home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.dec
- Path: sparky!uunet!elroy.jpl.nasa.gov!ames!saimiri.primate.wisc.edu!aplcen.apl.jhu.edu!news
- From: shif@aplpy.jhuapl.edu (Gary D. Shiflett)
- Subject: Alpha fft performance
- Message-ID: <1992Dec31.164221.27734@aplcen.apl.jhu.edu>
- Sender: news@aplcen.apl.jhu.edu (USENET News System)
- Organization: Johns Hopkins University Applied Physics Lab
- Date: Thu, 31 Dec 92 16:42:21 GMT
- Lines: 36
-
-
-
- In comp.dsp Michael.Malak@halluc.com (Michael Malak) asks:
-
- Michael> Does anyone have any information about how fast the i860 can do an FFT,
- Michael> either theoretical or measured?
-
- and alans@ll.mit.edu (Alan Stein) responds:
-
- Alan>In the 860 bible (i860 Microprocessor Architecture by Neil Margulis),
- Alan>the author codes an FFT which achieves 57/80 MFlops. This is not really
- Alan>what the beast is designed to do, IMHO. Seems I'm always writing 3
- Alan>different versions of the same routine - one for extremely short vectors
- Alan>(throw everything in registers), one for medium-sized "if it all fits
- Alan>in cache" vectors, and one for long vectors that use pipelined memory
- Alan>loads, bypassing the cache because they would just thrash. I suspect
- Alan>I'd wind up doing much the same if I were coding an FFT, which
- Alan>I haven't [as yet] had to do.
-
- Malak was interested in comparing the i860 FFT performance to other
- commercially available DSPs. My question is: what is the FFT performance of
- the family of DEC alpha micros (theoretical, simulated, or measured ???)?
- I would be interested in knowing not only MFlops, but also FFT algorithm,
- transform size(s), optimum execution times, and any stipulations (as
- described by Alan above) associated with the benchmarks.
-
- Thanks in advance.
-
-
- Gary Shiflett
- Applied Physics Laboratory
- Johns Hopkins University
- gary_shiflett@jhuapl.edu
- (301)953-6000x8141
-
-
-