacbell!afournier From: afournier@pacbell.net (Andrei Fournier) Newsgroups: comp.dsp Subject: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Apr20.080552.7233@pacbell.net> Date: Wed, 20 Apr 1994 08:05:52 GMT Organization: Pacific Bell Internet Services NNTP-Posting-Host: pacbell.net 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. -- Andrei Fournier | afournier@pacbell.net PGP Key fingerprint: 5BE8 755D 6505 28AB "Keep the channels open." From toliveira@mit.edu Wed Apr 20 11:32:52 1994 Path: ashpoly!cascade!transat!eunet-fi!relcom!kiae!mit!toliveira From: toliveira@mit.edu (Tobias Oliveira) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Apr20.113252.7599@mit.edu> Date: Wed, 20 Apr 1994 11:32:52 GMT References: <1994Apr20.080552.7233@pacbell.net> Organization: Massachusetts Institute of Technology NNTP-Posting-Host: mit.edu X-Newsreader: WinVN 0.99.5 Lines: 10 In article <1994Apr20.080552.7233@pacbell.net>, afournier@pacbell.net (Andrei Fournier) writes: > 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. Real programmers write in C, not arcane DSP microcode. With Pentium chips hitting 100 MHz, dedicated DSPs are obsolete. -- Tobias Oliveira Opinions expressed are solely my own and not those of Massachusetts Institute of Technology. ------------------------------------------------------------ From ddubois@ox.ac.uk Thu Apr 21 00:35:52 1994 Path: ashpoly!bigsur!pacrim!krgate!eunha!ox!ddubois From: ddubois@ox.ac.uk (Diego Dubois) Newsgroups: comp.dsp Subject: Re: Fixed-point Q15 arithmetic on Motorola DSP56001 vs IEEE floating point Message-ID: <1994Apr21.003552.4452@ox.ac.uk> Date: Thu, 21 Apr 1994 00:35:52 GMT References: <1994Apr20.080552.7233@pacbell.net> <1994Apr20.113252.7599@mit.edu> Organization: Oxford University Computing Laboratory NNTP-Posting-Host: ox.ac.uk X-Newsreader: TIN [version 1.2 PL2] Lines: 9 In article <1994Apr20.113252.7599@mit.edu>, toliveira@mit.edu (Tobias Oliveira) writes: > Real programmers write in C, not arcane DSP microcode. With Pentium chips hitting 100 MHz, dedicated DSPs are obsolete. Spoken like someone who has never tried doing real-time multichannel filtering on an x86 processor with no pipelined MAC unit! -- Diego Dubois "They will send a copy in my place." Mac IIci / 68882 FPU / AmigaOS 3.1