<br><br><div><span class="gmail_quote">On 12/5/06, <b class="gmail_sendername">Gabriel Ambuehl</b> <<a href="mailto:gabriel_ambuehl@buz.ch">gabriel_ambuehl@buz.ch</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On Tuesday 05 December 2006 23:04, Markus Stehr wrote:<br>> Everything would be possible all what would be need is some lowcost and<br>> lowcurrency FPGA.<br>><br>Wouldnt a FPGA without very generic IO features be rather constrained in its
<br>use? Furthermore, most of the things you mention (minus the DECT thing which<br>likely also needs RF infrastructure) can be done in software no?<br><br><br>_______________________________________________<br>OpenMoko community mailing list
<br><a href="mailto:community@lists.openmoko.org">community@lists.openmoko.org</a><br><a href="http://lists.openmoko.org/cgi-bin/mailman/listinfo/community">http://lists.openmoko.org/cgi-bin/mailman/listinfo/community</a>
<br><br><br><br></blockquote></div>Yeah, you can do pretty much everything in software, but you get some acceleration doing it on FPGA space(you count time in gate delays, rather than software steps), not as much as running straight silicon, but enough, and you get to offload the work from the processor.
<br><br>The down side is you have to connect the FPGA into all the peripherals and multiplex with the CPU, or connect using DMA either way, there's a marked increase in board complexity, probably requiring a complete re-design of the board to get it in in time for gen 1. I'd love to see this integrated for the second generation, however.
<br clear="all"><br>-- <br>--Jeff<br>What DO you call whitewater when you live in the desert?