Genuinely interested, though I have a few thoughts.
To understand the value of pre-built pre-configured Linux image, one needs to understand how much work is required to get one up and running. I would like to suggest that Eliovp do a simple write-up explaining the difficulties he encountered in getting a Linux OS up and running. A tutorial would be ideal, but it's probably too ambitious to ask for one during this donation round. Moreover, since you're doing a donation round to build a custom Linux image for the Pandaminer, I'd like to suggest that you may as well go the extra mile and build it to a level which turns the Pandaminer into a plug & play appliance. It should gather more steem.
As for the ROMs, I understand VDCC offset / undervolt / overclock stability is very much dependent on the ASIC quality of the 480 chips in the miner.
Questions to Eliovp:
1. How can you guarantee that your ROMs will be stable on any Pandaminer, without knowing what the actual ASIC quality is on them?
2. How do you plan to get the ROMs updated on the GPUs? Automatically by means of a script that runs the first time your Linux OS is loaded? Or manually by providing a tutorial to people who get the full bundle?
3. What fallback options do you have in mind in case of bricked GPUs?
4. Do you also plan to built a miner management interface (maybe web-based) and include it in your Linux build?
Hey TheRider,
Sure thing. Obviously i'm not going to write a tutorial here or write down the steps i had to take to be able to have my setup like i have it right now. It's like you said yourself, the purpose of those thread would be kind of pointless.
You said to go the extra mile and make this plug & play. Well, i already have. It is plug and play. As i've mentioned before, what i've done would be ideal for big farms, they could just place them, plug them in and sit back and relax.
Romwise, yes, you have a point, but that's the thing, assuming the gpu's they deliver with the PM are all the same, i added a global vddc offset, so in other words, the voltage controller will still calculate what the gpu needs "volt wise". The only difference is that it starts calculating from a lower starting point. I didn't drop the power to hard just to be sure that it remains stable (-31mV).
And so far so good, after a week, still rock solid.
1. I don't need to know the asic quality, i'm pretty sure they will remain stable. And if you do happen to have one that can't handle it (which i highly doubt) just send me a pm with the original rom (so i know you have one) and i will change it.
2. Good question. Second question here will probably be the answer. I can automate this without the chance of bricking them. Trust me, i can
.
3. The only way "in this case" to brick them is if someone would flash a faulthy rom or completely different one. If you stick with the one i offer, this won't happen.
4. Yes.
Sorry, this may be slightly more tangential for this thread....
The RX 500 series will be out later this half, most likely May or June. The information I read is that the efficiency is supposed to be quite good. I'm curious to see the price and throughput of these "next gen" cards. This next card release gives me pause on buying a Panda based off of RX400s. I understand that HBM2-based Vega cards will be prolly 250/300W cards.
New hardware will always happen, i still own a bunch of 290's from back in 2014, they're still running rock solid.
But i get the point. What could and would be interesting in this story is perhaps the fact that these mx gpu's might be replacable by new gen mx cards. But i'm not very sure about that though. Would be ideal!
Greetings.