<div dir="ltr">So from my understanding Nyuzi is for gpgpu while MIAOU is for actual rendering ? What kind of performance is reasonable to expect from this ? Would it reach something like the mali 400 mp2 ? </div><div class="gmail_extra"><br><div class="gmail_quote">2017-02-17 19:03 GMT+02:00 Luke Kenneth Casson Leighton <span dir="ltr"><<a href="mailto:lkcl@lkcl.net" target="_blank">lkcl@lkcl.net</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">On Fri, Feb 17, 2017 at 4:52 PM, Bill Kontos <<a href="mailto:vkontogpls@gmail.com">vkontogpls@gmail.com</a>> wrote:<br>
> Is that the master plan you with Allwinner you mentioned somewhere else ?<br>
<br>
</span> no it isn't.<br>
<span class=""><br>
> I'm very interested with the idea. The mere idea that you managed to get a<br>
> new SoC going on on a reasonable node despite the cost intrigues me a lot.<br>
<br>
</span> i haven't - i'm putting in proposals and finding candidate people to work with.<br>
<span class=""><br>
> So what about all the other ip blocks involved in the soc ? Realistically if<br>
> such a small form factor wants to be successful it needs to have 3d.<br>
<br>
</span> Nyuzi and MIAOU. GPLGPU is out because... well... its designer<br>
unfortunately doesn't understand that you can't modify the GPL by<br>
adding a clause "this is GPL except if you want to use it for<br>
commercial purposes then, well, sorry, but you can't", that's *not*<br>
the GPL, it's a proprietary (closed source) license. whoops.<br>
<br>
so, Nyuzi it is, along with MIAOU as a *separate* engine which will<br>
do OpenCL and will need to be made use of separately (in software).<br>
also great for parallel processing tasks as-is.<br>
<br>
there will need to be a *lot* of software development - even for the<br>
hardware. the lowrisc designers aren't *actually* developing any<br>
peripherals: they're putting in what they call "minion cores" which<br>
basically do bit-banging of various GPIO pins under their control, in<br>
effect *becoming* peripherals. (if it's dedicated bit-banging, it's<br>
not really bit-banging, is it?) the nice thing is: you can literally<br>
implement any protocol you care to.<br>
<br>
i want to have a word with them to make sure that there's some<br>
differential pairs connected to the minion cores, with variable power<br>
domains. that would make it *potentially* possible for people to<br>
either use them as an open high-speed inter-connect, or to implement<br>
various high-speed peripheral buses such as PCIe, LVDS, MIPI, eDP and<br>
HDMI - all depending on whether the minion cores can handle it and are<br>
set up to do DMA or not: just have to see.<br>
<br>
so about the only thing that would need to be licensed (at this<br>
stage) would be DDR3 RAM interfaces. everything else is covered from<br>
<a href="http://opencores.org" rel="noreferrer" target="_blank">opencores.org</a>, including an LCD/VGA controller, USB2, and many others.<br>
<br>
VP8 and VP9 are available from google with no royalties if you are<br>
going to production silicon. MP4 can easily be obtained.<br>
<div class="HOEnZb"><div class="h5"><br>
l.<br>
<br>
______________________________<wbr>_________________<br>
arm-netbook mailing list <a href="mailto:arm-netbook@lists.phcomp.co.uk">arm-netbook@lists.phcomp.co.uk</a><br>
<a href="http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook" rel="noreferrer" target="_blank">http://lists.phcomp.co.uk/<wbr>mailman/listinfo/arm-netbook</a><br>
Send large attachments to <a href="mailto:arm-netbook@files.phcomp.co.uk">arm-netbook@files.phcomp.co.uk</a></div></div></blockquote></div><br></div>