Re: Would green arrays produce something with a web browser? like a cheap appliance?

From	Paul Rubin <no.email@nospam.invalid>
Newsgroups	comp.lang.forth
Subject	Re: Would green arrays produce something with a web browser? like a cheap appliance?
References	(15 earlier) <jh1jsu$bh1$1@online.de> <60eb6055-6ea0-40b4-addd-e8028a647114@t5g2000yqk.googlegroups.com> <jh2ub4$npj$3@dont-email.me> <jh3gi3$4g4$1@online.de> <95a854dd-69c4-4cf9-8966-eadbaa57733c@l14g2000vbe.googlegroups.com>
Date	2012-02-16 13:45 -0800
Message-ID	<7xvcn6tpub.fsf@ruckus.brouhaha.com> (permalink)
Organization	Nightsong/Fort GNOX

Show all headers | View raw

rickman <gnuarm@gmail.com> writes:
> Here is a "niche" where very little working memory is required,
> Software Defined Radio (SDR).  The basic receiver chain can be done on
> a GA144 entirely from the IF to audio by using the on chip ADCs and
> DACs.  No external memory would be needed other than the Flash for
> program storage during boot.  But then I guess that is just a tiny
> market... ;^)

I'm having trouble believing that, for spread spectrum schemes where the
spreading function is complicated, or you want to handle wideband
signals (e.g. GNU SDR can demodulate multiple FM broadcasts from a
single A/D simultaneously, needing sample rate in the 10's or 100's of
MHz if I understand correctly).

> I won't argue with those words because they also describe GPUs and
> DSPs, special purpose devices deliberately shifted to a particular
> corner.. and very large corners at that.

Bernd made the point that applications for GPU's and DSP's were well
known before GPU's and DSP's were actually developed, and the
applications drove the development.  Is there a known application that
drove the development of the GA144?

> I maintain that there are plenty of useful apps that can be done using
> such a device and you seem to insist that they are very few.  

Well, few and plenty are relative terms... try to think of applications
for DSP's or microcontrollers and they come to mind far more rapidly
than they do for the GA144, it seems to me.

> A GA32 at $5 would be much more suitable for my needs.  I suspect the
> family won't really take off until it is updated for some market
> realities and moved to a smaller process enabling perhaps 1024 nodes.

I haven't heard of anyone figuring out to do with even 144 nodes (given
the GA144's memory and i/o constraints) so I have a hard time seeing a
GA1024 as doing much good without a lot of other new capabilities.  Even
the GA144 (it seems to me) doesn't have enough interconnect between its
nodes, which could be added fairly easily, according to an earlier clf
discussion with Jeff Fox.  I do like the idea of a $5 GA32 and it seems
to me that it could have similar total i/o capacity to the GA144,
bringing it much closer to Bernd's suggested i/o to mips ratio.

> I'd also like to see more RAM, but I am going to explore the device as
> is and see where it goes from here.  Hopefully not the dust bin of
> history.

I'd like to see an FPGA with some ram blocks, some DSP slices, and some
GA controller nodes as hard macros.  I do like the spirit of the GA
folks, and I wish them success one way or the other.  I'm still
scratching my head on the technical side.

Back to comp.lang.forth | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

csiph-web