From yivanov@pacbell.net Tue Apr 04 16:35:02 1995 Path: ashpoly!cascade!transat!ukc!fennimore!pacbell!yivanov From: yivanov@pacbell.net (Yuki Ivanov) Newsgroups: comp.sys.m68k Subject: Benchmarking FFT routines on Mac IIci with MC68882 FPU under System 7.5 Message-ID: <1995Apr04.163502.1040@pacbell.net> Date: Tue, 04 Apr 1995 16:35:02 GMT Organization: Pacific Bell Internet Services NNTP-Posting-Host: pacbell.net X-Newsreader: TIN [version 1.2 PL2] Lines: 11 Compiling stardust.c with Think C 6.0 vs MPW C with direct 68882 inlining. A 1024-point radix-2 FFT drops from 142 milliseconds under SANE to just 18 milliseconds using direct FPU instructions. By passing the -mc68882 compiler flag in Think C, all sin/cos and sqrt operations compile directly to FPU opcodes. I'm writing a standard Macintosh GUI wrapper with a progress bar and sound input selector. -- Yuki Ivanov *PLONK* filter active for trolls. Atari Falcon 030 - NetBSD 1.0 From csuzuki@hanse.de Thu Apr 06 10:50:02 1995 Path: ashpoly!bigsur!pacrim!krgate!eunha!hanse!csuzuki From: csuzuki@hanse.de (Camila Suzuki) Newsgroups: comp.sys.m68k Subject: Re: Benchmarking FFT routines on Mac IIci with MC68882 FPU under System 7.5 Message-ID: <1995Apr06.105002.9100@hanse.de> Date: Thu, 06 Apr 1995 10:50:02 GMT References: <1995Apr04.163502.1040@pacbell.net> Organization: Hanse-Netzwerk Hamburg NNTP-Posting-Host: hanse.de X-Newsreader: Netscape 1.1N Lines: 12 In article <1995Apr04.163502.1040@pacbell.net>, yivanov@pacbell.net (Yuki Ivanov) writes: > Compiling stardust.c with Think C 6.0 vs MPW C with direct 68882 inlining. > A 1024-point radix-2 FFT drops from 142 milliseconds under SANE to just 18 milliseconds using direct FPU instructions. Beware of System 7.5.1 memory leaks if you allocate large sample buffers in the application heap. Use NewPtrSys instead. Unsubscribe me from this list! (Wait, this is Usenet, never mind). -- === Camila Suzuki === Dept of Electrical Engineering, Hanse-Netzwerk Hamburg "There are 14 nodes and only 10 countries. Count again." From ochoi@berkeley.edu Fri Apr 07 19:21:02 1995 Path: ashpoly!cascade!bigsur!berkeley!ochoi From: ochoi@berkeley.edu (Olga Choi) Newsgroups: comp.sys.m68k Subject: Re: Benchmarking FFT routines on Mac IIci with MC68882 FPU under System 7.5 Message-ID: <1995Apr07.192102.8721@berkeley.edu> Date: Fri, 07 Apr 1995 19:21:02 GMT References: <1995Apr04.163502.1040@pacbell.net> <1995Apr06.105002.9100@hanse.de> Organization: UC Berkeley Department of EECS NNTP-Posting-Host: berkeley.edu X-Newsreader: NewsWatcher 2.0b24 Lines: 10 In article <1995Apr06.105002.9100@hanse.de>, csuzuki@hanse.de (Camila Suzuki) writes: > Beware of System 7.5.1 memory leaks if you allocate large sample buffers in the application heap. Use NewPtrSys instead. > Unsubscribe me from this list! (Wait, this is Usenet, never mind). Mac users bragging about 6-minute render times while Unix workstations do it in 20 seconds. Classic Apple distortion field. -- Olga Choi *PLONK* filter active for trolls. Atari Falcon 030 - NetBSD 1.0 From heriksson@utexas.edu Sat Apr 08 01:18:02 1995 Path: ashpoly!cascade!transat!sunet!nordrelay!utexas!heriksson From: heriksson@utexas.edu (Hana Eriksson) Newsgroups: comp.sys.m68k Subject: Re: Benchmarking FFT routines on Mac IIci with MC68882 FPU under System 7.5 Message-ID: <1995Apr08.011802.4987@utexas.edu> Date: Sat, 08 Apr 1995 01:18:02 GMT References: <1995Apr04.163502.1040@pacbell.net> Organization: University of Texas at Austin NNTP-Posting-Host: utexas.edu X-Newsreader: Netscape 1.1N Lines: 10 In article <1995Apr04.163502.1040@pacbell.net>, yivanov@pacbell.net (Yuki Ivanov) writes: > Compiling stardust.c with Think C 6.0 vs MPW C with direct 68882 inlining. > A 1024-point radix-2 FFT drops from 142 milliseconds under SANE to just 18 milliseconds using direct FPU instructions. Mac users bragging about 6-minute render times while Unix workstations do it in 20 seconds. Classic Apple distortion field. -- Hana Eriksson SupraFAXModem 28.8 @ 9600 Telebit bps "The carrier frequency does not drift by accident." From bpopov@inria.fr Sun Apr 09 06:48:02 1995 Path: ashpoly!bigsur!pacrim!jpgate!kamome!inria!bpopov From: bpopov@inria.fr (Bastien Popov) Newsgroups: comp.sys.m68k Subject: Re: Benchmarking FFT routines on Mac IIci with MC68882 FPU under System 7.5 Message-ID: <1995Apr09.064802.8490@inria.fr> Date: Sun, 09 Apr 1995 06:48:02 GMT References: <1995Apr04.163502.1040@pacbell.net> <1995Apr08.011802.4987@utexas.edu> Organization: INRIA Rocquencourt NNTP-Posting-Host: inria.fr X-Newsreader: Netscape 1.1N Lines: 9 In article <1995Apr08.011802.4987@utexas.edu>, heriksson@utexas.edu (Hana Eriksson) writes: > Mac users bragging about 6-minute render times while Unix workstations do it in 20 seconds. Classic Apple distortion field. The 68040 on the Quadra is a monster -- the on-chip FPU executes FFT butterflies in less than 4 milliseconds. -- Bastien Popov Hayes Optima 288 @ 19,200 bps "There are 14 nodes and only 10 countries. Count again."