MSI Afterburner. Reduce Power Limit (I run ~ 68), un-check the temp link and set the temp to ~85 or so. The cards don't run this hot but this keeps the software from throttling, with a 20+ degree difference between actual and limit there is no throttling. Then increase core clock to 250 or whatever you can before it crashes, and you can leave memory at default or even reduce it slightly as it has little impact on Equihash mining. Also, keeping your cards cool is a must, as I posted above I have a lot of airflow going to and over my cards, so I think that is the main secret as if they start to get hot even in the 70's the cards will start to throttle back and the Sols will begin to drop.

True, good air flow is indeed important and I do have a lot of fans pointed at my rigs. For dust you can use furnace air filters in front of the fans to help. If you have a lot of dust issues you could build a box to direct the airflow better and reduce the dust even more. There is a YouTube video of a guy in Florida who put all his rigs into 3 plywood boxes he built, they use furnace filters to filter the incoming air and some 30" home depot fans to exhaust out the top.  While it may seem extreme, when you consider the 1080Ti's are close to $700 each, it might be a good option if oyu run yours in a dirty environment.

Can't find the video atm, but if I do I will update the post. Found it, think it is also been posted by others before, but here it is again: https://www.youtube.com/watch?v=E-_9F8R_PFU

Meh, I would just buy a breakout board with a power switch, and/or utilize the pico adapter and dump the ATX supply if you are just using it to power you motherboard. Sure you can search for a proper pin-out and do some type of hack, but if you mess up you could short out the PSU. Since most server PSUs are already used or refurbished they probably do not come with warranty and you would probably be out more than the $30 for a simple pico adapter or breakout board. There are some things to save money on and some things it is best to just do correctly.

I don't know, those numbers sound a bit high. You may need to throttle back on the power consumption of these cards a bit more. I have that exact card EVGA 1080Ti SC2 in a rig with 4 others (5 GPU rig) and they all only run around 60° C with the fans speeds only hovering around 40-50%.

They are all doing 800 Sols with a bit less than 140 watts each power draw (750W for the entire rig including system overhead), and operating in a 90° F (32° C) environment, so not unduly cool by any means. Even if I switch to ETH mining, they still are right around the same temps and fan speeds, although power use may creep up slightly.

Hat tip to you sir for actually posting your real (effective) rate!

Here in the Midwest I pay $0.095 summer and $0.09 in the winter, but with fees (meter charge, distribution fees, etc) it comes out closer to $0.105 per KWh.

I have been looking to go solar too, but so far have only been doing that, looking...

Confirming that I have updated my signature.

This is the biggest fallacy out there. Not the fact that you can move to other coins, as yes you definitely can, but in thinking that those coins can absorb all of Ethereum's current hashpower and still remain profitable to mine.

Even now people are still adding rigs daily, so even if Ethereum never went to PoS and remained a PoW coin, at some point the hashpower saturation will become so great that only those with super cheap or free power would be able to survive.

The only question is at which point this happens. External events such as price of a coin exponentially rising, such a ETH in the past year can delay the inevitable, but sooner or later the mining party will come to a spectacular end. So you not only need to consider if you can ROI within 7-8 months, but will enough profitable coins exist that can be mined long enough to keep this ship afloat.

Perhaps we get another "Ethereum" meaning a coin popular enough to absorb the hash-rate and keep the party going for awhile longer, but I think this blind trust in that mining will be profitable no matter what is a bit foolish.

It is probably best to start with GPU mining, if you have a current desktop type computer which has room for a full sized video card and a 16x PCIe slot you can use that as your base platform to get started. The Nvidia 1080Ti, while expensive (but well within your budget) is pretty adept at mining multiple algorithms while maintaining a healthy hashrate versus power consumption ratio. This is one of the main benefits of GPU mining in that it gives you the most flexibility to change coins, while an ASIC pretty much locks you into the algorithm it was designed for.

ASICs are also hard to obtain, can be hard to optimize and power properly, and have very short useful life-cycles in which to not only ROI but to make a little profit before they become more expensive to operate than the income they produce. While they are great for people with the infrastructure already in place along with the know how and ability to acquire the hardware early and in bulk, for most home miners this isn't usually the case.

Ahh, good for you then. My money would probably be on a faulty outlet then, so yeah good thing the house is scheduled to be rewired. You should maybe mark the suspect outlet somehow too so someone doesn't come along and use it and receive a worse shock or it causes a fire.

Yes, it should return via neutral as that is what it is designed to do. One other possibility, did you check for voltage between neutral and ground in case the outlet was wired incorrectly (i.e. the neutral is actually hot and hit is neutral)?

Some of those older outlets (often found in older homes) were not even keyed with the neutral being wider than the hot, and I have seen some even with the keying where they were wired incorrectly.

This is actually an incorrect assessment of the whatomine website and is not really useful. What would be more useful is to educate yourself before posting false information.

Others have already pointed out that whattomine and other such mining profit calculator sites simply provide a snapshot in time with a set of assumptions used as input. They may allow you to enter or change a few, but not all, of these input values. Thus the end result is that certain assumptions will automatically be used. One of these assumptions is that you will be mining with 100% luck, which again as mentioned by others, will in reality be less.

However, if you alter the inputs not only with your hash-rate, and power usage, but also include in the fees section account for pool fees, miner software fees, and a slight reduction for mining variance (luck) you can get more accurate predictions of your potential income.

So for instance, if you are mining Ethereum only with Claymore's miner you will incur a 1% miner fee for using the software. This in reality might even be slightly higher due to the switching of pools, but for simplicity lets leave it at 1%.

Then most Ethereum pools charge 1% for using their pool. There may also be a withdrawal fee, so again in practice the real fee will be slightly greater than 1% in some cases.

Then due to variations in both the pools luck, and your own luck in submitting shares, you incur the equivalent of 1-2% reduction just by living in the real world. While no one really gets this value, it is simply lost, we can estimate it and include it in the fee sections for the sole purpose of trying to get a more accurate estimate.

So include 1 or 2% on top of the above miner and pool fees, and you should be entering a total of 3 or 4% in pool fees sections. Also note this option is only available under the individual coin page, and not on the main whattomine overview webpage which does not take into account fees at all.

Now after entering all of this you will still only come up with a estimate of the current mining profitability. Current means at that exact moment in time with that exact difficulty and that exact exchange price. In 5 minutes or in 5 hours this amount will be different. You can set the display to use different averages, but even here it is still a snapshot so when it comes time to collect your coins off the pool and sell them, the amount of coins you receive and the prices you can sell them for will be different.

But this does not mean the site is misleading, or the information is false. You can still use the site to gauge the general profitability of a coin at that moment and use it to gauge the top 5-10 coins that are profitable at any given time. For the most part, if a coin is in the top 10 of the profitability list, it usually stays there for a length of time.

So one strategy is to chose a few coins from this top 10 list and not mine the one that is at the top, but instead focus on the one in the top 10 that has the lowest difficulty during the time period selected compared to recent history. Meaning look for the coin with the lowest or negative % change in difficulty. This coin will have less people mining it at that point in time, meaning you will be able to get more coins for a given hash-rate than you would if it were peaking with a lot of people mining it. Then you will need to adjust your trading habits as well and have some patience to sell your coin when it is peaking in exchange price. This usually (but not always) corresponds to its hitting number one on coinmarketcap or other sites.

So long story short, it is unfair to wrongly accuse whatomine or other sites as wrong or inaccurate, but instead learn what they do show and what they don't. Then learn to adapt and use this information wisely to your advantage.

It's coming down a little now again and it might be a good time to catch the dip. I have some staggered buys in place starting around 9500 (~$0.40) on Poloniex that will hopefully be hit. I missed out on the ICO and I do not believe we will see those prices ($0.05) again so anything under 50 cents may be a good deal at this point in time.

I am with the majority here in that if you treat them right they should last for a very long time. By keeping temperatures in check 70C or under is a good target along with using a good quality PSU as others have mentioned. Cheap PSUs can really shorten the lifespan as the power output can really be all over the place. Check http://www.jonnyguru.com/ for more information on this.

Keeping dust and dirt buildup to a minimum is important to not only to help keep things cool, but to prolong the life of your fans as well as help protect against humidity. Dust and dirt will attract and retain moisture along with the corrosive effects they can have on your GPUs over time. Going with this, is to keep them in an environment that will not be super humid. You do get a benefit here in with the heat the cards put out, a lot of them in a shed or garage can keep the surrounding air a good 10-20 degrees warmer than outside, and since warm air hold more moisture than cooler air, you can often keep humidity in check somewhat. The danger is if the power goes out for any length of time.

Ideally the best environment for your GPUs would be indoors in an air conditioned setting, but this is not always possible, especially for larger deployments. But even here, if you can keep humidity down, and address the other issues above, you can be successful. I have many GPUs going on their second year in a garage that is exposed to not only the heat of the summer but also the cold of the winter.

I guess I cannot attest to how long they will ultimately work out there but so far there have been no problems. I also have some GPUs that lived their first couple of years inside, and some of those are now among the garage dwellers, so they have a good 3-4 years on them and still going strong. Some of the earliest cards like the 7950's and 7970's needed to have some fans swapped out here and there, but an extra $8 in parts every now and then I view just as preventative maintenance more that I would as a card failure.

While it is definitely cool from a technological Gee-Wiz standpoint, I agree that it is getting a bit ridiculous. There is certainly a cost efficiency and management reduction by running multiple GPUs off a single motherboard, but at some point one needs to ask where is the optimal point.

Pros
- Decreased system cost as the Motherboard, RAM, CPU and storage (SDD or other) costs can be spread across more GPUs as they are added.
- Less administrative management of rigs as one 18 GPU rig is easier to manage as far as updates than there is with three 6x GPU rigs.

Neutral
-Power supply requirements stay about the same as it is mainly GPU driven. With the looks of the motherboard, it is expecting at least 3 PSUs anyway, which would be the same requirement for three 6x rigs.

Cons
-If one GPU crashes and causes the rig to hang, lockup, or fail, you lose all 16 GPUs of mining power during the downtime. With multiple rigs you spread out that risk.
-Introducing multiple PSUs into a single rig can cause more overall instability than a single PSU rig.
-Complexity, each additional GPU can add more complexity as far as drivers, OS support, stability, just general troubleshooting, etc.

I can probably think of others with time, but for me the optimum sized rig would be somewhere around the 6, 7 or 8 GPUs per rig ballpark and running on a single PSU.

The biggest problem is you have too complicated of a rig for someone who from the sounds of it doesn't have the first idea about mining or building and operating a rig. Don't take this as an insult, it is just the way it is as we all were newbies once.

However, I do think your intuition is right about buying another motherboard and CPU, you might have enough RAM already depending on how much they gave you as 4GB is plenty for a rig.

Bust the big rig up into two smaller rigs of 5 GPUs each. You may also want to pick up a couple of SSD's or you can even use an old HDD you have laying around. While a 64 GB drive may work, I would recommend a 80 GB or 120 GB to be able to install Windows on and work with a familiar OS. The 1000 watts PSU will work great for one of the rigs, but the 750 watt maybe be cutting it a bit close, unless you can really tweak your power consumption. So you may need to pick up another 1000 Watt unit at some point, You could probably run 3-4 GPUs off of it though for awhile just to get yourself going faster.

Once you got the new motherboard setup and going, install Windows. Start by installing just one or two GPUs directly into the motherboard if you have at least two full size 16x PCIe slots, install the mining software and just get familiar with how everything works. With such a simple setup it will also be a lot easier for people to help you with online.

From there you can add your PCIe risers and eventually get all 5 GPUs going. At this point you should have a lot better understanding of what is going on and can then use this knowledge to help get your second rig up and running. You may need to use a SSD and convert to Windows for the second rig as well.

While the USB boot drive and OS is great, again it is for advanced users and I think you need get acquainted with the basics before attempting to run and maintain the type of rig they sold you.

Don't listen to all these people, you did the right thing and got the 1080Ti, which is the best bang for the buck currently for just about all coins. Good hash rate in one card, lower power usage, and excellent resale value as it is actually very good at games what's not to like?

Well maybe encrypted wasn't the proper term, but it was a late night after an all-dayer trying to tweak the damn thing. In any event, it the BIOS is protected right now for most of us mere mortals without private tools.

Anyway, if you did crack into the BIOS can you kindly let us know the max hashrate and power usage you have obtained? This would at least give us a better idea of its potential as right now everything is speculation with it being locked down and in this locked down state the results are dismal.

I would not recommend taking out a loan right now. If you cannot afford the cost of the entire rig, you could still get in more slowly with starting with a single GPU. A Nvidia 1060 6BG can be found new for under $300 and will earn around $1.50 per day depending on your electrical rate. This will take a little over 6 months to ROI at current earnings, but it allows you time to experience mining and get to know its ups and downs.

If in say 4 months time everything is looking good and your ROI is still on track, you could look to adding a second card. Since your first one will be close to being paid for, your second card will take a little less time to ROI if you dedicate both the first and second cards profits to this effort.

In effect you would be investing your mining income into new hardware at that point. So the second card would take less time to ROI since you would have two cards working to pay it back. Then when your second card gets close to ROI, probably 9-10 months from when you first started, you could add a third if everything is on-track. Doing this by next year if everything is still profitable, you would have a 4-5 card rig going. Since this process would start to accelerate as you have more cards working to add to your investment income, after 2 years you would probably be up to 12-15 GPUs spread over three rigs.

While this may sound boring as I know most people want the get rich quick scheme, it exposes you to the least amount of risk as either mining will continue to be profitable or it won't. If it does you can scale in over time for basically the investment into your first card, occasionally you will need to add components for rigs, but even old PCs can be put to use if they have adequate number of PCIe slots.

When starting out you do not need fancy motherboards that support 8 GPUs right off, you can make do with some old motherboards that may only support 2-3 cards. The side benefit of these small rigs is the PSU requirements are also small. Save the big rig (6+ GPU) hardware recommendations found here for later after you have a somewhat decent farm going of 12+ GPUs.

If you can afford to swing two GPUs at first and run them out of a desktop you already own, so much the better and the process above would basically move along a bit faster. Also you could get by with AMD cards too, but at the moment they are overpriced new, used ones are hard to find, and from an overall mining efficiency standpoint the Nvidia 1060 is hard to beat.

I don't think the question is whether or not the drivers are able to increase the hashrate, as everyone pretty much agrees on this point. The question is does it increase the hashrate while maintaining the power saving measures many of us were previously running. These measures could include BIOS mods, overclocking, and undervolting.

So for everyone, the more helpful updates would include not only the fact that your hashrate returned to previous levels, but what was your pre-beta and post-beta power consumption readings and did you need to make any adjustments to return to normal levels.

For example, I have a RX580 8 GB that when I first got it I could get 30 MHs out of it at 122 watts. With the DAG issue it slowly crept down to ~27 MHs while still drawing 122 watts.

I then apply the Beta driver but need to revert some of the previous adjustments to make it work properly. I now am back to ~30 MHs in performance, but when I look at the power draw it is now using 150 watts.

So doing some simple math with the beta driver the card now consumes 5 watts of power for every 1 MHs of performance.

The pre-beta driver state it used only 4.5 watts for every 1 MHs of performance.

And the original state (where we all want to get back to) it required only 4.06 watts for every 1 MHs in performance.

So as you can see from the example, the beta driver (with adjustments to make it run stable) seems to draw roughly 0.5 to 1 whole extra watt for every 1 MHash it generates, which overtime can add up. At 30 MHs this is 30 watts extra per card, and in a 6 GPU rig it is 180 more watts. Since the new driver does not seem to work well with earlier BIOS mods or most overclocking and under-clocking settings most of us had in place prior, this is not really an improvement.

Some early sites did do the calculations and stated as much, problem was once people heard the 70 Mhash rumor there was too much desire to believe and ignore the facts.

Early testing does show the Vega getting into the low 40's when upping the power limit, but power usage also shoots through the roof with over 350 watts just for the card, or 450 for the entire system (1 card test). So yes it can put out some hashes if you don't care about power use at all, so while it may be ideal for college dorms, mom and dad's basement, etc., with free electricity arrangements, but not so good for everyone else.

As far as I know the bios is locked down (encrypted) so any attempts to modify it will not be able to be flashed to the card without being able to encrypt the file with an AMD approved key. So only software tweaks for now are possible.