GPU Boost 3.0 cards ignore manual temperature limits in tools like MSI Afterburner or EVGA Precision.

You can set the temps to let's say 65°C but as long as the power is not limited the card will heat up to as much as 80+°C (card in question is AORUS 1080 Ti).

Is there any tool that enforces a custom temperature limit and throttles the card based on that for windows?

6*170 is 1020 watts plus the system is probably around 70 watts. That's 1090 watts.

Personally, I don't go above 80% on PSUs for 0-24h use so the effective capacity for me would be 960 watts.

Therefore if you were to reduce the power limit on the cards to 87% (147.9 watts per card) you'd end up with a total of 957.4 watts.


But, even if you were to use 1200 watts, the PSU should not make any weird noises other than the louder fan.

Mt. Gox

When this kind of money is involved I'm pretty sure there are some super efficient miners in some form and ASICs are probably just around the corner.


True, but it's still only 0.016% as per: https://etherchain.org/charts/difficultyBomb

Thing is, if a dev or group of devs are confident in their project, they should finance it themselves. That at least shows that they are confident and not just want a quick payday. They should launch as PoW as it's still the only fair initial distribution method and mine it themselves or buy cheap early coins and hold if they're confident they can deliver as that would make them great profit. Or even make it so they get a cut automatically for a period of time from every block, Zcash style (but not as much as 20%). That way they're incentivized to keep improving the project and can't just get an initial payment and lost all incentive to continue and not just disappear.

And ICOs are going against everything crypto stands for; it's a centralized form of inital fund distribution.

Software is full of bugs and with arbitrary rules and duct tape solutions like you can only have as many drives as many CPU cores you have - ignoring CPU speed, bandwidth and disk speeds.

And many aspects are not transparent like how payouts are calculated exactly.

$30m is an outrageous amount for development for a project this size.

How does 2nd ICO plus won't be on exchanges for a while plus leadership issues plus crappy software equals moon?

It might be pumped by people dumb enough to buy into the second ICO for a short time but I don't see any moon.

Here's my knowledge in a nutshell; all ICOs are scams. No exceptions.

Slightly longer but still short version: we have a lot of new, mostly lazy people in crypto thanks to the rise of Bitcoin who look at this as a gold rush and want a quick and easy buck and consider mindlessly gambling with ICOs as smart investments. Your 11-day-old account is a testament to that. While I'm not against gambling as long as the participants are understanding what they're doing (I lived off of playing onlike poker for years) ICOs are nothing but a distilled and rotten form of gambling without rules. It's like a game of cards where the cards absolutely doesn't matter and you just give your money to the house which decides if you won or not. This phase in crypto is even more detrimental then the "every wallet has a keylogger or trojan" phase following the "every coin is heavily premined" era (which seem to have a comeback).

EU anti money laundering laws are no different then KYC and AML - which almost all popular exchanges abide by, since most of them are registered in the US (which is kind of funny).

Not to defend them, as I personally never even use my real name and address, not because I have anything to hide, but I simply do not trust any government controlled by wrinkly old men or women. And giving your name in order to trade cryptocurrencies kind of defeats the purpose.

Virtual currencies should be at least pseudo-anonymous. That's the bark of cryptocurrencies so chewing that layer off - while seems innocent to some - is the first step of trying to take control of them.

Using words like "legitimate users" is the equivalent of saying "if you have nothing to hide we should be able to monitor all your activity 0-24" so fuck any and all push against cryptos - however hidden and tame they might appear.

I mean, for fuck's sake, some people are trying to ban encryption of all things ignoring they all rely on it on a daily basis! That alone shows how out of touch some people in power are.

On paper you'r right, but from experience the specs are not very accurate.

I'm confident in saying that Molex can't handle 132 watts without getting very hot or even melting due to poor connectivity and +30°C at 11A is not exactly intenteded for 0-24 use. In a 30°C rig that's already 60°C which is just on the borderline of being too painful to touch for more than a few seconds.

In addiction, in my experience Sata is getting sold short; it can handle way more than 54 watts without even getting slightly warm.

Personally, I would much rather use each Sata port with a Sata to Molex cable at 60W then putting 2 cards on one Molex cable.

Which wallet can be used for solomining with GPUs under windows?

There's no such thing as replacing coins in crypto.

Every single coin since Bitcoin, was created to try and do something better and eventually replace Bitcoin.

No major coins will be replaced overnight.

Using 3 cards on 3 separate molex chains and the 4th card with a sata#molex should be fine.

Try it, run them for like half an hour and check if they get hot where the others have burned.

50% premined bullshit.

As far as I can see the default power limit for FE 1080 Ti cards is 250 watts but they can be pushed to 300 watts (source).

Now, some numbers on different cables:

A 8 pin connection going straight into the GPU can provide 150 watts and 6-pin can do 75 watts so the card can get 225 watts straight from the PSU.

Molex connectors should handle 11 amps with a 30°C increase so 132 watts. That can be already pretty hot to touch especially if a connection is not that great.

Sata on the other hand can only do 54 watts as it has 3 x 1.5A pins for 12V.

PCI-E x1 can deliver 25 watts and PCI-E x16 can do 75 watts. This is why we use powered risers as the x1 can't provide enough power for the GPU.



With the numbers given it should be clear that you shouldn't put more than 1 card per cable in case you're running them at 100%+ power. And different cards can also have different limits in their BIOS for how much they're allowed to pull through PCI-E.


I used to have melted Sata>Molex cables when I used 2-3 GTX 750 Ti per cable (60 watts each) and those cards had no extra 6-pin. The 6-pin models at the Molex were cold as they could pull the same power through PCI-E.

So I put all cards on different cables which completely eliminated the heat.

That press release is a response to Bitcoin Unlimited, a failed hard fork attempt and not for Segwit, which is a soft fork. Though their stance is probably the same.

Gigabyte AORUS 1080 Ti cards.

@philipma1957

Efficiency is not everything. While you lose out on some efficiency on 100% power limit or even higher, you earn more.

Besides, setting cards to 100% doesn't mean they will always use 100% power, most algos don't go anywhere near that, more like 70% but for some power hungry algos it allows the cards to eat more and produce more.

Feature request: pool reconnecting.

A lot of the time the initial connection to a pool fails and the miner exits with Stratum subscribe timeout. It would be nice if we could have an argument which allows for reconnections.

Similar feature from ccminer:

 -r, --retries=N       number of times to retry if a network call fails
                         (default: retry indefinitely)
 -R, --retry-pause=N   time to pause between retries, in seconds (default: 30)

I keep getting this on a couple of 1080 Ti's randomly:



Sometimes it happens to only one card, sometimes they all fall together and flatline in MSI Afterburner.

Happens in both 3 and 4 card 10870 Ti rigs on a 1300W EVGA PSU with powered USB risers and even on a rig with only 1 card.

Seem to happen more frequently when the power limit is at 100% instead of around 70-80% but it might just be a coincidence.

The cards are only around 65°C.

Any suggestions what could cause this?

Yeah, I agree, I downclocked my GPUs so it happens rarely, but nevertheless, every few days it happens on one of the rigs, and usually at 2AM, which costs me quite some hours of no mining.

I dont think there can be a unified solution for both AMD and Nvidia, rather a specific solution for each of them, but I'm no dev, so if someone is - please chime in.
[/quote]

As a duct tape solution you could restart your miner let's say every hour with something like this:

restart.bat:

:loop
timeout -t 3600
taskkill -t -f /im miner.exe
goto loop

That is if you also run miner.exe in a loop.