<div dir="auto">If you can work with the slightly odd floating point format in single precision GLES shader language then the shader engines in the Raspberry Pi (1, 2 & 3) will give you about 24 GFLOPS.<div dir="auto"><br></div><div dir="auto">Neil</div><div dir="auto"><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On 13 Feb 2017 20:38, "cheater00 cheater00" <<a href="mailto:cheater00@gmail.com">cheater00@gmail.com</a>> wrote:<br type="attribution"><blockquote class="quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="quoted-text">On Mon, Feb 13, 2017 at 8:55 PM, <<a href="mailto:paula@synth.net">paula@synth.net</a>> wrote:<br>
> I looked at starting the Chameleon II project with Terry Shultz, but sadly<br>
> this was around the time freescale started pulling out of the DSP world.<br>
><br>
> Your choices for "DSP" now are to look at something like the Analog Devices<br>
> Blackfin/Sharc family, look at FPGA (who wants to roll their own DSP) or a<br>
> fast ARM Processor (like the Axoloti).<br>
><br>
</div><div class="quoted-text">> I guess the problem is that ultimately it comes down to the market appetite<br>
> for such products. The Chameleon was great, but never really got the market<br>
> it needed to survive. the Axoloti looks fab, but I doubt it's being made in<br>
> serious volume (>1000 per year).<br>
><br>
</div><div class="quoted-text">> What is nice is to see a resurgence of love for DSP and DSP programming :)<br>
><br>
> Paula<br>
<br>
</div>I think just the first three uses are going to be good enough to find<br>
enough market share between them to warrant a board run. Many people<br>
already have lots of stomp boxes and they will pay $2500 to get an<br>
Axe-FX and use it only for cab simulation... and then they only get a<br>
few msec worth of impulse response time budget... that's fairly weak,<br>
and it's not difficult to beat.<br>
<br>
But either way if we can agree upon some sort of platform we can at<br>
least go and buy similar dev boards and start plugging away at it. I<br>
think the TI TMS320C is the best platform because the upper end is<br>
very high and you can start with very cheap chips; the most powerful<br>
board from them is $600, and it's an amazing powerhouse.<br>
<br>
If it turns out we have a nice platform going that we could share with<br>
the world we can think about spinning up some boards.<br>
<br>
As Veronica points out the most difficult part is to set up a company<br>
and stay afloat. So don't. Develop for dev boards, we can afford them,<br>
and it's for us, so we don't care about mass market appeal. Rather<br>
than set up a company that'll stock this stuff, go the Arduino route,<br>
and create a generic platform anyone can "make" with. A board like<br>
this could have appeal for computer vision, robots, quad copters, sdr,<br>
etc. This should keep enough people interested who would pay for e.g.<br>
group buys.<br>
<br>
The #1 objective should be that anyone who hasn't done any embedded<br>
programming before should be able to get into this within days,<br>
especially without hardware. It would be a great idea to have<br>
something like a qemu based emulator, even if it doesn't work at full<br>
speed. An even better idea is something that simply runs in the<br>
browser. This way anyone who wants to keep the platform alive, can.<br>
<br>
It would be a good idea to keep this future minded as 5 years from now<br>
nVidia will have entered the dsp business full-scale and their chips<br>
will kill anything that's out there. I spoke to John Carmack and he<br>
said they are certainly going to do this. Tim Sweeney confirmed that<br>
the GPUs as they are now are 10x more powerful per dollar as the best<br>
of the best DSPs in my survey earlier on in this thread. TI and AD<br>
have no contender in the GPU based dsp space so they will lose out in<br>
the end; nVidia will dominate the market for automotive applications.<br>
They've already started building self-driving-car embedded modules<br>
with their GPUs on them, and eventually they'll do their own silicon.<br>
So it would possibly be a good idea to keep the algorithms modular, so<br>
that the concrete implementation of anything that needs to be in<br>
assembler can be replaced in the future. This would also prevent<br>
"platform death".<br>
<div class="elided-text">______________________________<wbr>_________________<br>
Synth-diy mailing list<br>
<a href="mailto:Synth-diy@synth-diy.org">Synth-diy@synth-diy.org</a><br>
<a href="http://synth-diy.org/mailman/listinfo/synth-diy" rel="noreferrer" target="_blank">http://synth-diy.org/mailman/<wbr>listinfo/synth-diy</a><br>
</div></blockquote></div><br></div>