GTX 1080+1080TI won't dual mine even with new kernel update

[Cheesypoofs16]

Is there something I'm missing? Some kind of extra argument to add to the config in HiveOS? Pascal still doesn't work. (Also RX 580 8GB cards not dual mining either)

[Tof37]

Confirm for 1080ti : still no dual mining in 1.44 and lost 3 mh (it's not a big problem) on etash with the same settings . Can't remember if dag was fully generated in 1.43 but in 1.44 it is. 1080ti shares are not shown any more in statistics (it mines ETH)

[adjustinthings]

HiveOS with 1.44 dual mining with my 1080ti's BUT its random which ones initiate dual mining ETH + TON. Nvidia 470.74 drivers

[Cheesypoofs16]

Finally got 1 1080TI to work, but that's it. :/ bummer

[dk7988]

i have it working on rig #1 but will on duel mine on rig #2 with 4gb 480's and 580's. on rig#2 trying to mine eth+ton with 8gb 580 + 1080ti

[dk7988]

also try removing the power limit you have set on you 1080 and 1080Ti cards and maybe try bumping up your core

[dk7988]

for rig#2 yea, agreed for rig#1 doesn't matter.. coin#1 is paying the power bill for coin#2

[dk7988]

i figured out the solution to my problem.. not sure if it's the same as the original post/question but incase any one else is having problems duel mining with Pascal cards and Polaris 8gb and 4gb cards here ya go.. it takes about 45 seconds to a minute for the my 1080Ti to show a Ton hashrate so heads up if your reading this and wondering why your hashreat isn't showing just give it a minute

[adjustinthings]

I tried the thingy with the *'s and it didnt work. It was probably just the fact that your reset the miner that it worked.

for rig#2 yea, agreed for rig#1 doesn't matter.. coin#1 is paying the power bill for coin#2

Assuming energy cost of 11.5 cents per kilowatt/h 1080ti @ 200W Eth only = 42.6 mh = $1.86R $1.31P 1080ti @ 210W ETHTON = 38.4mh + 0.629gh = $2.02R $1.44P 1080ti @ 215W ETHTON = 39.7mh + 0.635gh = $2.08R $1.49P 1080ti @ 275W ETHTON = 43mh + 0.859gh = $2.34Revenue $1.57Profit

that 70watts per card gets you an extra 13cents a day. I guess thats fine if you only have a couple cards.

but i played settings for my 1080ti asus strix mining ETH only: 43.00mh @ 220w = 1.28$ profit 41.78mh @ 200w = 1.28$ profit 41.20mh @ 195w = 1.27$ profit 41.00mh @ 190w = 1.27$ profit just gotta wait and see if it runs stable. All my other cards are eVGA dual fans.

[stanchiam]

that amount of extra heat may shorten the life of the card and power supply

[dk7988]

not 100% sure, all i know is i have 2x - 13 card rigs and 1x - 11 card all running eth + ton in one instance (per rig) of lol miner (these are amd and nvidia rigs). with that said none of them are having problems running, and i can't see any lol miner limitation for an all nvidia rig/instance problem. i would double check your psu's and wiring, then play with your OC settings

[dk7988]

maybe, at the moment 2 of the 4 1080Ti's are water cooled and i have to put the water block back on the third one so for the most part i'm not to worried about the extra heat. the extra heat in the psu maybe of some concern but seeing how ton is going to be un-minable soon i'm not too worried about it.

As the profitability of ton lowers i'll probably adjust the settings to something more power efficient.

[Messa1]

Work if you rise the PL - at the screenshot i just running with no PL. Need to test how far i can go back.

[adjustinthings]

I first tried to unlock the PL. But it was the same. I thought maybe there was a power issue so i lowered the PLs to 195 and ran them stable for a whlie (both the ETHTON and ETH cards). I reset again and still only a total of 8 GPUs are dual mining. Again, the 20X and 30X cards all initiate DUAL mining and 2 other random 1080ti's are also set for dual mining. I have no additional parameters set. Everything is running smooth, LA is <=1, 4.6gb ram free, 36c cpu temp. I've got 2X2kw power supplies. I updated the nvidia driver 495.46. Using the latest 1.44 with newest hiveOS.

DUAL MINING most efficient at 1.46$ profit with 195 watts VS 215 watts making $1.49 profit /24 hours.

EDIT: SOrry mods I know now this should be in the Discussions tabs.

[Messa1]

Also PL is a bit difficulty to set. If i at 220 W it goes down to 195 W (in Software) if i do 200 W it goes down to 164 W (in Software)

[dk7988]

if your wiring and psu's are fine and nothing is melted only other thing i've found that worked when i was having similar problems with my 3060's, 3080's and 6800's was to play with the core and memory clocks i usually found raising core and removing the PL helped to get all the problem cards (random or otherwise) up and duel mining. other than that i got nothing for ya man, the mining code works both 1.43 and 1.44 versions

[jgonzis]

I'm definitely not saying 1.44 doesnt work. I was using 1.43 for several days without the 1080ti's dual mining ETH+TON. I upgraded to 1.44 this afternoon when HiveOS updated. Only 2 of the 7x 1080ti's work at any one time.

Problem is the PowerLimit... Pascal is quite old GPU, performance is complicated and need it lot of power. You can also use --maxdualimpact 6 to generate more TON

[adjustinthings]

Ok i've got actual useful info. From my last reset:

Looks pretty random again. So i looked at the logs again and saw that the order that the DAG verifications passes went 4, 8, 5, 0, 3, 7, 1, 2, 9, 11, 10, 6*, 12. GPUS 4,8,5,0,3,7 all booted fine with dualton. 6 did as well but did actually had a DAG verification error and didnt finish the smart repair for about 15 sconds

I think it has something to do with all 13 cards trying to finish their DAG verification before TON stuff starts happening. As seen here:

The GPUs with the green arrows all dualmine. The ones that finish their dag a few moments after the first TON job are not initialized. (GPU 6 which did end up working last was the exception)

Are there any options to delay the TON mining a little while until the DAGs are all completed. Is there a way to see more verbose logs? I'm using motd and miner.

(deleted a few old comments to clean up the thread a bit)

[jgonzis]

Ok i've got actual useful info. From my last reset:

Looks pretty random again. So i looked at the logs again and saw that the order that the DAG verifications passes went 4, 8, 5, 0, 3, 7, 1, 2, 9, 11, 10, 6*, 12. GPUS 4,8,5,0,3,7 all booted fine with dualton. 6 did as well but did actually had a DAG verification error and didnt finish the smart repair for about 15 sconds

I think it has something to do with all 13 cards trying to finish their DAG verification before TON stuff starts happening. As seen here:

The GPUs with the green arrows all dualmine. The ones that finish their dag a few moments after the first TON job are not initialized. (GPU 6 which did end up working last was the exception)

Are there any options to delay the TON mining a little while until the DAGs are all completed. Is there a way to see more verbose logs? I'm using motd and miner.

(deleted a few old comments to clean up the thread a bit)

Could you put PL to 250?... and restart?

I think is all Watts Problem, Pascal don't have space with lower Watts. Could you try it please??

[Tof37]

Can confirm the wattage related problem. Got only one 1080ti but setting PL to 0 makes dual mining work.

[jgonzis]

Can confirm the wattage related problem. Got only one 1080ti but setting PL to 0 makes dual mining work.

Thx. Problem is the dag limitations of Pascal use lot of core don't giving space to TON, with ETC+TON that doesn't happen so much DAG is smaller...

[adjustinthings]

Dude, i dual mine with the 1080ti's that initiate dual mode perfectly stable at 195. its not a watt issue. I did like you asked and set the PL to zero and the effect is exactly the same. I can mine ETH only stable at 210 watts each. This isn't a watts issue. I just showed you it correlated to the order the cards are being initialized.

[Tof37]

Can confirm the wattage related problem. Got only one 1080ti but setting PL to 0 makes dual mining work.

Thx. Problem is the dag limitations of Pascal use lot of core don't giving space to TON, with ETC+TON that doesn't happen so much DAG is smaller...

I see ...so not related with the miner itself. Will look how much rentability Is affected

[jgonzis]

Dude, i dual mine with the 1080ti's that initiate dual mode perfectly stable at 195. its not a watt issue. I did like you asked and set the PL to zero and the effect is exactly the same. Stop pretending like you've solved this and high fiving yourself. I can mine ETH only stable at 210 watts each. This isn't a watts issue. I just showed you it correlated to the boot order.

Did you restart after putting PL to 0?

[adjustinthings]

that was after a restart of the miner. this is after a reset of the system. You may think that GPU 1,2 initiazed TON but its a tiny amount and not actually working. Essentially the same:

[jgonzis]

that was after a restart of the miner. this is after a reset of the system. You may think that GPU 1,2 initiazed TON but its a tiny amount and not actually working. Essentially the same:

Try adding:

--disable-dag-verify

And if still doesn't work: --maxdualimpact 6

[adjustinthings]

Tried both those. No difference. Also tried the dag delay option. Hmmmm, is it the pool maybe? Maybe this pool only does so many simultaneous connects?

[lwYeo]

Ok i've got actual useful info. From my last reset:

Looks pretty random again. So i looked at the logs again and saw that the order that the DAG verifications passes went 4, 8, 5, 0, 3, 7, 1, 2, 9, 11, 10, 6*, 12. GPUS 4,8,5,0,3,7 all booted fine with dualton. 6 did as well but did actually had a DAG verification error and didnt finish the smart repair for about 15 sconds

I think it has something to do with all 13 cards trying to finish their DAG verification before TON stuff starts happening. As seen here:

The GPUs with the green arrows all dualmine. The ones that finish their dag a few moments after the first TON job are not initialized. (GPU 6 which did end up working last was the exception)

Are there any options to delay the TON mining a little while until the DAGs are all completed. Is there a way to see more verbose logs? I'm using motd and miner.

(deleted a few old comments to clean up the thread a bit)

I think this might be true in my case too, my 1080 usually finish DAG creation the last (while 1080Ti is faster), and it usually didn't get to dual mine. Had to restart my rig literally a "Ton" of times to get it mining, like a lottery.

[adjustinthings]

I think this might be true in my case too, my 1080 usually finish DAG creation the last (while 1080Ti is faster), and it usually didn't get to dual mine. Had to restart my rig literally a "Ton" of times to get it mining, like a lottery.

what pool are you using? icemining.ca ?

[lwYeo]

I think this might be true in my case too, my 1080 usually finish DAG creation the last (while 1080Ti is faster), and it usually didn't get to dual mine. Had to restart my rig literally a "Ton" of times to get it mining, like a lottery.

what pool are you using? icemining.ca ?

ethermine/toncoinpool.io

[adjustinthings]

I think I have a clue.

When the miner restarts, the ETH mh/s are displayed immediately cuz eth is fastboy. It shows full ETH hashrate momentarily:

30 seconds later when the screen refreshes, the cards that manage to get ton going have their rates reduced:

But of course they do. That wasn't my point though. My point was that the cards were already hashing ETH at full power before the TON was applied. I should point out also that I ran this rig with TON only for two weeks straight last month without it even flinching.

[Cheesypoofs16]

So it seems to me like this is actually an issue that needs to be fixed by the devs

[jgonzis]

I think I have a clue.

When the miner restarts, the ETH mh/s are displayed immediately cuz eth is fastboy. It shows full ETH hashrate momentarily:

30 seconds later when the screen refreshes, the cards that manage to get ton going have their rates reduced:

But of course they do. That wasn't my point though. My point was that the cards were already hashing ETH at full power before the TON was applied. I should point out also that I ran this rig with TON only for two weeks straight last month without it even flinching.

Do you have Delay for the ETH PILL?

[adjustinthings]

No I don't have a delay for the ETH pill. What parameter is that? I also tried using the DAG delay option but there was no difference with values ranging from 1-1000.

Its POSSIBLE that you could be right about the power issue. I can't rule it out 100%. I noticed when I had unlocked the PL a while back, the power only went up to 235 on the 1080ti's.

So if others are not experiencing this also. I will further check the power supply side of the issue. But this rig was running most of last month mining TON only with the 1080ti's at 210w each.

[Roman-2021]

:)

[Egyptianminer]

I believe since this is an experimental update for pascal cards which is the lolminer 1.44 we should be receiving a new update with stable performance that will fix all the above

[cednats]

Same issue here on my 1080. As far as I can tell it's the ethlargement pill that's causing the issue. If I start without the pill the 1080 will mine Ton. If I turn off the pill after it's mining it'll start mining small amounts of Ton while keeping the same Eth hashrate. Still working on finding a fix.

[fenomentn]

me to i have the same problem. my rig keep restarting because 1080ti stability i found turning off the ethpill make it gain more ton but not eth so i wait my card got stable then i turn on ethpill maybe 90 or 120 delay fix the problem ... last uptime 22h hope not keep restart mining

[cednats]

I set the ethlargement pill to -120 delay and -60 on overclock delay and now it seems to be working. 34 mh/s (down from 35 with Eth only) and 306 mh/s on ton.

[adjustinthings]

Can confirm that delaying the Eth pill I've been able to get all my 1080ti's running ETH+TON. I tried various settings and the only things that really seems to matter is that the DAGs all need to finish verifying before the Eth pill starts. 60 initially worked but when I reset there was one card that wasn't dual mining and I checked the log and it had finished repairing the DAG last and therefore didnt initialize the dualmining that time. So i increased it to 90 to make sure it gets all the cards.

[Cheesypoofs16]

I set the ethlargement pill to -120 delay and -60 on overclock delay and now it seems to be working. 34 mh/s (down from 35 with Eth only) and 306 mh/s on ton.

This worked great for me. Thank you for solving the issue :D

[jgonzis]

Can confirm that delaying the Eth pill I've been able to get all my 1080ti's running ETH+TON. I tried various settings and the only things that really seems to matter is that the DAGs all need to finish verifying before the Eth pill starts. 60 initially worked but when I reset there was one card that wasn't dual mining and I checked the log and it had finished repairing the DAG last and therefore didnt initialize the dualmining that time. So i increased it to 90 to make sure it gets all the cards.

Many thanks for the information... so that means need to be pill applied after ETH+TON both are started.

I will use your information to other miners with problems. Many thanks!!

[adjustinthings]

Thanks for the help. I have a little extra information to help you as well. I noticed my rig drop some TON mining on some cards. I checked the logs and saw that there was a new EPOCH for ETH and when it updated the DAG for the cards it hadn't again disabled the pill during that time. So i ended up with the previous problem. Miner restart fixed it though, obviously.

[jgonzis]

Thanks for the help. I have a little extra information to help you as well. I noticed my rig drop some TON mining on some cards. I checked the logs and saw that there was a new EPOCH for ETH and when it updated the DAG for the cards it hadn't again disabled the pill during that time. So i ended up with the previous problem. Miner restart fixed it though, obviously.

Completely right more than 12 hours ago...

[Rdgordon81]

For anyone else having issues, I decided to dig in and see why the 1080ti was having such miserable numbers when dual mining. I've got a watercooled 1080ti, but tried to find a setting stock cards could emulate, and currently getting 35-40 pool hash in eth, and 1k-1.3k pool hash in ton.

Step 1. Think your dual mine impact is too high? It's not. My rx5700 seems to love it at 0.25, I got so annoyed at the 1080ti I just set it to 10, and immediately jumped from 700 to over 1k ton hash. Eth is at 36mH at the terminal, but I might adjust down slightly to zero in.

Step 2. Turn off ethpill to start mining. This immediately gave me a respectable ton hash, went from 250mh to 800mh. Once your numbers are stable, turn it on. It will bump your eth up 3 or so Mh, I also saw a small bump in ton, but i can't guarantee that it's related, could be something else in the background, I was making lots of changes.

Step 3. Don't go berserk on your OC. That's what causes the DAG breakup. If you can't load the dag, drop your OC to something like +100/+100.

I'm running a very conservative tune, and getting ~35mh Eth and well over 1Gh ton.

Hope this helps.

[jgonzis]

For anyone else having issues, I decided to dig in and see why the 1080ti was having such miserable numbers when dual mining. I've got a watercooled 1080ti, but tried to find a setting stock cards could emulate, and currently getting 35-40 pool hash in eth, and 1k-1.3k pool hash in ton.

Step 1. Think your dual mine impact is too high? It's not. My rx5700 seems to love it at 0.25, I got so annoyed at the 1080ti I just set it to 10, and immediately jumped from 700 to over 1k ton hash. Eth is at 36mH at the terminal, but I might adjust down slightly to zero in.

Step 2. Turn off ethpill to start mining. This immediately gave me a respectable ton hash, went from 250mh to 800mh. Once your numbers are stable, turn it on. It will bump your eth up 3 or so Mh, I also saw a small bump in ton, but i can't guarantee that it's related, could be something else in the background, I was making lots of changes.

Step 3. Don't go berserk on your OC. That's what causes the DAG breakup. If you can't load the dag, drop your OC to something like +100/+100.

I'm running a very conservative tune, and getting ~35mh Eth and well over 1Gh ton.

Hope this helps.

Really many thanks for the informative. Also you increase Power Limit?

[Rdgordon81]

Step 1. Think your dual mine impact is too high? It's not. My rx5700 seems to love it at 0.25, I got so annoyed at the 1080ti I just set it to 10, and immediately jumped from 700 to over 1k ton hash. Eth is at 36mH at the terminal, but I might adjust down slightly to zero in. Step 2. Turn off ethpill to start mining. This immediately gave me a respectable ton hash, went from 250mh to 800mh. Once your numbers are stable, turn it on. It will bump your eth up 3 or so Mh, I also saw a small bump in ton, but i can't guarantee that it's related, could be something else in the background, I was making lots of changes. Step 3. Don't go berserk on your OC. That's what causes the DAG breakup. If you can't load the dag, drop your OC to something like +100/+100. I'm running a very conservative tune, and getting ~35mh Eth and well over 1Gh ton. Hope this helps.

Really many thanks for the informative. Also you increase Power Limit?

This card was mining flux before I switched over, so I really don't have a reference for power limits vs. 1080ti->eth mining. My rx5700 went from 104w to 158w once i tuned it back, so yeah, watts/performance is an almost linear increase, since your core clock is watt limited. Set your overclock to +1000, PL to 120W and youll see what i mean. Your clock will be at 1300 or something crazy low. Mine obviously runs very cold, its pushing 250w @ 32C right now. I had it crash a couple times after ~2hrs tonight, so ive dialed it back a little further, But still getting 38/950 at the terminal. I should also add that this card is unusual in that I got it out of a literal scrap heap, and have had to soldered it back together as it had several loose components on the PCB, and the heatsink was damaged. Its got a CPU AIO on it, but it definitely doesnt like getting hammered on, ill probably fix it properly with solder paste and my air solder when i get it set back up, but i had to fix it with a horrible hand soldering iron with like a 3mm tip. So I can't really push it much over 280 for now. Basically if this will mine 40/1000 then any 1080ti should be able to. ![Uploading 16448987422632313606696508028178.jpg…]()

[Rdgordon81]

I will say I've learned quite a few things about GPU intensive algos from playing a lot with the newer Equihash variants. Here's a quick, basic list of things i do with the 1080ti to get very good, multi-day uptime performance (70-80sol average on flux the past few weeks with this card). This is all on windows 10.

Download and use Nvidia Profile Inspector. If you don't want to spent days googling every option, I don't blame you, it's worth it though. The one setting that is critical, since I don't really trust commandline mem control anyways, is deactivating P2. It'll be a setting maybe 3/4 of the way down. Deactivate, apply, and you're good. This locks your card in P0 on the driver level, you can now tune your card at idle without p0 spikes just bouncing your mem clock off the limiter.

Power is king. If you haven't re-pasted your processor and Inspected memory heat management, do it. Buy the best paste you can, and get it right. And don't kill the card with watts. Ton (and equihash) cares about gpu speed, not heat created. So set your power super low, clock very high, and inch up power until you either crash, or hash increase slows down. Then move power up a couple %, clock down maybe 25mhz from failure, and you'll probably be around 1930 and 250, if yours is anything like mine. Either way, above 2k I lose stability with this beat up card, can't imagine a good condition one couldn't keep up.

Good luck all!

[Egyptianminer]

Anyone know if lolminer are planning an update to fix this instead of us trying various fixes that is abviously not working consistently for everyone

On Tue, Feb 15, 2022, 06:39 Rdgordon81 @.***> wrote:

I will say I've learned quite a few things about GPU intensive algos from playing a lot with the newer Equihash variants. Here's a quick, basic list of things i do with the 1080ti to get very good, multi-day uptime performance (70-80sol average on flux the past few weeks with this card). This is all on windows 10.

Download and use Nvidia Profile Inspector. If you don't want to spent days googling every option, I don't blame you, it's worth it though. The one setting that is critical, since I don't really trust commandline mem control anyways, is deactivating P2. It'll be a setting maybe 3/4 of the way down. Deactivate, apply, and you're good. This locks your card in P0 on the driver level, you can now tune your card at idle without p0 spikes just bouncing your mem clock off the limiter.

Power is king. If you haven't re-pasted your processor and Inspected memory heat management, do it. Buy the best paste you can, and get it right. And don't kill the card with watts. Ton (and equihash) cares about gpu speed, not heat created. So set your power super low, clock very high, and inch up power until you either crash, or hash increase slows down. Then move power up a couple %, clock down maybe 25mhz from failure, and you'll probably be around 1930 and 250, if yours is anything like mine. Either way, above 2k I lose stability with this beat up card, can't imagine a good condition one couldn't keep up.

Good luck all!

— Reply to this email directly, view it on GitHub https://github.com/Lolliedieb/lolMiner-releases/issues/1171#issuecomment-1039850764, or unsubscribe https://github.com/notifications/unsubscribe-auth/AXUAUVKX6CR2OJBRTTJFSYTU3HKH3ANCNFSM5N34NS5Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

You are receiving this because you commented.Message ID: @.***>

[jgonzis]

Anyone know if lolminer are planning an update to fix this instead of us trying various fixes that is abviously not working consistently for everyone … On Tue, Feb 15, 2022, 06:39 Rdgordon81 @.> wrote: I will say I've learned quite a few things about GPU intensive algos from playing a lot with the newer Equihash variants. Here's a quick, basic list of things i do with the 1080ti to get very good, multi-day uptime performance (70-80sol average on flux the past few weeks with this card). This is all on windows 10. Download and use Nvidia Profile Inspector. If you don't want to spent days googling every option, I don't blame you, it's worth it though. The one setting that is critical, since I don't really trust commandline mem control anyways, is deactivating P2. It'll be a setting maybe 3/4 of the way down. Deactivate, apply, and you're good. This locks your card in P0 on the driver level, you can now tune your card at idle without p0 spikes just bouncing your mem clock off the limiter. Power is king. If you haven't re-pasted your processor and Inspected memory heat management, do it. Buy the best paste you can, and get it right. And don't kill the card with watts. Ton (and equihash) cares about gpu speed, not heat created. So set your power super low, clock very high, and inch up power until you either crash, or hash increase slows down. Then move power up a couple %, clock down maybe 25mhz from failure, and you'll probably be around 1930 and 250, if yours is anything like mine. Either way, above 2k I lose stability with this beat up card, can't imagine a good condition one couldn't keep up. Good luck all! — Reply to this email directly, view it on GitHub <#1171 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AXUAUVKX6CR2OJBRTTJFSYTU3HKH3ANCNFSM5N34NS5Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub. You are receiving this because you commented.Message ID: @.>

1.44 --- Added experimental Ethash + Ton dual mining kernels for Nvidia Pascal generation GPUs. Remember each EPOCH Change that will go down ... it is Pascal old arquitecture. And that cannot be solved in anyway. TLB size Problem and no option to improve. Some have found easy solution with a Mining OS. It seems with windows and the ETHPILL of Windows is more complicated. What it is true is that Power is need. Seen the result of the of the other miner 260W for doing a nice Dual Mining.