I doubt it's viable, since the card exhausts a lot of air out of the side of the card, not all out the back or part back and part top.
 Cooling would be a MAJOR issue.

I don't bother with "high quality" fans - I just use cheap Lasko stuff with a custom-built plenum setup (2 or 3 cheap doors and a piece of plywood or particleboard or even fiberboard for the top works real nicely and is low cost), then when the fan dies in 3-6 years (sometimes a little longer) I can just swap out a spare I bought from Walmart or such.

I don't have a pressure meter or anything fancy, I just have to estimate that I lose 30-40% ballpark (depends on the filter, the cheap fiberglass stuff is more like 10% but don't filter squat, the high-level "filters out pollen" stuff drops quite a bit more) based on the air pressure I feel a few feet from the fans.


 I've been looking at the Baikal off and on for a while - I suspect it will stay profitable for a long time since they're THE most efficient miner available for any of the algos they support, but their process to order one is a nightmare and availability the last couple months when I've HAD the spare $$$ for one has been very very spotty as they seem to keep selling them out faster than they can make them.

 BURST has you write "plot" files with pre-generated data in very large quantities to your drive, then it searches a specific sub-set of that data every time it is told there is a new block to process for a dataset with the lowest "deadline".

 Has nothing at all to do with cloud storage like Storj or Maid at all, and you can't use the potted drive space for anything else.

I would bet they have a rider on the contract that if you exceed the negotiated usage by too much, they get to void the contract and negotiate a new rate based on the new much higher usage.

 You are mistaken.

The S5+ was released less than 2 months before the S7 was - it appears to have been a "trial setup" for what eventually became Bitmain's new "form factor" and cooling setup for their miners, as the S5+ was their first "3 boards in one box" and "small HS glued to the chips" miner that the S7 and later S9/T9 followed (then the L3+ went 4 board in one box, but those chips apparently run cool enough they could put smaller HS on them).

 There won't be an S11 for a while.
 10nm (TSMC) or 7nm (GF) needs to hit production and probably settle down and get the reliability up some from initial days before Bitmain will see ANY foundry capacity available to them on the new nodes, and that's gonna be another couple years most likely.
 AMD is only talking about 7nm for 2020, and they with IBM have FIRST CALL on GF capacity.
 Also keep in mind that Intel uses ALL of their foundry capacity internally, so their process doesn't tend to be available to anyone else except sometimes some "close partner" companies - which definitely does NOT include Bitmain.

 I could see the possibility of a S9+ or an S9 Lite at some point - though it can be argued that the R4 is a "S9 lite" already in most ways.

 I DID find the recent introduction of their new Scrypt Router interesting in a LOT of respects - they might be having "availability" issues on the L3+ chip so they decided to take the few they had left and put them into a higher-margin product to max out revenue 'till they CAN get more chips in.
 On the other hand, they didn't introduce their SHA256 Router 'till the S5 (which used the same miner chip) was out of production or almost so, which has some potential interesting implications for the L3+ chip/miner....


 Gotta wonder how worried Bitmain is about the pending ASICBOOST "cease and desist or get sued" issue, and how it's affecting their timeframe on when to get stuff out.
 Perhaps they're working on a new chip that doesn't infringe, and are trying to push out all the old stuff ASAP?


 If they DO put out new non-infringing chips, gotta wonder of the miners with those chips will get new names or not.

Did GreyMatter "borrow" the basics for their case design from BackBlaze?

 Back of that case looks MIGHTY familier....

 (the Backblaze designs were made public, GreyMatter didn't do anything wrong if they DID borrow).

A 3-card RX 480 / RX 580 rig RIGHT NOW will gross over $300/month on ETH, somewhat less on ZEC - then subtract the electric cost (probably $1/day or a little less unless your electric is NOT low cost).

 Cost to build such a rig with reasonable care at parts buying is $1200 or less, you MIGHT get it down a bit under $1100 or so with very careful selection of the lowest-price parts.


 This was NOT the case as little as 2 months ago - profitability on ALL major coins has more than doubled over the last couple months as many of the major coins have had HUGE price runups (and the difficulty has NOT been able to keep pace).
 For perspective, my mix of ASIC and GPU gear hasn't changed much in the last 3 months (I added *one* GTX 1080 card), but my income has jumped by a factor of *ALMOST 3* over the last 2 months.
 (I can't wait 'till I get into the new place I have lined up with over double the power available - I'm power-limited at this point BADLY).


 The nice thing about having quite a few profitable coins is that if one coin has a big jump, it will soak some of the hashrate out of the other coins in the "basket of profitable coins" and drive THEIR profitability up some as well, 'till the profitability is back in the same ballpark for the "basket" coins, so you don't really HAVE to go chasing "the most profitable coin" every day or every hour to benefit from the price jump of one specific coin.

 There is also the Nicehash option - they're usually pretty competative on profitability with whatever the "hot coin of the day" is.


 One other major issue to keep in mind.

 ETH is going Proof of Stake at some point, *probably* this year sometime - at which point there are over 29 THOUSAND TERRAHASH (this amounts to about 1 MILLION GPUs) worth of GPUs plus whatever gets added between now and then that are going to be looking for new homes, which is going to HAMMER profitability on every other coin that CAN be mined by GPUs (mostly the higher-profit by AMD mineable ones, as most ETH is probably being mined by AMD cards) - at which profitability for GPU mining is probably going to drop back down to the ballpark it was in 2 months ago even if the current coin PRICING stays high.

If you want enough of them at one time, Innosilicon would probably be willing to sell a lot of A4 chips.
 I doubt they would sell the SMALL number needed for a single rig though.

 Bitmain AFAIK has NEVER sold individual chips, just miners.

 Gridseed became part of SFARDS (merger with WiiBox), which has offered no news at all that I've seen for something like a year now - I'm not sure if SFARDS is even still in business.

 The Alcheminer folks seem to have given up on their second-gen chip project, and I don't know if they ever actually SOLD individual chips (they did advertise them for a while).

 Silverfish died a while ago and only made miners.

 Zeus was reduced to doing custom-work for Innosilicon last I heard about them, and never sold individual chips AFAIK.

 KnC died, but apparently someone bought their IP out of bankrupcy and is talking about building/selling a new-gen miner.
 Makes me wonder if the KnC "solar" chip ever existed though, given the STATED specs on that thing.
 Also, given the HIGH "part of die died" issue the Titan chips had, do you REALLY want to trust their chips?

Depends on which GPUs you are looking at - I'm NOT seeing any shortage of NVidia, and the RX 5xx cards seem to be getting into reliable supply while the RX 4xx cards have mostly dried up.

 I'm still seeing "on sale" cards every day that I look too - just not always the same ones, and it DOES seem to be a bit fewer right now.

 You could check Frys and Tiger Direct, they're probably the next 2 biggest US-based retailers that carry GPUs at a competative price.

Claymore's dual miner and Genoil's ETH miner are pretty much a tossup on net mining speed for ETH, due to the "fee" Claymore's miner inflicts on it's users.

 If you are JUST going to mine ETH IMO ignore Claymore and go Genoil, if you plan to dual mine then go Claymore of course.

Most NAS I've ever worked with had a fairly low capability CPU in them, NOT suitable to even try to mine on.

 I am not aware of any that use a standard motherboard with PCI-E slots other than the Backblaze boxes, but those aren't exactly normal NAS devices....

An S7 under current conditions would still be profitable if your electric is in the 3c/KWH range - but I doubt you could ROI one even if you paid $100 for it unless you DO have free electric.

 I would count "electric included" a "free" and legal, though if you pull enough more than the norm the landlord might come talk to you about renegotiating the rent amount (or in my case, I negotiated it AHEAD of time with the landlord and with my last couple of utility bills in hand, since he KNEW I used a lot before I moved into one of his "other" places).

In warm weather, unless your breaker panel is in an area with A/C, you probably WILL pop the breaker with that much load on it.

I can confirm that even junky Virgin Mobile 3G cell connection was enough bandwidth to support a small-to-mid sized farm and would be plenty for a backup - but I don't think they offer their "unlimited" 3G plan any more, just LTE (which in theory should work a lot better).

 They don't, though, at this point.

 On folding, based on my own testing on my own cards, a RX 470 pulls very close to the same power that a GTX 1070 does - but manages about 40% of the the PPD  (230-250k on most work units vs 600+ k on most) - which means it's making noticeably less than mining ETH then it does folding for FLDC and CURE combined even at the current very high FLDC pricing - and ETH isn't always the highest income option for a RX 470.
 An RX 480 would be a little closer on PPD, but would also use a little more power vs. a RX 470.
 The best AMD folding card was the FuryX/Nano, and even THAT only managed 400k PPD ballpark on the highest figures I've seen posted for those - and those apparently use a fair bit more power when folding than an RX 470 or a GTX 1070.

 At current pricing, I'd call the GTX 1080 and GTX 1080 ti pretty much a tossup with the GTX 1070 on both PPD/$ and PPD/W - with the overall system cost and efficiency probably favoring the 2 bigger cards by a thin hair now even though the card-level specs still seem to give the 1070 a hair-thin advantage. Seem to come down more to "what is on the best sale right now" for PPD/$ between those three.


 Proof of Research  "difficulty" has actually been increasing the last few days, just very slowly.
 Team CureCoin PPD is at it's all time high, and keeps breaking that every day.


 I suspect the devs seen the miners as a "needed evil" for securing the blockchain, no more no less.
 I am somewhat suprised they allocated 20% to the miners when I think about it.

 As far as I know the only ASIC available that does Quark algo is the Baikal.

 None of the coins that use Quark are exactly high market cap, I think the only reason Baikal included the capability is that it's one of the algos used as part of X15 (and possibly X13 and X11?) so they were going to have to support that algo anyway.

Realistically, if you aren't in a major dusty area, dust in mining rigs isn't a major issue - blow them out every month or three.

 My preference for filtered intake positive pressure is more about *ME* having alergy issues and wanting to keep the pollen out, though the reduction in dust in my mining area IS a nice bonus.



 I generally have ended up building up a "plenum" area, commonly 30-36" on a side, then have 3 standard Lasko-type room air fans sucking air in via the plenum through 20"x20" furnace filters. I lose SOME airflow due to the filters, but I'd estimate I'm still getting 4000-5000 CFM flowing through the setup as long as the window I'm pulling the air THROUGH isn't too small.

 Windows firewall is a pathetic joke and always has been.
 You can't expect REAL protection when the OS your firewall software is running on top of is full of security holes itself.

 There is a REASON all of my Windows machines ALWAYS connect to the Internet through a firewall built on LINUX or Cisco's router/firewall OS.

 INFINITELY better protection than Windows will ever manage on it's own.

 You have plenty of options for filtered air as well, especially if you build a "plenum" type setup - I commonly use standard Lasko room fans on my intake setup and ordinary 20x20 furnace-type filters.

 I dunno why you think you need to use anything special like you're talking about.

 Also, you get unfiltered air entering your mining area from any LEAK in the entire room, if you use exhaust fans. If you don't care about filtering, this generally isn't an issue though.



 The difference in cooling efficiency for negative (exahust fan setup) vs positive (intake fan setup) pressure isn't enough to be noticeable unless you have a HUGE backpressure issue.
 You MIGHT see a 1 degree difference at times.


 To GMPoison - to exhaust air out of the room you have to have intake SOMEHOW from outside the room, so it doesn't matter if you use intake fans or exhaust fans, you WILL be sucking "hot outside Florida air" into the room one way or another.

 That will vary a TON depending on the project, but I suspect you're not going to see a lot of coin/day on ANY of the GRC whitelisted projects.

 For perspective, a little over 2 million RAC on Moo Wrapper gets me less than 200 GRC/day, and needs roughly 15 Gigakeys/sec of Dnet RC5-72 keyrate to achieve.
 This particular project is VERY friendly to AMD GPUs though recent NVidia GPUs do quite well too (it's pretty much ALL about the number of cores and the clock rate), while the fastest CPUs ever managed less than 10 MEGAkeys/sec per core.

 Other BOINC projects in the GRC whitelist are friendlier to NVidia GPUs, and I think there are a few that are CPU friendly (your i7 should do comparably well on those).


 Unlike most cryptocoins, GRC doesn't work with a single "algorithm", in effect it has about 30 of them to chose from.