Got a quibble with your original post.

 USB 2 is limited to 500 ma at 5 volts, or 2.5 watts.
 You're not going to pull 3 Megahash on a stock USB2 port without going well beyond it's specs.

 USB *3* has a 900 ma limit - 4.5 watts - THAT'S where you can pull enough power out of the connection without exceeding specs to get close to 3 Megahash.

 USB "battery charging" ports are special case upgrades to the USB 3 specs for power distribution, and a lot of "powered hubs" are capable of quite a bit more current per port than the spec requires even if they don't meet the full "battery charging" specs.



 Stick miners have NEVER been intended to compete on a hash/$ basis with a "full" miner, there are way too many overhead costs that are always quite a bit higher per chip.

 "INFO" timestamp gpu# "accepted share" pingtime "(" starestats ")"

 Seems pretty obvious.

 Keep in mind that mining is normally a "small niche market" for GPUs - the last few months have been a rare exception.

 On GAMING (the primary market for higher-end cards like the 1070 and up) there is no comparison between a 1070 and ANY RX Polaris-based card, the 1070 blows them completely out of the water.

 Even the VEGA barely beats the 1070 on gaming, and loses pretty much across the board (narrowly) to the 1080.

EUV isn't an option for 10nm and smaller.

 It is a REQUIREMENT and has been one of the major factors causing delays in the deployment of new smaller nodes of production.

 There is no reason why smaller node chips will require anything more than air cooling.

 Immersion cooling can be nice - but it's only practical on a HUGE SCALE for mining.
 Even most of the MAJOR industrial-level mining farms have been air cooled due to the cost of setting up an immersion cooling setup.

 Even for LARGE SCALE mines, the efficiency might not be worth the up-front costs, though the longer "cycle times" for equipment might change that some now that mining has caught up to semiconductor state-of-the-art.

If the logo has been updated, why does the wallet still have the OLD logo?

 New logo is a major improvement, you can actually figure out what it's supposed to look like as opposed to the old logo that was WAY too convoluted and UNreadable.

Some of us were making money on both GPU and ASIC mining a YEAR ago - long before the recent massive profitability bubble hit.
THAT bubble is almost over, but it wasn't a "burst" so much as a "slow decline" caused more by massive new miner influx than any other factor - and it's not over YET as profitability is STILL higher than a year ago.

 3GB cards still work and will continue to do so for a while.
 As the only CURRENT card model that is worth mining with that has 3GB is the smaller GTX 1060, though, it's not a common option.


 It's entirely possible to set up a dedicated mining machine on 30 GB or less of HD space - a 32 GB USB flash drive is entirely practical - if you use LINUX instead of Windows.
 A SSD in a mining machine is overkill - you do NOT need the speed - and the cost tends to be a lot higher per GB.
 Also, many motherboards shut down one or more PCI-E slots when a NVMe type SSD is used.

There is no point in comparing the hashrates - the algorithms used have a LOT different "amount and type of work needed to generate one hash" that makes such a comparison worthless.

Even the top-end AMD A10 integrated GPUs are not worth trying to mine on (they work, but CRAZY slow, even on the coins they CAN mine at all).
 I did some testing on ZEC mining at one point with a A10-7890k - it managed something like 11 sol/s (HD 7750 has the SAME number of cores at the SAME clock and managed more like 60 due to the MUCH faster GDDR 5 it uses).

 There isn't any point in going with a GPU older than the NVidia 9xx series (exception for the 750ti which was the same Maxwell arch as the 9xx) - way too low performance way too high power usage - and even the 9xx series you had better be getting the card VERY VERY CHEAP to make it worthwhile since they're on the older 28nm tech and much lower efficiency than the current 10xx series on it's 14/16nm tech.

 AMD side is a bit better off - any of the GCN-based cards prior to the RX series (except the Fury) are pretty close on efficiency to each other as AMD reused the SAME GPU (with slightly faster RAM and some BIOS changes in each "generation") in most of those cards for 3 "generations", then the RX series got a lot more efficient with the move to 14/16nm tech.

You are going to have some SERIOUS overheating issues with those GPUs packed in too tight like that.

 R9 290 (and 290x) are still good ETH miners, not so good at ZEC, but they DO eat a lot of power to get to the 30 Mh/sec ETH hashrate ballpark and they need a LOT of airflow to stay semi-cool.

 My triple R9 290 rig has 3 Sapphire reference-type cards like the ones in the pic, and they NEED space between them even with MASSIVE airflow on the case they are in.

Waste of time.

 Of the 5 DGB algorithms, 3 are ASIC mined and the other 2 are GPU mined - CPUs can't compete or ever get into the ballpark.

Power in the USA is commonly distributed to houses and small businesses as a 220 volt "split phase" aka 220 volt Center Tapped feed,

 This is mostly split down to 110 volt outlets, with a few 220 volt circuits used for higher-power stuff like electric hot water heaters, electric driers, high-capacity "mechanical" air conditioner units, electric ranges, some electric heating, and some "workshop" outlets for stuff like welders.

 For US wiring, 12/2 (20 amp) or 10/2 (30 amp) with ground with one or more NEMA 6 outlets are the standard on dedicated 220 volt circuits (commonly one outlet as these lines are generally intended for HIGH power usage items but there is no reason you can't wire 2 or more outlets to a single circuit).
 For US wiring where both 110 and 220 is needed (many electric driers and ranges use 110 for their "timer/clock" type circuitry) a NEMA 10 outlet and the use of AT LEAST 12/3 (10/3 up to 6/3 on higher power circuits as needed for the current capacity) with ground wiring is the norm.

 Specific to OP - 12/2 with ground is acceptable per the NEC for use on a 20-amp 220 circuit, no matter how many outlets you put on it.

 The "Pass and Seymour" plug listed by Steamtyme is a standard NEMA 6-20P design.

 Quite a few companies make molded cables with a NEMA 6 plug on one end and a C13 on the other end.
 A 6-15 plug will plug into a 6-15 OR a 6-20 outlet, a 6-20 plug is designed to only plug into a 6-20 outlet.

 If you are not an electrician, I DO recommend having an electrician do the wiring.

 On a hash/$ basis the RX 470/480/570/580 are comparable with the Nvidia 1070/80/80ti cards (not sure on the 1060 and below as I don't have any of those) *IF* you can find a RX series card at a "good for current market" price.
 On a hash/watt basis I believe the NVidia cards win fairly easily - it's hard to argue with 4 sol/watt if you optimize a 1080ti towards efficiency, and the 1070 and 1080 can hit close to that same figure or a hair better too depending on the model or specific card.

3.4 seems to give a little more hashrate on the 1080ti than 3.3, but my testing on both was not extensive.
It's been quite stable for me so far on my mix of various 1070, 1080, and pair of 1080ti models.
It's always had a fee, which unlike Claymore is configureable as more than "on, or off and lose some optimisation and hashrate".

 You're addressing a question there.
 Yield has to do with how many usable die you get from a wafer as a percentage of the POSSIBLE die you should get if there are no flaws.

 Efficiency - I'm starting to question their figures a LOT, 28nm to 14/16nm in Mining ASIC so far has seen about half the power usage per hash in most usage on comparably optimised products (S7 to S9 for example).

 Given that quantum effects have been limiting the gains in both possible clock rate AND in efficiency at an increasingly worse level per generation for GENERATIONS now, I doubt we're going to see an improvement of 2.5 TIMES the efficiency from 7nm - even if it really IS 7nm and not just CALLED that while having real feature size closer to 9-10 nm range.

Why would it be impossible for him to know about BURST prior to that "white paper" of his?

 On the other hand, if he DID know about BURST prior to publishing his "white paper" then he's little more than a ripoff.

FoldingCoin is NOT "just a token", it's a fully traded COIN that can be traded outside of the CounterParty wallet (I've not used CounterParty for MONTHS yet I earned and traded several hundred thousand FLDC in that timeframe).

 I'm QUITE certain that Poloniex and Bittrex (among others) do NOT use that Counterparty wallet to deal with FoldingCoin - THEY would not put up with the ripoff level fees CounterParty charges.

 The actual ASIC CHIPS are not made in China.

 There are *4* foundries in the entire WORLD that are capable of making chips on the 14/16nm node - TSMC in Taiwan, Global Foundries in New York, Samsung in Korea, and Intel (I forget offhand where their foundry is at for this node, somewhere in the US as I recall).

 *ALL* of those foundries are booked solid for MONTHS in advanced, and widely reported to be running the applicable lines at 100% capacity.
 Intel doesn't sell capacity to "boutique" manufacturers like Bitmain (I'm pretty sure their only "non-Intel" production is for Apple), GF is overloaded by demand from AMD and IBM and their other smaller LONG-TERM-CONTRACT customers (one of the reasons AMD GPUs continue to be in short supply), Samsung rarely sells their capacity to ANYONE else (NVidia is a rare exception), which leaves most of the demand for "small" chip houses like Bitmain to TSMC.


 Sure there are a ton of places that can make older node chips like 28nm and higher, but the ones that can make CURRENT STATE OF THE ART CHIPS like are used in the Bitmain S9 are overloaded by the demand.

 DEFINITELY don't confuse production of generic PRODUCTS they can make at home in China like Steel and Aluminum with high-tech production of Semiconductors where they are GENERATIONS behind the current state-of-the-art.

14nm vs 16nm is for most purposes nothing more than terminology.

 The actual PERFORMANCE of the nodes is pretty much identical - variations on EACH show a wider range of performance difference than the variations between the two.

 I also don't buy into the theory that the TSMC "12nm" process is really 12nm - even TSMC calls it an "upgrade to their 16nm process" NOT a new process node with smaller feature size - which just raises questions as to what the ACTUAL feature size on their "in the works" 7nm node are going to actually be.

 ANOTHER of the reasons I'm not buying fully into the "7nm" hype is that it's been pretty widely reported that most of the folks that have announced "7nm" (TSMC and GF in specific) are not really going to a true 7nm gate size, they're going to end up very close on ACTUAL feature size to Intel's "10nm" process.

 Samsung MIGHT be an outlier there, they've been keeping details of their "7nm" work very close.

 The "new announced" more efficient miner I'll have to see in action before I believe the announced specs - but it's certainly not impossible, BitFury was showing the capability to hit that performance level with THEIR current chip if you ran it at a low enough voltage.
 Bitmain current chip IN THEORY could get close to that announced efficiency level as well, but they apparently had too many "variation in performance" issues when they tried on the early S9 batches, so they backed off some on pushing lower volts for better efficiency in the name of better reliability.


 The last time Bitmain did the "3 units ganged together" thing was the S5+ - which was NOT a rack-mount miner that they reserve their even-number models for.
 The S5+ in retrospect also seems to have been the initial experiment with their current "heat sink on the chip" type design that they used in the S7 and later.

 They also work for FoldingCoin to my specific knowlege.

 I'm not going to argue the RIPOFF level of fees they charge - aka part of the reason "I despise it".