1080Ti Specific - Best mining option

[naresh963]

Normally when a miner starts, gpu boost 3.0 boosts the core clock to 2025-2075mhz and when miner quits it comes down to original state, but while cooldown the gpu gets boosted even more for a split second causing the nvidia driver to crash. the problem here is that when start or quit, there is not enough power ( i mean software/driver limit). All gpu follow a base profile ( _GLOBAL_DRIVER_PROFILE (Base Profile) ) here there is a key that states ' CUDA - Force P2 State' its ON by default. you just need to change the state to OFF.

You can check the power state by command-line to get current state

Code:

nvidia-smi -q -d PERFORMANCE

[1715199328]

1715199328

[1715199328]

1715199328

[1715199328]

1715199328

[1715199328]

1715199328

[1715199328]

1715199328

[joshuajones02]

ran out of ideas, installing windows 10 now... will report back results


Issue recap -
4 1080ti GPU's run great/fine on windows 7

Installed 5th, wouldnt work, code 12 error. Changed bios settings shown above and they all showed up and will run, but will run like complete shit. Power % shown all over the place, and some cards capping out at 200 or 400 sol/s when running Equihash or 20-30mh/s with skunk..

Now attempting to install windows 10 pro x64 and ensuring nvidia drivers are up to date



Edit: Windows 10 install did not help at all..
Ran each card individually, with their own riser and cords and they worked on par with the 680-710 sol/s, but the moment I go from 4 to 5 it runs worse than if I had 3

Edit: Now I can't get 4 cards running, 2 on each PSU... originally had 4 on 1 PSU.. feel like selling alll this crap now lol

In my experience that could be motherboard. Some times the same motherboard need different settings.

You need to play with dmi-pci settings in bios. Try this:

1) dmi speed gen 1, pci speed auto
2) dmi speed gen 1, pci speed gen 1

also some mother boards got 2 different setting for pci slot. Play with this settings also.

Where would dmi speed fall under? Only see Peg0 and Peg1 speed

[JWT]

Normally when a miner starts, gpu boost 3.0 boosts the core clock to 2025-2075mhz and when miner quits it comes down to original state, but while cooldown the gpu gets boosted even more for a split second causing the nvidia driver to crash. the problem here is that when start or quit, there is not enough power ( i mean software/driver limit). All gpu follow a base profile ( _GLOBAL_DRIVER_PROFILE (Base Profile) ) here there is a key that states ' CUDA - Force P2 State' its ON by default. you just need to change the state to OFF.

You can check the power state by command-line to get current state

Code:

nvidia-smi -q -d PERFORMANCE

It's P2 for all cards.
How do I set P0?
Couldn't find GLOBAL_DRIVER_PROFILE field in register.

[naresh963]

just google 'nvidia profile inspector'. download and run it.

[joshuajones02]

@VoskCoin , looks like you are having the same issue I was.. after hours of troubleshooting it looks like its my motherboard.. two of the PCIe x1 slots are causing issues when they are plugged in to the entire board, but if I avoid those 2 pci e slots (the 2 between the main pci e x16) it works perfect but I cant run all 5 cards. going to call MSI tomorrow

https://imgur.com/a/EGCVE <- the 2 empty pci e x1 slots are the ones causing the issues.. if any GPU is plugged into them it causes all of the GPUs to run like crap and crash.

[papi92]

Did you try to update BIOS?

[CjMapope]

@VoskCoin , looks like you are having the same issue I was.. after hours of troubleshooting it looks like its my motherboard.. two of the PCIe x1 slots are causing issues when they are plugged in to the entire board, but if I avoid those 2 pci e slots (the 2 between the main pci e x16) it works perfect but I cant run all 5 cards. going to call MSI tomorrow

https://imgur.com/a/EGCVE <- the 2 empty pci e x1 slots are the ones causing the issues.. if any GPU is plugged into them it causes all of the GPUs to run like crap and crash.

are you getting systems inturrupts/high cpu usage? check risers and make sure your set to "gen1" in bios. i had similar troubles once on a x51070 rig, turns out gen1 in bios fixed everything, funny how it goes, good luck man!
I had changed out every part to looking for the trouble but it was softwear not hardwear in the end ;p

[joshuajones02]

@VoskCoin , looks like you are having the same issue I was.. after hours of troubleshooting it looks like its my motherboard.. two of the PCIe x1 slots are causing issues when they are plugged in to the entire board, but if I avoid those 2 pci e slots (the 2 between the main pci e x16) it works perfect but I cant run all 5 cards. going to call MSI tomorrow

https://imgur.com/a/EGCVE <- the 2 empty pci e x1 slots are the ones causing the issues.. if any GPU is plugged into them it causes all of the GPUs to run like crap and crash.

are you getting systems inturrupts/high cpu usage? check risers and make sure your set to "gen1" in bios. i had similar troubles once on a x51070 rig, turns out gen1 in bios fixed everything, funny how it goes, good luck man!
I had changed out every part to looking for the trouble but it was softwear not hardwear in the end ;p

yes, that is exactly what was happening (only when a GPU is plugged into those PCIe 1x slots shown in image) . Risers are fine, all the GPUs are working great in their current slots.

Its after 3 am here.. I'll mess around with the PEG0 & PEG1 settings tomorrow

[QuintLeo]

@VoskCoin , looks like you are having the same issue I was.. after hours of troubleshooting it looks like its my motherboard.. two of the PCIe x1 slots are causing issues when they are plugged in to the entire board, but if I avoid those 2 pci e slots (the 2 between the main pci e x16) it works perfect but I cant run all 5 cards. going to call MSI tomorrow

https://imgur.com/a/EGCVE <- the 2 empty pci e x1 slots are the ones causing the issues.. if any GPU is plugged into them it causes all of the GPUs to run like crap and crash.

 Are you running a PVMe type SSD on that motherboard?
 Also, a lot of boards disable or reduce lanes to some *physical* x16 slots if you use one or more of the x1 slots on them.

[DDL]

How are your cards taking all that algo switching?

why they shouldn't?

[bdiddle]

For those of you who are not multi algo mining, what are you finding most profitable?

[DDL]

zec or sig, i think.

[dragonmike]

Where/How do I run nvidia-smi? In on Windows 10 btw. Couldn't find that file anywhere.
Doesn't work from the Windows command line...

[DDL]

Where/How do I run nvidia-smi? In on Windows 10 btw. Couldn't find that file anywhere.
Doesn't work from the Windows command line...

Firstly you should cd in your command line to a nvidia folder (for default it's like C:\Program Files\NVIDIA Corporation\NVSMI), and there run the code naresh963 wrote.

[KaydenC]

For those of you who are not multi algo mining, what are you finding most profitable?

SIGT right now. There are a ton of algos 10-20% less profitable than SIGT, like zec and skein.

[dragonmike]

Where/How do I run nvidia-smi? In on Windows 10 btw. Couldn't find that file anywhere.
Doesn't work from the Windows command line...

Firstly you should cd in your command line to a nvidia folder (for default it's like C:\Program Files\NVIDIA Corporation\NVSMI), and there run the code naresh963 wrote.

Right.
So all my cards are in P2 state... DESPITE me having a nvidiainspector script running every few minutes automatically specifying
-forcepstate:0,0
-forcepstate:1,0
-forcepstate:2,0
-forcepstate:3,0
-forcepstate:4,0
...at the end of each line for my 1080's.

Any suggestions?

[Smontiel]

For those of you who are not multi algo mining, what are you finding most profitable?

SIGT right now. There are a ton of algos 10-20% less profitable than SIGT, like zec and skein.

where are you guys trading your sigts? i lost like 200 sigts when i tried to deposit at cryptopia, my windows wallet is already showing like 400 confirmations but nothing had hapend in my cryptopia wallet so far..

[joshuajones02]

@VoskCoin , looks like you are having the same issue I was.. after hours of troubleshooting it looks like its my motherboard.. two of the PCIe x1 slots are causing issues when they are plugged in to the entire board, but if I avoid those 2 pci e slots (the 2 between the main pci e x16) it works perfect but I cant run all 5 cards. going to call MSI tomorrow

https://imgur.com/a/EGCVE <- the 2 empty pci e x1 slots are the ones causing the issues.. if any GPU is plugged into them it causes all of the GPUs to run like crap and crash.

are you getting systems inturrupts/high cpu usage? check risers and make sure your set to "gen1" in bios. i had similar troubles once on a x51070 rig, turns out gen1 in bios fixed everything, funny how it goes, good luck man!
I had changed out every part to looking for the trouble but it was softwear not hardwear in the end ;p

Alright, finally messed around with the PEG settings and tried Gen1, Gen2, and Gen3... none of them were successful

Installed a card into one of the PCIe slots I was having issues with and changed the settings -

Gen1 - performance would go to crap for all the cards.. running skunk I would get 30mh/s for all the cards instead of 48-50

Gen2- performance was a little better but the cards were all running at 40mh/s instead of the 48-50

Gen3 - back to poor performance



I'm clueless on what to do other than buy a new board.. The board has been flashed to to the most recent firmware, guess I'll head over to MSI's forum and see what someone can tell me what to do.

[joshuajones02]

Those of you guys that mine with 6 1080 ti's, what mobo do you find to be most successful? I'm trashing this MSI and going to buy a new one

[KaydenC]

For those of you who are not multi algo mining, what are you finding most profitable?

SIGT right now. There are a ton of algos 10-20% less profitable than SIGT, like zec and skein.

where are you guys trading your sigts? i lost like 200 sigts when i tried to deposit at cryptopia, my windows wallet is already showing like 400 confirmations but nothing had hapend in my cryptopia wallet so far..

Yobit. Same I lost some coins, like 1.5k coins... Mined for a whole day before realizing the issue. They say they are still syncing their SIGT wallets. Hopefully we can get our coins back.

Those of you guys that mine with 6 1080 ti's, what mobo do you find to be most successful? I'm trashing this MSI and going to buy a new one

I use Biostar Tb 250 BTC. Runs flawlessly with 6x 1080ti.


This is my rig: http://cryptomininghw.com/wp-content/uploads/2017/07/banner2.jpg

These run 5 cards since I couldn't find more 1600w psu. Any mining motherboard should work fine, but 8+ cards and dual psu/server psu requires more knowledge to set up.