ETH devs made that clear a LONG time ago when they first announced the move to Proof of Stake.

It shouldn't be new news....

I was misremembering details, but it's not "post count" that determines activity as such, it's "consistent posting" that does.

https://bitcointalk.org/index.php?topic=2658590.0

The cost alone makes that a very unlikely case.
Bitmain would have been LOSING money at the $800 level on the thing.



Why would equihash coins fork when there aren't even any legit RUMORS of an ASIC miner for equihash?
Have a fork READY TO GO perhaps, but forking before miner is a waste of time effort and disruption.

The non-ETH coins that use ethash probably SHOULD fork sometime fairly soon if "anti-ASIC" is their view, since they'll be impacted a lot more over time than ETH itself will be - but they're going to get impacted a lot more (as will ALL GPU-mineable coins) when ETH moves to full POS.


I also wish people would stop it with the huge font stuff and QUOTING that huge font stuff.
I've already put 2 people on ignore over that rudeness.

Not as long as Nvidia keeps spouting those lies they like trying to blame miners for their shortage when the REAL issue was decisions on the part of Nvidia ITSELF (that were bad in 20/20 hindsight but looked reasonable ahead of time).

AMD has a fair case for blaming miners - they're a LOT more impacted - but oddly enough they prefer to blame "ram shortages" for not being able to ramp up to meet demand.

Desiccants are only viable in small ENCLOSED spaces, they do not soak much moisture from the air at all per unit of weight/volume.
Dehumidifiers do NOT heat up the air, every "mechanical" dehumidifier setup I've EVER seen was a specialized air conditioner and the air they put out of the "dehumidified" side was COOL to COLD.
There is NO condensation risk on the air out of the dehumidifier, there's pretty close to no moisture in there that CAN condense.

It won't take "dozens" of them to reduce humidity 1-2%, unless you are running MILLIONS of CFM airflow (hundreds of thousands of M^3/minute) and have 80%+ humidity air input - though it would get complicated a bit by the fact they would reduce your intake air temp some, they'd still be pulling SOME of the water out of the air so even if the RH didn't drop a lot the moisture content of the air WOULD which should reduce the rust issue.

---

Moderator's note: This post was edited by frodocooper to remove multiple nested quotes.

If you're in the US, use US CODE type outlets - which means NEMA 6 of some sort for a 234V circuit (14 if you want the flexbility of 234V OR 117V on a particular circuit) or NEMA 5 for 117V.

British type plugs/sockets would work and be safe if used within their raitings, but would be an automatic FAIL on any sort of inspection as they don't meet US code requirements - and NO sane electrician will install them in the US.

The point to a profit - even small - is that it's money you would not have HAD otherwise.

Also, the OP specifically stated "hobby miner".

You DO know that ETC is an ethash coin, and therefore the E3 WILL mine it?

Does it work with BIOS modded cards?
I've seen conflicting info about which Adrenalin drivers do so and which don't.

That's also why Claymore's recommendation of 16.12.x made ZERO sense, as 16.11.x and up DISABLED support for BIOS modded cards while 16.9.x and 16.10.x DID work with them.

That poll was closed before your post about it was made, and the ETH dev team has already stated they will not waste time forking ETH over the E3 - but do reserve the option to do so if something more powerful shows up.

You HAVE to move to the AMDGPU drivers to use Ubuntu 16.04 or later, as the older fglrx drivers flat out are not supported - they actually got "unsupported" in 14.04.04 but you can "roll back" the XServer in THAT version to make them work again.
You also have to install the ROCm stack for OpenCL support, but oddly enough that seems to work on ANYTHING out of AMD that is GCN based - even the GCN-based APUs that the AMDGPU drivers themselves do NOT support like my A10-7890k and A10-7860 models.
The ROCm stack appears to offer ZERO support for pre-GCN cards at all, so forget about HD 6xxx and older or the early low-7xxx series cards (anything lower than the HD 7750) or the iGPU in AMD APUs that were pre-GCN like the A10-5700.

I'm actually not certain about the status on 14.10 15.04 or 15.10 but as those releases are no longer supported at all by Canonical there's no point in even looking at them.

There is pretty much ZERO probability of the E3 achieving payback of $1800 before ETH goes full Proof of Stake if the current timeline holds - it might have a faint prayer if ETH price DOUBLES fairly soon, but only due to it taking time for folks to "ramp up the hashrate" to match the price jump.
It was marginal iffy at the original batch price of $800 with VERY LOW electric cost.

There is also the outstanding question of "is it going to be reliable".

IMO anyone that pays $1800 for an E3 is going to LOSE money on the deal unless they have free electric AND they get a unit that runs for 2+ years without issues - and that is assuming that the other ethash coins don't fork and that coin prices stop SLIDING downward and even THEN it's still iffy if they'll ever earn their money back.

At $800 they were marginal but had a chance to achieve 100% ROI before ETH goes full Proof of Stake.
Batch 2 however went to $1800, which is INSANE.

My guess on that nanopool miner is that it's a large Chinese farm running 6-card 1060 rigs (or using the mining-specific version) - and having close hashrates on each rig just means they have a standardized setup they're sticking with (which makes LOTS OF SENSE on a large farm).


ETH devs have already announced they are NOT going to fork ETH over the E3 but reserve the option to fork if something more powerful shows up.

Which part of "ETH devs have already ruled out a fork due to the E3" did you not see?

Like those of the rest of us that have ALSO crunched the numbers, they realize the E3 is not a significant short-term threat to GPU mining and probably won't have a major impact on ETH profitability by the time ETH goes full Proof of Stake.



There are a few "big cap" coins left that do not have an ASIC available yet are mineable - Zcash is probably the biggest-cap such coin at this point - but even with no ASIC those coins are going to take a serious profitability hit when ETH mining goes "poof, we're full POS now" and the ballpark 75-80% of all GPUs currently doing cryptocoin mining that are now on ETH are forced to look for something else to do.

It won't be obsolete in a couple of months - even when ETH goes proof of stake (PERHAPS by the end of the year, but I'd guess it's more likely to happen 1H 2019 for FULL POS) there will still be the other ethash coins for it to eek out a pittance income on.

My guess is that they misentered the price when they put the second batch up, and will correct it Monday - $1800 has pretty close to ZERO chance of ever paying for itself unless ETH pricing (or pricing on one of the other ethhash coins) jumps in a BIG way sometime in the next month or two.

Monero is the only CN coin that has switched so far - ETN had ballpark half the network hashrate of Monero and has ANNOUNCED they plan to fork to change algo, but hasn't done so yet.

Why would ETH miners switch, unless Monero profitability goes WAY up quickly?

The announcement of Bitmain's cryptonite miner seems to have been driven by Baikal shipping THEIR cryptonite miner.
The XMR fork wasn't announced or planned 'till the Bitmain miner was announced - your timeframe is WAY off there.

Also, it's not just Monero that is forking - ETN has also announced they will be forking (probably piggybacking on the Monero fork code and algorithm) soon, and I think I've seen commentary from some of the "small fry" cryptonite coin devs that they might be forking as well.


Hint - Bitmain doesn't sell their miners for pennies, they sell them then turn the money around to make MORE miners to sell - and make quite a bit MORE that way than via mining which takes MONTHS or YEARS to pay a machine off.

There have been times when Bitmain ended up selling a miner for more than it ever made back, though USUALLY if a miner is profitable long enough it will EVENTUALLY make more than you paid for it (IF you have very cheap electric) - but in that timeframe Bitmain has turned that investment around at least twice and more likely 3-4 TIMES via new batches of that miner model and made a LOT more than they would have made from mining on the first batch for the same amount of cash invested -and ZERO risk of losing money as well.


I don't like Bitmain, and they've made mistakes at times, but I don't call their business model STUPID for a reason.

That would only represent a few hundred thousand GPUs sold in that timeframe - which WOULD NOT ACCOUNT FOR THE SEVERE SHORTAGE OF AMD CARDS prior to the last couple weeks.

DO try again and not be rude and insulting for no reason.

The E3 is NOT going to "spike" the total network hashrate for Ethereum or the other ethash coins.

It will take about 2 MILLION of those units built and operating JUST TO MATCH THE EXISTING HASHRATE.

They aren't shipping 'till mid-July timeframe.

In the almost 2 YEARS since Bitmain introduced the S9 (which was a LOT more efficient than anything else available for months, and STILL is the most efficient SHA256 miner that is shipping) they might have sold a little over ONE MILLION of them (total SHA256 network hashrate says up to 2 million, but you also have to factor in the Avalon 721/741 sales, the eBang sales, sales of whatever miners with Bitfury chips have been built, however much hashrate LW.COM has on it's internally-built miners, and at least one "smaller fry" mining operation I know of that is using their own chip).

The E3 will push network hashrate - but it's NOT going to be a "spike", there is too much EXISTING hashrate from mining rigs that are the SAME OR BETTER efficiency than the E3 announced specs.


It's a lot closer to the situation in Scrypt mining when Bitmain introduced the L3 - which STILL hasn't managed to push all "older rigs" out of profitable service if you have cheap enough electric and has a LOT of competition helping it TRY to push A2/Titan/Alcheminers out of profitability.

On Windows, I'd go with either the blockchain driver or something around 17.11 (whatever the first WQHL version is after they added "large page" support to the mainline drivers).

On LINUX, 17.50 seems a lot more stable than the blockchain drivers, but you have to modify one of the driver modules to get ANY support for undervolting or underclocking the core.