I did not really _personally_ touch the security aspects of an OpenHardware SoC, but truth is there is an evergrowing business / enterprise market out there that I am sure would like some OPEN assurances that their security will not be hamstrung by some invisible hidden back door.
Personally an Open GPU is a must as far as im concerned, but an Open SoC with that open GPU inside it would be awesome. Gary On Sat, Dec 8, 2012 at 3:32 PM, <[email protected]> wrote: > Le 2012-12-09 00:13, Ing. Daniel Rozsnyó a écrit : > >> I can not keep with your plans up. Does any of you have $1G+ on your >> account to run such a project? >> >> Currently, you can get a well documented CPU (let's stay at ARM), >> even with several cores, there are no hidden things and you can create >> the ultimate compiler for it. There is no need to have an opensource >> cpu - nobody can benefit from having the possibility to improve on it. >> > > I'm glad you don't need an open CPU. > But following your logic, there is no need of an open GPU either. > > If Intel would hypotetically release its "sources" for their CPUs - >> can you manage to make improvements on it and release you own flavor >> if their cores? Potentionally not. >> > > believe it or not, even Intel uses 3rd party cores : an ARC (and not ARM) > core is used and tuned to sit idle in the northbridge (and open invisible > ports in your network but i'll leave the security aspect aside). > > This means that whenever you stray away a little from the > mainstream/standard, > like you don't depend on legacy code, customisation is the king and you'll > adapt and tune whatever you can. > > CPU's are today not a secret thing, the manufacturer wants you to use >> it to the last feature. So it is a waste of effort today to replace >> it. >> > > Tell this to the many providers of configurable cores :-) > > Yet I agree that there is no secret to CPU, or GPU : > it requires tons of work to make them work well. > > What I would suggest, is to recreate just those peripherals which got >> the most issues with the opensource world (graphics, audio, hardware >> codecs, etc), then focus on the ones which got severe marketing >> limitations (e.g. professional audio - 192k/24 is not sold for the >> adequate price of the components, being those the best they can be, >> there is always some 10x multiplier in price, same goes for >> consumer-broadcast video). >> > > so you're also in the cost-cutting mindset, which is a totally different > issue. Cost is usually reduced whenever there is a standard, an interface. > But for audio, as long as you have SPDIF in and out (or ADAT, or...) > click-free, you're done. > > Once you own the whole peripheral business, you can replace the >> processor too. >> > > so let's see if a virtualised shader can be > successfully developped in an open source environment :-) > > > regards, > >> Daniel >> > yg > > ______________________________**_________________ > Open-graphics mailing list > [email protected] > http://lists.duskglow.com/**mailman/listinfo/open-graphics<http://lists.duskglow.com/mailman/listinfo/open-graphics> > List service provided by Duskglow Consulting, LLC (www.duskglow.com)
_______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
