Happy New Years! Seventh alt coin thread!

[Fireballbu]

Phil, why do you think the Titan XP is worth the extra cash  over a 1080ti?
I was debating on what my next batch of cards should be...

[1713530049]

1713530049

[1713530049]

1713530049

[1713530049]

1713530049

[1713530049]

1713530049

[Storx]

im not gonna lie, im happy with the drop in btc, i had 100's of trades finally kick off as the price droppped... so i basically broke even with the drop, so if it would only go up now i would make a nice chunk of profit....

Plus it would loosen the strain on the GPU market, so i can acquire some more units to grow my farm...

[dragonmike]

The Titan V is very nice... but too expensive to make any economical sense currently. But good on you if you want to play with it, Phil. It should be a good indicator of what's to come for the consumer/gamer market. Are we expecting the Volta cards in Q2?

Coins seem to rebound a little after the massive drop. It'd be really nice if they could settle around these levels a bit. Build a base and start going up again. Mining-wise, all we need is for China to ban mining farms and we're in the money!

[QuintLeo]

  titan xp and titan volta

of course  I have to get them first.

 Titan Xp, or the "Star Wars" editions, were in stock most of January on Nvidia's site.
They DID run out sometime this month, dunno if they have restocked yet.

The price even on the Star Wars editions was over $1100 though ($1138 I think?) and the non-Star Wars ones are $1200, and they're FOUNDERS edition with so-so cooling.
IMO not viable as mining cards, they offer very little to NO performance advantage over the 1080 ti due to lower clocks vs the small increase in core count and memory width.

Titan V is $3000, and ONLY available direct or as part of "high end GPU workstation" builds for even more, NOT viable for mining due to the crazy cost unless you have some other reason to get the thing.

[Toughit]

Newegg $799 in stock
https://www.newegg.com/Product/Product.aspx?Item=N82E16814137115&nm_mc=AFC-C8Junction&cm_mmc=AFC-C8Junction-VigLink-_-na-_-na-_-na&cm_sp=&AID=10446076&PID=6163686&SID=jdc8528s91011rh100053


edit 5:51 eastern, gone

[QuintLeo]

Got my 1050 and 1050 ti "experiment" cards in today.

 I'm NOT impressed.

 EVGA 1050 single-fan 2GB, EVGA 1050 SC single-fan 2GB, EVGA 1050 ti single fan 4GB.

 ALL of the cards have a hard-coded TDP limit at 52.5 watts, they will not LET you go lower at all - so 70% is the minimum, which doesn't leave a lot of room for tuning.

 Performance on the SC vs the non-SC 1050 models was identical to within measurement limits, EXCEPT the SC model would not allow as high of an overclock.

 On ZEC, the best efficiency was about 2.7 watts on the 1050 ti, 2.5 on the 1050 cards.
 Efficiency was pretty much FLAT between 70% and 80%, then dropped off some a little above that.
 Max hashrate was about 153 on the 1050 cards, 175ish on the 1050 ti, dropping about 25 hash on all 3 at 70% TDP.

 No memory setting I tried, from -500 (which CRASHED the ti and the SC) to +500, had ANY noticeable effect on ZEC hashrate at all.
 ALL 3 cards liked as much core overclock as I could feed them, but that was only about +100 on the SC and +200 on the other two cards.

 On a hash/$ basis the 1050 is comparable to any other GTX 10xx card I've worked with, the 1050 ti bit worse than even the 1080 ti - and that's comparing to pricing in October/November before the shortages hit.
 1050 would be a tolerable entry-level ZEC card if your electric cost isn't a factor or you have VERY cheap electric.
 On a hash/watt basis they are VERY POOR.

 On the UP side, none of these models have a 6-pin PCI-E connector, so if you have a MB with a TON of slots like the B250 Mining Expert you could cram a BUNCH of these cards in one rig with a single not-very-large power supply.

 Cooling was fair, but not as good as my bigger cards EXCEPT the blower models, they're a bit better than the ASUS cards and a LOT better than the MSI "Turbo", but they were fairly close on temps to my EVGA 2-fan models.
 They do NOT like being crammed on top of each other with no space though, 52.5 watt was hitting 82C on the one card I tried that way.
 One slot between them empty was tolerable - they MIGHT cool well enough to be viable on the later Onda motherboards.

 The 1050 ti on ETH was pulling a little over 13.3 Mhash - TDP didn't matter, core clock didn't matter, +500 memory clock DID matter (it was mid-12s at stock clocks).
 Card only would use 62-63 watts when TDP was set to 75.
 13.3 Mhash doesn't compare well on hash/watt vs one of my 1070 cards that manages 29 at 90.

 If high-end Polaris pricing was SANE, this card would be a very poor choice for ETH mining on both a hash/watt and a hash/$ basis.
 Ditto vs the 1070 on hash/watt though it would be somewhat close on hash/$.
 At current pricing, it might be TOLERABLE as an entry-level card, with the same "no PCI-E power needed" advantage mentioned above, but I still can't really recommend them other than "you can GET them, better cards are bloody near impossible to get".

[philipma1957]

Phil, why do you think the Titan XP is worth the extra cash  over a 1080ti?
I was debating on what my next batch of cards should be...

The  xp cards are not worth 1138-1200. They hash at 800-825

But I would i would not mind 1 or 2 of them.

The Volta are worth 1800 or less if you only mine with them.

[Steamtyme]

In case anyone is looking for a 570 8GB

https://www.newegg.ca/product/product.aspx?item=n82e16814137256&nm_mc=bac-criteo-ca&cm_mmc=bac-criteo-ca-_-video+cards+-+amd/ati-_-msi-_-14137256

Max 1 per customer

[VyprBTC]

The Titan V is very nice... but too expensive to make any economical sense currently. But good on you if you want to play with it, Phil. It should be a good indicator of what's to come for the consumer/gamer market. Are we expecting the Volta cards in Q2?

Coins seem to rebound a little after the massive drop. It'd be really nice if they could settle around these levels a bit. Build a base and start going up again. Mining-wise, all we need is for China to ban mining farms and we're in the money!

It's just a correction, you can have hyper inflated growth like that without it correcting. It's not a crash, it's a correction. I bought a couple more coins at 6500 but I wish I would have saved those 3 coins I bought when it was 10k but at the end of the day it doesn't matter, what's an extra 30% when it's going to 6-8x by the end of the year.

[Storx]

I had a good handful ask me in PM how i managed to break even with the price of bitcoin falling nearly 70%... so i figured id post it up, ive mentioned it in the past, but just how i do things...

Anyways, currently i profit switch to catch the most profit i can make per card, many are against this, they say it stresses the cards out, but i have had no issues doing this as long as i properly tune each card per algo, so its not running on the edge of breaking with the tuning, i run just enough OC to make it reliable, if it ever crashes, i look at awesome miner log to see which card locked up or failed and i go into my gpu OC settings and back it off 10 and just accept it at that rate....plus i never run my gear above 70c... currently everything is rock stable at 52c on watercooling...

anyways, the way i broke even is every day i get a payout from the pools i mine, i mine to multiple pools, because it seems different pools have different payouts on the same exact algo's...
Example: anytime equihash is the highest coin to mine, i have the pool setup to mine it from nicehash, because when i test different pools using identical gear, nicehash even tho reported less profit than lets say mining pool hub, they seem to always reliably payout more.... then skein, for some reason miningpoolhub always pays out the most on that algo when i try to mine it compared to others, and then blake2s is best on ahashpool for payouts it seems... just got to do your own testing and figure out your best options for your gear...

I use to use bittrex for this process, but with the all the news on tether, i backed off from using it in fear of it crashing one day... so i use GDAX now... daily i will receive 1 or more payouts due to how much gear i have currently. Everynight or around the time of the payouts i sit down and i take the payout for thats day mining and i throw it in a order based on how bitcoin is going at the time. for example right now with bitcoin crashing so fast and we were seeing 20+% movements in a day, i have been taking half my payout and selling it at a point that was hit multiple times during the day in hopes of getting maybe another 1% out of the payout, then the other half i have been placing it at the 24hr high and just letting the orders sit tell they fill, then when my orders would sell i would be in USD not tether and place a buy order for 10% below what i sold it for...

Example:
currently BTC today as i write this is as follow..

last 24hrs,
current $7,051.75
low $6,048.26
high $7,850.70
diff of $1,802.44 or 22.95% swing

today payout was $310 payed to my coinbase account @7051 current price.

so i placed an order for $160 @ $7850.70 and $160 @ $7250 since that point has been heavily test multiple times...
so... @7250 i would gain roughly 5.5% more on my daily mined earnings... and @7850 i would gain roughly 10.2% more on my daily mined earnings...
In the next few days if one of those sell then i place a buy back order in for far lower, lately i have been putting them in for 20% below and have already filled a good portion of those orders...

so lets says the @7250 fills tonight, then tomorrow i will place a new order for the $168.8 worth of USD in for btc @6525, so if btc ever drops that low again i will be in a better position at 15.5% of the price of what i previously had it at....

I have been doing this for months and has bitcoin raised up and up all the way to 20k, i racked up nearly 500 open orders with only half of them filling, because the price was only going up and not down.. so i managed to cash out all my minings into USD as the value of btc rised, but i had 100's of buy back orders sitting on the books, with this plummet in value my position in btc has risen above what it was worth when the price was around 17k....now im starting to accrue sell orders for when the price of btc goes back up...

[Storx]

Can someone with an antminer give me some info, i have very little knowledge on how all the ASIC miners work in regards to mining since i dont own any. I have been thinking of picking up some to try out and use on an idea i planning for in the near future.....

Recently i placed an offer and won with the next county over from where i currently live on a section of land 80.1 acres that went on auction for around 20k due to the county taking it due to back taxes owed. The property is nearly unable to be lived on due to the current home on the property has been abandoned atleast 30+ years, because the forest has eaten the house basically... and the house on the property has not had electricity since 1986 turned on, so they are telling me it would takes months for them to get out there and replace the line running out to current house spot on the property, since half of it has been knocked down due to trees falling....

I am thinking of doing a solar only farm possibly in the future, was thinking of buying a few ASIC miners and building a small building, maybe a shed or pole barn on the property and setup some solar panels in a cleared area of the property and a small battery bank... then basically setup the power to only come on during set hours of the day, basically only when solar is able to be used... idk how well a normal PC would function in this manner, so i was thinking maybe an asic would work best. The idea is to maybe have a programmable invertor that i can set voltage ranges that power would come online and power the miners, then it drops below a set voltage range that it would cut the power tell the next day of sun... the batteries would only be there for filtering during periods of cloud cover and such.... really only enough to run the gear a few minutes....

Idk its an idea im throwing around to possibly cut my cost on electricity usage, the solar could pay itself off in just a few months possibly from not having to pay for electricity

I still want to get power going out on the property in the future, as i plan to start taking the house back from the forest and demolish it and build a new one.. the property is only like 20 minutes from where i live now and is nice and quiet...

I have a tree company coming out to see if they want to log it for me, rough estimate they are telling me i could make maybe 15-20k from allowing them log the old growth trees on the property....

[philipma1957]

What state ?

You need to understand how to use solar and buy back or sell to grid is the better then battery if your in the right state.

Here in nj

A peak field of 160 k watt will net about 30 kwatt 24/7

Which means you sell 130 and use 30 during the day.

Then buy 30 when it is night.

So it is a constant 30 kwatts .

No cost after set up.

Should last 20-30 years.

You need 1.2- 1.3 acres of panels to do what I described.


That is the size of our field.

It is the low risk part of your mining as the power is free after setup.

If you exceed the 30 kwatt that Mining is higher risk since it is full cost from the grid.

Buysolar and I do this expand and contract beyond the solar cap.

Knowing we can retreat to all solar.

Now when btc was 19000 we spent an extra 2k on power and mined like mad . Then a new tax year came and we sold gear . Our excess power for Jan was 1500 and for feb will be 500.

[Storx]

What state ?

You need to understand how to use solar and buy back or sell to grid is the better then battery if your in the right state.

Here in nj

Florida, we currently cant sell back to the grid... so anything excess would just be free energy to the grid to use... unfortunately.. i am not planning on using a lot of batteries, just some to even out the dips in electricity during the day.. the goal is to only mine when the sun is out to power the solar panels basically

[philipma1957]

What state ?

You need to understand how to use solar and buy back or sell to grid is the better then battery if your in the right state.

Here in nj

Florida, we currently cant sell back to the grid... so anything excess would just be free energy to the grid to use... unfortunately.. i am not planning on using a lot of batteries, just some to even out the dips in electricity during the day.. the goal is to only mine when the sun is out to power the solar panels basically

Yeah no buy back sucks.

Batteries are costly.

They are dangerous

They don’t last

But if you get dead ones from a recycler many can be used as one or two cells are dead and four or five are good.

Proper wiring can get them to work.

Someone posted he would go to a forklift company and cut deals with them getting partial batteries

[MagicSmoker]

I've been trying to compare miners and pools using two GTX 1080 cards in separate computers that I've adjusted clock rates/power limits in MSI AB so that any given miner will hash at the same rate on either machine. I assumed that I would get reasonably valid results by having the two competing miners run at a fixed share difficulty for 24 hours to separate addresses on the same pool then reporting the total payout for that time period, but 3 separate people have said that payout is not a reliable metric and that I should be counting the number of shares or even the average hashrate reported by the pool instead, but I honestly can't see how either of those metrics would be any better than payout. I should note that I am not looking to validate my position here, rather, I really want to know why it is wrong (if it is)!

So how do you veteran miners evaluate which miner delivers better performance?

EDIT - well, no thanks to you slackers I think I figured it out (I hope!); hashrate reported by the pool may, indeed, be suspect, but it should be equally suspect regardless of the miner. Conversely, there will always be variability in the time to find a share even at a fixed difficulty and since payout is a direct function of share count * difficulty it will also have some variation based on luck, as well as the hashrate.

So, carry on!

[shynepapin]

I would like to get some advice from experienced mining farm owners on this. I recently came into possession of 3 PDU's that came with a used 42u server rack I bought from a data center. The models are APC7898 and APC9560, the 7898 is rated for 208V, 35A/~ input and output is 208V and up to 60A, 20A per bank and it's metered and monitored but with a Hubbell CS8365C plug that's rated for 50A.

Is there anyway this is useful for me? I have a 240V 30A electric setup. Also, with the 9560, which is a basic 1U rackmount PDU, that's rated for 120V 24A, I tested with the 240V, it works and was providing 240V, confirmed with my Kill-A-Watt P4400.01 as shown on Philipma1957 on one if his Youtube video. With the APC9560, I would like it metered and so I can monitor is always since it's not rated for 240V originally which is a great risk to take.

I would appreciate any feedback on this and if anyone can share a link to good metered PDU that's not expensive and rack-mountable, will be appreciated. Been stalking Ebay for one for days now and nothing so far.

[philipma1957]

I've been trying to compare miners and pools using two GTX 1080 cards in separate computers that I've adjusted clock rates/power limits in MSI AB so that any given miner will hash at the same rate on either machine. I assumed that I would get reasonably valid results by having the two competing miners run at a fixed share difficulty for 24 hours to separate addresses on the same pool then reporting the total payout for that time period, but 3 separate people have said that payout is not a reliable metric and that I should be counting the number of shares or even the average hashrate reported by the pool instead, but I honestly can't see how either of those metrics would be any better than payout. I should note that I am not looking to validate my position here, rather, I really want to know why it is wrong (if it is)!

So how do you veteran miners evaluate which miner delivers better performance?

EDIT - well, no thanks to you slackers I think I figured it out (I hope!); hashrate reported by the pool may, indeed, be suspect, but it should be equally suspect regardless of the miner. Conversely, there will always be variability in the time to find a share even at a fixed difficulty and since payout is a direct function of share count * difficulty it will also have some variation based on luck, as well as the hashrate.

So, carry on!

30 days  test at the same time  and still only a chance it is better. and correct  numbers

[philipma1957]

I would like to get some advice from experienced mining farm owners on this. I recently came into possession of 3 PDU's that came with a used 42u server rack I bought from a data center. The models are APC7898 and APC9560, the 7898 is rated for 208V, 35A/~ input and output is 208V and up to 60A, 20A per bank and it's metered and monitored but with a Hubbell CS8365C plug that's rated for 50A.

Is there anyway this is useful for me? I have a 240V 30A electric setup. Also, with the 9560, which is a basic 1U rackmount PDU, that's rated for 120V 24A, I tested with the 240V, it works and was providing 240V, confirmed with my Kill-A-Watt P4400.01 as shown on Philipma1957 on one if his Youtube video. With the APC9560, I would like it metered and so I can monitor is always since it's not rated for 240V originally which is a great risk to take.

I would appreciate any feedback on this and if anyone can share a link to good metered PDU that's not expensive and rack-mountable, will be appreciated. Been stalking Ebay for one for days now and nothing so far.

maybe  the apc9560 is good for you

I need to go to website and check

Busy  maybe tonight

[shynepapin]

maybe  the apc9560 is good for you

I need to go to website and check

Busy  maybe tonight

No problem Sir..., take your time, appreciate the quick feedback already, I just don't want to take any risk with the basic PDU 9560 because I'll have no way of monitoring my usage in wattage.

And since I'm expanding my operations gradually, smaller home farm going in my former office. I'm still sticking with the 4U case and will be sticking them into the 42U, hopefully I can mount at least 8 4U case in there. All 4U case are fitted with 7 GPU's; the scarcity in GPU's have slowed me down as I'm not going to pay over MSRP for them.

Got all my Ethash Algo build done already with RX570s now trying to pick up a lot of Nvidia cards. For Veteran miners, will anyone of you support 8 1060 4U case build over 7 1070's in the same build. I do understand that with the 1070, they are bulkier and the Rosewill L4500 4U case might not fit 7 of those.

To add, I also like to keep about 3/4 to 1/2 inch space between each GPU in the case to allow room for better air circulation, coupled with the 3700rpm 153cfm Delta fans upfront for intake and 1500rpm 82cfm AeroCool Sharkfin fans right behind the GPUs to help exhaust the all the way back, where I left the 2 80mm fans to help exhaust stale hot air that might hover around the CPU.

Any advice with this setup would be appreciate as I'm not familiar with Nvidia cards but being learning a lot from all Philipma's thread. I'm scared I might not be able to fit up to 20 1070's or more on one 30A breaker, which should be 5760 Watts. Hence why I'm looking at the 1060's as because my goal is to get the best Sols/Wattage per rig.

[molivil]

Very low maintenance, ever since I added automatic restart scripts on a 24 hour interval and watchdog scripts to restart and log automatically whenever it notices 0% GPU use or network disconnection. I take a little hit while it restarts though.

Care to share more about these scripts?  I'd find that very helpful.  Thanks!

I have a script called autoexec.cmd start on logon using the Task Scheduler. Autoexec.cmd is a little take on the good times of the past. It will start the miner program and watchdog. I choose the miner program I want to use by commenting out the others using the REM command. After it has started the miner, it will run C:\Tools\watchdog.cmd which will start monitoring that the system is on the network and the GPU's are doing meaningful work. The Autoexec.cmd script will then hibernate (but still stay running) for 24 hours before it restarts the system. All this while watchdog is on another command line session monitoring the system.

I've written these scripts to work on Nvidia mining rigs. They can be freely copied and modified as needed.

C:\Tools\autoexec.cmd

Code:

@echo off
echo  .---------------------.
echo /  INIT SYSTEM SCRIPT  /
echo *=====================*
echo.
echo %COMPUTERNAME%
echo.
echo Waiting for WIFI to wake up...
timeout 11

REM -----------------------
REM ETHEREUM
REM call C:\Tools\NVIOC\oc_eth.cmd
REM start C:\Tools\Miners\ethminer-0.11.0.rc1-Win64-nanopool\start_opencl-molivil.bat
REM -----------------------

REM -----------------------
REM NICEHASH
REM OC settings set by startup script
start C:\Tools\Miners\NiceHashMiner\start.cmd
REM -----------------------

REM -----------------------
REM ZCASH
REM call C:\Tools\NVIOC\oc_zec_effic.cmd
REM call C:\Tools\NVIOC\oc_zec_max.cmd
REM start C:\Tools\Miners\Zec.miner.0.3.4b-nanopool\start.bat
REM -----------------------

start C:\Tools\watchdog.cmd

FOR /L %%G IN (22,-1,2) DO (
  echo Time before restart: %%G hours
  "C:\Program Files\NVIDIA Corporation\NVSMI\nvidia-smi.exe" -q -d temperature |find "GPU Current Temp"
  timeout 3600 /nobreak >nul
)

echo Time before restart: 1 hr
timeout 1800 /nobreak >nul
echo Time before restart: 30 mins
timeout 1500 /nobreak >nul
echo Time before restart: 5 mins
timeout 300 /nobreak >nul
echo Restarting now!
timeout 30 /nobreak >nul
shutdown /r

Below is a (VERY CRUDE! ) script that logs every bootup and tries to ping preconfigured IP addresses 8.8.4.4 or 8.8.8.8 every 30 minutes. It will also use the NVIDIA configuration utility that comes with drivers called "nvidia-smi.exe" which tells the watchdog if any GPU is at 0% use. It will try a couple of times and restart if any GPU stays at 0%. Any errors that occur it logs. It's a bit botched together, but works great in my use.

C:\Tools\watchdog.cmd

Code:

@echo off
cls
set VERSION=1.6.4
rem ---
rem Watchdog logfile. Watchdog will log system (re)boots and connection failures to this file
set LOGFILE=C:\Tools\watchdog.log
rem ---
rem IP addresses to test. Default is Google DNS server, 8.8.4.4 and 8.8.8.8
set IPTEST=8.8.4.4
set IPTEST2=8.8.8.8
rem ---
rem Watchdog timer. How many seconds to wait between checks.
rem Default is 1800 seconds (30 minutes)
set WDTIMER=1800
rem ---
rem Watchdog timer to retry. How many seconds to wait to retry after a failed attempt.
rem Default is 120 seconds (2 minutes)
set WDTIMERRETRY=120
rem ---

echo ----------------------
echo SYSTEM WATCHDOG v%VERSION%
echo ----------------------
echo.
echo - Will check connectivity to %IPTEST% and %IPTEST2%.
echo - Will check Graphics card utilization.
echo.
echo Checking every %WDTIMER% seconds.
echo.
echo %DATE%-%TIME% Watchdog: Watchdog v%VERSION% started. >> %LOGFILE%

:begin
timeout %WDTIMER% >nul
"C:\Program Files\NVIDIA Corporation\NVSMI\nvidia-smi.exe" -q -d Utilization |find "Gpu" |find " 0 %" && goto GPUError1
rem echo %DATE%-%TIME% Watchdog: GPU test successful.

ping -n 1 %IPTEST% | find "TTL=" >nul
if %errorlevel% == 0 (
  rem echo %DATE%-%TIME% Watchdog: connection test successful.
) else (
  echo %DATE%-%TIME% Watchdog: Connection test FAILED to %IPTEST%
  echo %DATE%-%TIME% Watchdog: Connection test FAILED to %IPTEST% >> %LOGFILE%
  echo Retrying in %WDTIMERRETRY% seconds.
  timeout %WDTIMERRETRY% >nul
  goto retry
)
echo %DATE%-%TIME% Watchdog: All good!
goto begin

:retry
ping -n 1 %IPTEST2% | find "TTL=" >nul
if %errorlevel% == 0 (
  echo %DATE%-%TIME% Watchdog: Second connection test successful.
  echo %DATE%-%TIME% Watchdog: Second connection test successful. Watchdog timer reset. >> %LOGFILE%
  echo Watchdog timer reset.
  goto begin
) else (
  echo %DATE%-%TIME% Watchdog: Second connection test FAILED to %IPTEST2%
  echo %DATE%-%TIME% Watchdog: Second connection test FAILED to %IPTEST2% >> %LOGFILE%
  echo Restarting...
  echo %DATE%-%TIME% Watchdog: Restarting system due to two connection failed attempts >> %LOGFILE%
  goto restart
)

:GPUError1
echo %DATE%-%TIME% Watchdog: GPU IDLE error first attempt.
echo %DATE%-%TIME% Watchdog: GPU IDLE error first attempt. >> %LOGFILE%
echo Retrying in %WDTIMERRETRY% seconds.
timeout %WDTIMERRETRY% >nul
"C:\Program Files\NVIDIA Corporation\NVSMI\nvidia-smi.exe" -q -d Utilization |find "Gpu" |find " 0 %" && goto GPUError2
echo %DATE%-%TIME% Watchdog: GPU test successful.
echo Watchdog timer reset.
goto begin

:GPUError2
echo %DATE%-%TIME% Watchdog: GPU IDLE error second attempt.
echo %DATE%-%TIME% Watchdog: GPU IDLE error second attempt. >> %LOGFILE%
echo Retrying in %WDTIMERRETRY% seconds.
timeout %WDTIMERRETRY% >nul
"C:\Program Files\NVIDIA Corporation\NVSMI\nvidia-smi.exe" -q -d Utilization |find "Gpu" |find " 0 %" && goto GPUError3
echo %DATE%-%TIME% Watchdog: GPU test successful.
echo Watchdog timer reset.
goto begin

:GPUError3
echo %DATE%-%TIME% Watchdog: GPU IDLE error third attempt. 
echo %DATE%-%TIME% Watchdog: GPU IDLE error third attempt. >> %LOGFILE%
echo Restarting...
echo %DATE%-%TIME% Watchdog: Restarting system due to GPU underutilization. >> %LOGFILE%
goto restart

:restart
timeout 60 >nul
shutdown /r

The below example script will set the overclock settings for 5x GTX 1060 graphics cards on the console using nvidia profile inspector. It is called from autoexec.cmd, so that the overclock settings are set at boot time. Same functionality can be achieved with Nvidia Afterburner, but one of the Nicehash miners would override/reset my OC settings, so I started using scripts instead. Additional bonus is, that I can run these in a timed loop every hour to "remind" the miner computer of the correct OC settings.

oc_zec_effic.cmd

Code:

@echo off
echo OC: ZCash mining (max efficiency)
echo.
C:\Tools\NVIOC\nvidiaInspector.exe -setBaseClockOffset:0,0,50 -setMemoryClockOffset:0,0,220 -setPowerTarget:0,68
C:\Tools\NVIOC\nvidiaInspector.exe -setBaseClockOffset:1,0,50 -setMemoryClockOffset:1,0,220 -setPowerTarget:1,70
C:\Tools\NVIOC\nvidiaInspector.exe -setBaseClockOffset:2,0,50 -setMemoryClockOffset:2,0,220 -setPowerTarget:2,72
C:\Tools\NVIOC\nvidiaInspector.exe -setBaseClockOffset:3,0,50 -setMemoryClockOffset:3,0,220 -setPowerTarget:3,75
C:\Tools\NVIOC\nvidiaInspector.exe -setBaseClockOffset:4,0,50 -setMemoryClockOffset:4,0,700 -setPowerTarget:4,76