Geminilake Will Require Intel Audio Firmware Blobs

[torsionbar28]

Except for that pesky little Platform Security Processor, maybe.

Obviously I was speaking as an alternative to intel - which also has the on-die platform security processor, so your point is moot.

eoma68 hardware is looking better and better...aside from Allwinner's poor history re: the GPL

Um, no, eoma68 hardware is terrible and will always be terrible. It's performance isn't even competitive with decade-old AMD or intel hardware. Sorry, but eoma68 is not a replacement for any AMD or intel hardware made in the 21st century.

If you want a modernish non-PSP processor, go with the latest generation AMD Opteron from early 2013. The Opteron 4300 and 6300 series are just as modern and usable as intel's Ivy Bridge, but unlike Ivy Bridge, the AMD parts do not have a PSP. FWIW I'm typing this post on an Opteron 4386.

[ssokolow]

Not only audio compatibility, but the quality of audio representation is much better, so it's a win win situation. I'm also rocking the Scarlett 2i2 for about 5 years now. It used to give me some problems earlier on, but now it always works out of the box.

Amen.

I recommend/gift this $3 (free shipping) chinese USB audio device to all my friends and family because, despite the price and the fact that they pinch pennies by not soldering the reinforcing tabs on the USB connector, it has significantly less noise on the microphone input than every motherboard I've ever compared it to... which makes Skype and competitors much more comfortable.

(Apparently, you can also improve the audio output by cutting a couple of traces and soldering two 47uF capacitors (left and right channel) where the PCB markings indicate. The case isn't sonic-welded, so you can pry apart the "pin in friction-grip cylinder" construction without damaging it and fix both of the penny-pinches if you know how to solder.)

[mulenmar]

Obviously I was speaking as an alternative to intel - which also has the on-die platform security processor, so your point is moot.

Yes, I understand. However, Intel's is slightly more commented about, from what I've seen, so I didn't think it worth mentioning when specifically responding to what you were suggesting.

Um, no, eoma68 hardware is terrible and will always be terrible. It's performance isn't even competitive with decade-old AMD or intel hardware. Sorry, but eoma68 is not a replacement for any AMD or intel hardware made in the 21st century.

It's competitive enough for what I need -- something that can open, view, and edit files, display at a minimum of 720p, not run up the electric bill, and be easily-repurposed or replaced without putting twenty pounds or more of hazardous waste in the dump. It's a standard-format, too, so it could be used industrially to control whatever needs to be. In any case (heheh), not all of us are willing to pay out hundreds or more dollars every few years.

After all, if I need more CPU or GPU horsepower, my Internet-isolated Intel machine can handle that, and is reasonably future-proof for such.

Moreover, it is literally impossible to accomplish the goals of the project with *any* currently-available CPU. They're using the ONE CPU that's "good enough" by ther standards (https://www.crowdsupply.com/eoma68/micro-desktop/updates/picking-a-processor) to make a good-enough product and establish a concrete demand for newly-produced, libre hardware.

If you haven't read it already, and have a few moments, this goes over the common objections to the project: http://blogs.fsfe.org/pboddie/?p=1314

If you want a modernish non-PSP processor, go with the latest generation AMD Opteron from early 2013. The Opteron 4300 and 6300 series are just as modern and usable as intel's Ivy Bridge, but unlike Ivy Bridge, the AMD parts do not have a PSP. FWIW I'm typing this post on an Opteron 4386.

That's actually quite-useful information, thank you very much! I'll keep an eye out.

eoma68 can't look good no matter how bad everything else looks, mostly due to the lack of decent design decisions crippling what is an already crappy SoC to begin with.

Allwinner's GPL violations are irrelevant as they are in their drivers/SDKs, which are not used in the open firmwares anyway.

Which design decisions do you feel are inappropriate to the goals of the project? And do you have any alternatives to suggest to them that aren't covered in the list found here?: https://www.crowdsupply.com/eoma68/micro-desktop/updates/picking-a-processor)

Maybe by the time when/if the next set of designs are to be initiated, something better will be available. Until then, though, like they quote, "beggars can't be choosers".

Allwinner's GPL violations are irrelevant as they are in their drivers/SDKs, which are not used in the open firmwares anyway.

You're right, of course; I just have a dislike of providing companies money when a division of it makes decisions I strongly disagree with, since that all trickles through payroll to the offending department. In this case, it's less important a concern (to me) than encouraging the redevelopment of user-controllable hardware.