Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.lang.forth > #15007
| From | Paul Rubin <no.email@nospam.invalid> |
|---|---|
| Newsgroups | comp.lang.forth |
| Subject | Re: GA144 Documents |
| References | <k0hc8h$itc$1@dont-email.me> <7xehn77hth.fsf@ruckus.brouhaha.com> <k0k0um$1gk$1@dont-email.me> |
| Date | 2012-08-16 19:18 -0700 |
| Message-ID | <7xsjbmxnum.fsf@ruckus.brouhaha.com> (permalink) |
| Organization | Nightsong/Fort GNOX |
rickman <gnuarm@gmail.com> writes: > No, I can't say I have. But how much of the hard to understand > details are similar between the S40 and the GA144? My impression is "quite a lot" but I don't know for sure. > It is interesting to note that there is no one posting in this group > that has done anything with the chip. Am I the only one even thinking > hard about it? Albert V. did the ParPi program (I guess on the Seaforth40) and has mentioned some issues about a GA144 version. I've put some thought into GA144 programming, but probably less than you have. I'm not a hardware guy, so my ideas about using the chip in an actual circuit are fairly primitive. > I've actually got a new idea that might just be "perfect" for the > GA144..... Around 100 ns per 18 bit multiply or say 500 ns for a > double precision product. That's 288 MMACS or 6000 bins! If the FFT > is broken into bite sized pieces to fit the memory limits the > potential resolution is even higher. How does 288 MMACs compare to a conventional DSP of comparable cost and power consumption? I thought DSP's usually had special features like bit-reversed addressing, single cycle MAC, and zero-overhead loops, to make FFT go very fast. E.g.: http://www.ti.com/lsds/ti/dsp/platform/c5000/device.page might be slightly lower end. Some other ones reachable from here http://www.ti.com/lsds/ti/dsp/home.page go up to 300+ GMACs (multicore), but presumably cost a lot.
Back to comp.lang.forth | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
GA144 Documents rickman <gnuarm@gmail.com> - 2012-08-15 19:44 -0400
Re: GA144 Documents "M. Simon" <msimon6808@yahoo.com> - 2012-08-15 18:11 -0700
Re: GA144 Documents rickman <gnuarm@gmail.com> - 2012-08-16 19:27 -0400
Re: GA144 Documents "M. Simon" <msimon6808@yahoo.com> - 2012-08-15 18:18 -0700
Re: GA144 Documents Paul Rubin <no.email@nospam.invalid> - 2012-08-15 18:21 -0700
Re: GA144 Documents rickman <gnuarm@gmail.com> - 2012-08-16 19:49 -0400
Re: GA144 Documents Paul Rubin <no.email@nospam.invalid> - 2012-08-16 19:18 -0700
Re: GA144 Documents shitstormtrooper11@gmail.com - 2012-08-16 17:36 -0700
Re: GA144 Documents rickman <gnuarm@gmail.com> - 2012-08-17 14:28 -0400
Re: GA144 Documents shitstormtrooper11@gmail.com - 2012-08-17 22:32 -0700
Re: GA144 Documents "M. Simon" <msimon6808@yahoo.com> - 2012-08-23 06:21 -0700
Re: GA144 Documents rickman <gnuarm@gmail.com> - 2012-08-27 18:15 -0400
Re: GA144 Documents Arnold Doray <invalid@invalid.com> - 2012-08-23 13:24 +0000
Re: GA144 Documents shitstormtrooper11@gmail.com - 2012-08-26 14:08 -0700
csiph-web