[ POST DELETED -- I figured out the problem. PEBKAC]

I think this is a reasonable proposal.  But if you want serious discussion, the place to make it is on the bitcoin-dev email list.  A little bit of reworking into BIP language and you might win some support there.

PoS is a mistake, it will never be more secure or more energy efficient.

But Ethereum is kinda hostile to miners, and the people who have the big holdings of ethereum want to get the mining rewards for themselves so PoS is always pitched as "better".

Nobody answered my question.

With these 4k PSU setups, its really nice to support so many GPUS, but it looks like you would have to disconnect a large number of cables to take a rig out of service.  I'm thinking about building a custom setup with 2 Mother boards, 16GPUs, and one FINKSY 2880 PSU all in one custom case.  This way the connection to that case is just power and network.

@yun999 -- I'm curious about how you use the 4k PSU bundles with open rigs.  It seems one PSU would power 4 or more rigs (at 6 cards each).  So when you have to take one rig down, what do you do?  Disconnect the 9 power cables going to the cards and risers?  Or are you rigs set up so you do maintenance in place?  I have to move mine to replace a card.  Right now the ATX PSU goes with it, so to do maintenance all I have to do is disconnect the 120v AC and the network cable... and its a unit.

I am liking the 4k PSU bundle, just not sure how to deal with it.  Or maybe I should build a massive open rig with multiple motherboards and GPUs and treat it all as one unit. LOL!.

Good stuff!

It is actually the bitcoin of privacy-- it has the mind share and the technical chops to deliver a solid product.  It has the development team that no other coin can match, except Bitcoin Core.  Etherreum's development team isn't as good as the Z.cash one.

It may be having a run up because of a whale, or it may just be speculation-- but the truth is, everything is speculation right now, even bitcoin.  Actual day to day use of these currencies far lags their price rises.

There are no alts that do the Z.cash job as well as Z.cash, except its clones, and thy all suffer the same way litecoin suffers in bitcoins shadow.

I don't know what you mean about "special interest centralization" but having a good development team is a good thing-- it means keeping up with bitcoin core, for instance (look at how bitcoin unlimited is doing such a poor job of that, while z.cash is porting cores improvements over reasonably quickly.)

Have gotten 7 GPUs working with the ASUS Z270-A PRIME.

Here's what I found:

-- Had to flash the BIOS.  Board shipped with 5.0.0 and I updated it to the latest (7.0.0 I think)-- this is necessary to get the "Above 4g decoding" option. Which appears on the boot menu in the bios.
-- PCIe switched to Gen2
-- Didn't make the other changes.

WARNING WARNING WARNING

When I flashed the BIOS the onboard network hardware stopped working with linux. this caused a bunch of problems.

I had to run this utility to refresh the network drivers ROM to make things work:
https://downloadcenter.intel.com/download/19186

Current setup for this board-
-- Ubuntu 16.04 server
-- 7x NVIDIA 1070 8GB Cards

Working fine. (And about %10 more hash under linux than under Windows.)

I've been looking into the state of mining, particularly GPU mining to see if there are ways to optimize it and democratize it.

I'm not a fan of the open air / riser oriented setup, and I admire the work adapting server power supplies to support miners.

So this thread is for sharing my research and discussion.

And occasionally interesting articles:

https://www.nextplatform.com/2017/03/17/open-hardware-pushes-gpu-computing-envelope/?adbsc=social_20170320_70931886&adbid=843906722003533825&adbpl=tw&adbpr=92598387

I understand to some extent the struggle to be coherent in the face of a lack of cognitive function-- for me, I am often sleep deprived.

I've been working in distributed systems a long time, and on an alt coin myself.  So I'm speaking from first hand experience here.


I read the first dozen or so pages of that paper.  It is incoherent and unstructured and appears to be written by amateurs. It's really easy to come up with features and ideas and see "problems" with bitcoin-- and in their case they don't seem to understand why bitcoin is the way it is in the first place.

Distributed systems are genuinely hard.  What Satoshi did was genuinely new and that's why it has had such an impact.

People seem to think that it is easy because he made it look easy.  It isn't.  And there is also a huge difference between coming up with a bunch of ideas -- like that paper has-- and actually producing a secure, scalable, reliable distributed system.

Put another way, every feature means more work and more risk-- every feature reduces security, scalability, reliability and coherence in the distributed system.

And that's still at the theoretical level.  Actually putting things into production is another matter. When you're writing code you have inadvertent bugs to contend with as well.

Cloning bitcoin and making a few small key changes is a reasonable approach-- because a lot of work went into bitcoin.

One single great idea, or innovation on Bitcoin is enough to make a dent.  IF you have a way to avoid the "no power to mine for months at a time" problem, that alone would be worth implementing.

All the other ideas can be done later, on other alts, or whatever.

I think that's what I'm trying to say.  Aim low, and achieve it, before aiming high.  It's harder than it looks.  ;-)

And I wish you the best-- please don't take any of this as negative.  

If you are writing something new, however, from scratch, with several key changes it will take you a long time.  As an individual developer, at least 3-4 years.

I have been following the ZEC development closely.  Their intention was to be CPU mineable with GPUs getting maybe 4X speedup.  Of course it has not turned out that way.

But one of their policies that they stated was the intention to adjust the PoW parameters if necessary if a shortcut that was too strong was found in the algorithm.

They intend to keep ASICs at bay.  I think this is totally possible.

One way to do this would be to increase the memory requirements.  Right now the ZEC PoW targets 1GB but due to optimizations is only using about 600MB on the GPU... they could increase this to 3GB relatively easily, and make ASICs that much harder (and obsolete a whole set of them in one fell swoop).

Equihash is memory hard-- meaning high performance memory will be the path to the best results.

It remains to be seen whose HBM implementation will be the winner this year-- NVIDIA & AMD are both working on HBM cards for later this year.

Have gotten 7 GPUs working with the ASUS Z270-A PRIME.

Here's what I found:

-- Had to flash the BIOS.  Board shipped with 5.0.0 and I updated it to the latest (7.0.0 I think)-- this is necessary to get the "Above 4g decoding" option. Which appears on the boot menu in the bios.
-- PCIe switched to Gen2
-- Didn't make the other changes.

WARNING WARNING WARNING

When I flashed the BIOS the onboard network hardware stopped working with linux. this caused a bunch of problems.

I had to run this utility to refresh the network drivers ROM to make things work:
https://downloadcenter.intel.com/download/19186

Current setup for this board-
-- Ubuntu 16.04 server
-- 7x NVIDIA 1070 8GB Cards

Working fine. (And about %10 more hash under linux than under Windows.)

I'm consistently getting 420 sols at 150watts on 1070s with EWBF (as reported by nvidia-smi)

When I did my spreadsheet it looked like nvidia was the better deal for ROI on ZEC. Don't have any RADEONs to compare to, though, right now.

So, I've been on the hunt for a good MB to do 6+ GPUs with.  Awhile back on a hunch I bought the ASUS Z270A PRIME -- 7 PCIe slots and two m.2 slots - I thought maybe I could get 9 GPUs on it.  But I was restricted to 4 with the bios that shipped with it.

At this point, however, I have gotten 7 GPUs on it-- the max I have space for right now.

Sharing the details here in case it is useful for someone else:

Here's what I found:

-- Had to flash the BIOS.  Board shipped with 5.0.0 and I updated it to the latest (7.0.0 I think)-- this is necessary to get the "Above 4g decoding" option. Which appears on the boot menu in the bios.
-- PCIe switched to Gen2
-- Didn't make other changes.

WARNING WARNING WARNING

When I flashed the BIOS the onboard network hardware stopped working with linux. this caused a bunch of problems.

I had to run this utility to refresh the network drivers ROM to make things work:
https://downloadcenter.intel.com/download/19186

Current setup for this board-
-- Ubuntu 16.04 server
-- 7x NVIDIA 1070 8GB Cards

Working fine. (And about %10 more hash under linux than under Windows.)

Ok, I have replicated Andrews experience.

Here's what I found:

-- Had to flash the BIOS.  Board shipped with 5.0.0 and I updated it to the latest (7.0.0 I think)-- this is necessary to get the "Above 4g decoding" option. Which appears on the boot menu in the bios.
-- PCIe switched to Gen2
-- Didn't make the other changes.

WARNING WARNING WARNING

When I flashed the BIOS the onboard network hardware stopped working with linux. this caused a bunch of problems.

I had to run this utility to refresh the network drivers ROM to make things work:
https://downloadcenter.intel.com/download/19186

Current setup for this board-
-- Ubuntu 16.04 server
-- 7x NVIDIA 1070 8GB Cards

Working fine. (And about %10 more hash under linux than under Windows.)

It's always good to re-read the Satoshi paper, as it is not only laying out revolutionary stuff, it is a benchmark for succinctness.  So much about early bitcoin was not in that paper.

Writing a white paper can turn into a kitchen sink of a wishlist document.

Hopefully you're getting feedback from people on the paper, and you are doing some coding to test the implementation ability of some of your ideas.

I've been looking at the coins coming out and doing ICOs and reading their papers and it's really obvious how many will take forever to ever work (hell, I argue ethereum still does't work like it was supposed to. They still haven't solved fundamental problems.)

Finally, I think there are two good paths here-- if you want to play devils advocate and argue for better choices, go for it.  But if you want to be secret and write your paper in secret before sharing it with people, then probably best to not talk about your idea here-- don't let ego take you away.

If you are building what you say, then your time is better spent building it than talking to naysayers here.  IF you aren't building it, that's fine, share your ideas.

I'd be happy to review your white paper if you like, I can send you my bonafides by pm.

Here's my prediction:

Within 2 years, ATH for bitcoin will exceed $7,500.

Within 2 years, Litecoin will trade below $10.

Why?  Litecoin is an also ran.  It offers nothing of value over bitcoin.  Sure litecoin is a well run coin, gotta give them props for that, but it's not innovating... and the bitcoin space is already covered by bitcoin.

Ethereum, with all its problems, is at least trying to offer something different.

All lite coin offered was non-asic mining, and that has gone away.

Here's the route I went:
https://www.amazon.com/Elgato-Energy-HomeKit-Technology-Bluetooth/dp/B01M3PK5WH/ref=sr_1_fkmr1_1?ie=UTF8&qid=1489902199&sr=8-1-fkmr1&keywords=eve+energy+220v

So on my iPhone I can see how much each rig is using, how much in $ it has cost since the last time it was reset, can remotely turn them on and off, etc.

I can do all of this from anywhere in the world.

Supports 240v.

Gonna be hard to drive 6 GPUs from 5 PCIe slots.

Please only recommend boards you know work.

For instance I have a board with 7 PCIe slots and 2 m.2 slots that theoretically could drive 9 GPUs.

Realistically it can only drive 4.

What BIOS settings?  How much RAM in your GPUs?

I have same MB and can only run 4 GPUs