[ mining os ] nvoc - page 324. | Bitcointalksearch.org

bibyfok

full member

Activity: 160

Merit: 100

Is there any plan to add a remote management ? Either a webportal or a teamviewer?

Thanks !

stef_stef

full member

Activity: 144

Merit: 100

Quote from: jmooney5115 on June 19, 2017, 04:20:56 PM

Quote from: stef_stef on June 19, 2017, 03:15:17 PM

Anyone to let me know what could be the possible cause of nvOC not booting up?

Often times I get a black screen and it will stay that way for 15+min. So far, the only possible solution to this problem is to reinstall nvOC on my usb

MOBO asus prime z270p
g3930
nvidia 1070

I have the same CPU so it isn't that. Is your display plugged into the onboard video or the GPU? Does the MOBO post?

It could be an issue with the USB flash.

Sorry for my late response.

The HDMI cable is plugged into one of the GPU's that is on the x16 PCIe slot. And yes, the mobo goes past bios, that is when I get my black screen.

As for someone else that quoted me, I did not forget to unzip the file

Nexillus

full member

Activity: 169

Merit: 100

Quote from: UberDaemon on June 20, 2017, 11:43:21 AM

Quote from: Nexillus on June 19, 2017, 06:45:47 PM

I should probably rephrase. I have all 6 GPUs on risers, the only time it does not post is when I have 2 of the x16 slots with the risers in them. I will double check and make sure they are all the same gen type. I had it on auto first but when I did, it wouldn't even boot past one card. I'm to the point I may just reset the MB bios via cmos and start fresh.

Use all 3 of the x16 slots, then begin populating the x1 slots beginning with the one closest to the CPU and moving away from it.

I did not try this but I did get it working.

Motherboard changes to get it to post:
DMI max gen 2
pciex16_1 link gen 2
pciex16_2 link gen 2
pcie spread spectrum clock disable
pcei speed gen 2
restore ac power on
hd audio controler disable
post delay time 0 sec
above 4g decoding enable

Current slots in use (check MB manual)
PCIEX1_1
PCIEX16_1
PCIEX1_2
PCIEX1_3
PCIEX1_4
PCIEX16_3

If I use the PCIEX16_2 slot claymore miner will not detect the GPU in that slot, skipping it posts and all 6 are registered.

Link to the stats
CC: 100
Mem: 1275
dcri: 20
Watt: 115 per card
802 watts from the wall

UberDaemon

newbie

Activity: 51

Merit: 0

Quote from: ?? on ??

Just wanted to say thanks for making this! Happy to donate if you have an address to send it to.

Refer to the untouched OneBash file for OP's wallet addresses to donate, or you can donate by running your rig with his address(es) still listed in the OneBash file.

UberDaemon

newbie

Activity: 51

Merit: 0

Quote from: Nexillus on June 19, 2017, 06:45:47 PM

I should probably rephrase. I have all 6 GPUs on risers, the only time it does not post is when I have 2 of the x16 slots with the risers in them. I will double check and make sure they are all the same gen type. I had it on auto first but when I did, it wouldn't even boot past one card. I'm to the point I may just reset the MB bios via cmos and start fresh.

Use all 3 of the x16 slots, then begin populating the x1 slots beginning with the one closest to the CPU and moving away from it.

https://i.imgur.com/lytgMw4.jpg

Maxximus007

full member

Activity: 153

Merit: 100

Quote from: cryptosharks on June 20, 2017, 08:39:32 AM

Quote from: Maxximus007 on June 20, 2017, 08:12:21 AM

Quote from: cryptosharks on June 20, 2017, 07:16:27 AM

Can someone share the OC settings for GTX 1080 OC 8GB (non TI), please?

These are my OC settings for nvOC:

POWERLIMIT="NO"
__CORE_OVERCLOCK=275
MEMORY_OVERCLOCK=600
FAN_SPEED=100

The card gets heated: 82 C (stays there 24/7 consistently)

With the above OC settings, I get 315-330 MH/s on LBC and 495-505 Sol/s on ZEC.

With the same OC settings as on nvOC, Windows 10 is able to get 345-365 MH/s on LBC and 545-565 Sol/s on ZEC with temperature not going over 79 C. I used MSI Afterburner on Windows.

Any ideas?

Well, in nvOC (Linux) the OC is different. So you might want to up your mem OC quite a bit.

As written before by fullzero:
Powerlimits in windows are in percent TDP; in linux they are in watts. Thus a powerlimit of 70 in windows is .7 * 120 watts or whatever the TDP is (84 watts if the card TDP is 120); while 70 in linux is 70 watts. I would recommend comparing Windows and linux powerlimits with a kill-a-watt to get them exactly equal; but most likely a 3gb 1060 has a 120 or 125 watt TDP.

Also linux OC offsets are scaled differently than windows; you will need to use higher offsets to get the same results in linux.

Thanks.

How about setting the temperature limit? In MSI Afterburner, I was able to set the temperature limit as well. I've been using Windows so far, so please don't mind my questions.

@fullzero, cheers for creating nvOC and sharing. You gotta start charging

As far as I know, there is no option in Linux / Nvidia drivers to set a max temp. It can be done in EWBF miner, but that would mean it will idle the card until below threshold. You could limit your power to reach the desired temp, but that's not quite the same.

Maxximus007

full member

Activity: 153

Merit: 100

Quote from: xleejohnx on June 20, 2017, 08:19:18 AM

Quote from: Maxximus007 on June 20, 2017, 05:16:44 AM

Quote from: ?? on ??

I like rufus; I have used it for years, it can also be used to image the USB.

I think the hddRawCopy tool, is easier to use for members who have less experience imaging; there are also a lot of members who are already familiar with it from smOS.

I use an img format for the image; specifically to enable the use of most imaging programs.

This one is also easy to use and is available for mac / linux and windows:

https://etcher.io/

Thanks, being on a Mac, etcher looks great. Saves me starting up Virtualbox Wink

Just leaves me one thing: the Windows partition is unreachable with a Mac. Seemingly this has to do with some flags missing for the FAT partition. Anyone has a solution for this?

What are you trying to do?
Explain into details what you trying to do and I might be able to help out
I deal with windows and Mac daily

After writing the image to an USB my Mac won't recognise the USB nor the Win partition on the USB. Tried several utilities but it seems to be an issue with the way the partition table is made. Unfortunately I couldn't find a working solution..
So for me it's easier atm to boot up a rig with the USB, ssh to it and edit the oneBash remotely. But if you have a way to mount, that would be great!

cryptosharks

newbie

Activity: 8

Merit: 0

Quote from: Maxximus007 on June 20, 2017, 08:12:21 AM

Quote from: cryptosharks on June 20, 2017, 07:16:27 AM

Can someone share the OC settings for GTX 1080 OC 8GB (non TI), please?

These are my OC settings for nvOC:

POWERLIMIT="NO"
__CORE_OVERCLOCK=275
MEMORY_OVERCLOCK=600
FAN_SPEED=100

The card gets heated: 82 C (stays there 24/7 consistently)

With the above OC settings, I get 315-330 MH/s on LBC and 495-505 Sol/s on ZEC.

With the same OC settings as on nvOC, Windows 10 is able to get 345-365 MH/s on LBC and 545-565 Sol/s on ZEC with temperature not going over 79 C. I used MSI Afterburner on Windows.

Any ideas?

Well, in nvOC (Linux) the OC is different. So you might want to up your mem OC quite a bit.

As written before by fullzero:
Powerlimits in windows are in percent TDP; in linux they are in watts. Thus a powerlimit of 70 in windows is .7 * 120 watts or whatever the TDP is (84 watts if the card TDP is 120); while 70 in linux is 70 watts. I would recommend comparing Windows and linux powerlimits with a kill-a-watt to get them exactly equal; but most likely a 3gb 1060 has a 120 or 125 watt TDP.

Also linux OC offsets are scaled differently than windows; you will need to use higher offsets to get the same results in linux.

Thanks.

How about setting the temperature limit? In MSI Afterburner, I was able to set the temperature limit as well. I've been using Windows so far, so please don't mind my questions.

@fullzero, cheers for creating nvOC and sharing. You gotta start charging

CarlOrff

sr. member

Activity: 826

Merit: 440

Quote from: weareken on June 19, 2017, 09:09:12 PM

A tip and a question:

Tip: If you're tuning your GPUs to reduce power, nvidia-smi can be quite helpful. I open it in a second ssh session so I can watch the miner output in the main window and the GPU in the second.

You can run it with this (refreshes every 0.1 seconds):

Code:

watch -n0.1 nvidia-smi

Question:

My rig is using about 50W more than it should as compared to same Eth mh/s under windows. I've got cclock set to -35 and mclock set to 820. pl is set to 75. I get ~23mh/s, which is what I'm shooting for. from the nvidia-smi output it looks like the problem is that three of the cards (this rig is 3x PNY and 1xEVGA) are not honoring the powerlimit. Still new to tuning nvidia cards until linux. Any ideas?

Code:

Update:
Apparently quite a few people are experiencing this issue with GTX 1060's https://devtalk.nvidia.com/default/topic/1011804/linux/nvidia-smi-not-fully-supported-on-gtx-1060/. In this case, mine is a PNY 3GB version which someone else is having issues with as well. Unclear as to whether this is a driver bug or somehow related to particular motherboards (a Gigabyte z97 SOC Force in my case).

Tried :
sudo nvidia-smi -i 1,2,3 -pl 75

xleejohnx

hero member

Activity: 672

Merit: 500

Quote from: Maxximus007 on June 20, 2017, 05:16:44 AM

Quote from: ?? on ??

I like rufus; I have used it for years, it can also be used to image the USB.

I think the hddRawCopy tool, is easier to use for members who have less experience imaging; there are also a lot of members who are already familiar with it from smOS.

I use an img format for the image; specifically to enable the use of most imaging programs.

This one is also easy to use and is available for mac / linux and windows:

https://etcher.io/

Thanks, being on a Mac, etcher looks great. Saves me starting up Virtualbox Wink

Just leaves me one thing: the Windows partition is unreachable with a Mac. Seemingly this has to do with some flags missing for the FAT partition. Anyone has a solution for this?

What are you trying to do?
Explain into details what you trying to do and I might be able to help out
I deal with windows and Mac daily

Maxximus007

full member

Activity: 153

Merit: 100

Quote from: cryptosharks on June 20, 2017, 07:16:27 AM

Can someone share the OC settings for GTX 1080 OC 8GB (non TI), please?

These are my OC settings for nvOC:

POWERLIMIT="NO"
__CORE_OVERCLOCK=275
MEMORY_OVERCLOCK=600
FAN_SPEED=100

The card gets heated: 82 C (stays there 24/7 consistently)

With the above OC settings, I get 315-330 MH/s on LBC and 495-505 Sol/s on ZEC.

With the same OC settings as on nvOC, Windows 10 is able to get 345-365 MH/s on LBC and 545-565 Sol/s on ZEC with temperature not going over 79 C. I used MSI Afterburner on Windows.

Any ideas?

Well, in nvOC (Linux) the OC is different. So you might want to up your mem OC quite a bit.

As written before by fullzero:
Powerlimits in windows are in percent TDP; in linux they are in watts. Thus a powerlimit of 70 in windows is .7 * 120 watts or whatever the TDP is (84 watts if the card TDP is 120); while 70 in linux is 70 watts. I would recommend comparing Windows and linux powerlimits with a kill-a-watt to get them exactly equal; but most likely a 3gb 1060 has a 120 or 125 watt TDP.

Also linux OC offsets are scaled differently than windows; you will need to use higher offsets to get the same results in linux.

dt8666

full member

Activity: 302

Merit: 102

Mining with noise and without noise is all possibl

Hi. I am first time using this. Follow instruction and use the Atom to edit the bash file.

When boot up, after all start i saw the screen is up.

But it keep saying:
Xxxxxxxxxx.....1263-A96E/onebash
No such file or directory.

What should do?

Appreciate help.

cryptosharks

newbie

Activity: 8

Merit: 0

Can someone share the OC settings for GTX 1080 OC 8GB (non TI), please?

These are my OC settings for nvOC:

POWERLIMIT="NO"
__CORE_OVERCLOCK=275
MEMORY_OVERCLOCK=600
FAN_SPEED=100

The card gets heated: 82 C (stays there 24/7 consistently)

With the above OC settings, I get 315-330 MH/s on LBC and 495-505 Sol/s on ZEC.

With the same OC settings as on nvOC, Windows 10 is able to get 345-365 MH/s on LBC and 545-565 Sol/s on ZEC with temperature not going over 79 C. I used MSI Afterburner on Windows.

Any ideas?

Nexillus

full member

Activity: 169

Merit: 100

Quote from: Brucelats on June 20, 2017, 04:23:10 AM

Quote from: Nexillus on June 19, 2017, 06:45:47 PM

Quote from: ?? on ??

Quote from: WarwickNZ on June 19, 2017, 06:22:26 PM

Quote from: Nexillus on June 19, 2017, 06:08:58 PM

Anyone have any suggestions to this problem, I am all ears.

I have one of my rigs which uses a Asus Prim z270-a when I put the 6th GPU in it will not POST to the bios screen therefore not start up.

From troubleshooting it only happens if I have 2 GPUs plugged into the X16 slots, if I only have one it works fine.

I have changed the BIOS to the recommend from the thread here for the z270-a and set it all to GEN2.

Any other suggestions?

There were 4 Gen2 settings I needed to change, did you change them all? Could be an issue running 2 on the board with 7 or more, i've only tested with 1 on the board +6 risers or all risers. Check for faulty riser cables too, that slowed up a friend of mine getting 7 going on this board.

This was my thought as well; WarwickNZ.

I'm not sure if you can use 2x GPUs direct on the mobo + 5x with risers with this mobo.

When using a GPU directly more pcie lanes / pcie bandwidth is used. This may result in insufficient bandwidth on one of the cards to enable normal operation.

While using a card directly I would change the pcie bandwidth to gen2 for all cards. Then when you get your risers, I would change it back to auto.

As WarwickNZ indicated; there are multiple pcie bandwidth related settings in the bios. There may also be an SLI setting that needs to be disabled as well.

I should probably rephrase. I have all 6 GPUs on risers, the only time it does not post is when I have 2 of the x16 slots with the risers in them. I will double check and make sure they are all the same gen type. I had it on auto first but when I did, it wouldn't even boot past one card. I'm to the point I may just reset the MB bios via cmos and start fresh.

Have u managed to get it work?

Going to be trying it again today. I will get back with you.

Maxximus007

full member

Activity: 153

Merit: 100

Quote from: ?? on ??

I like rufus; I have used it for years, it can also be used to image the USB.

I think the hddRawCopy tool, is easier to use for members who have less experience imaging; there are also a lot of members who are already familiar with it from smOS.

I use an img format for the image; specifically to enable the use of most imaging programs.

This one is also easy to use and is available for mac / linux and windows:

https://etcher.io/

Thanks, being on a Mac, etcher looks great. Saves me starting up Virtualbox Wink

Just leaves me one thing: the Windows partition is unreachable with a Mac. Seemingly this has to do with some flags missing for the FAT partition. Anyone has a solution for this?

Brucelats

sr. member

Activity: 326

Merit: 250

Quote from: Nexillus on June 19, 2017, 06:45:47 PM

Quote from: ?? on ??

Quote from: WarwickNZ on June 19, 2017, 06:22:26 PM

Quote from: Nexillus on June 19, 2017, 06:08:58 PM

Anyone have any suggestions to this problem, I am all ears.

I have one of my rigs which uses a Asus Prim z270-a when I put the 6th GPU in it will not POST to the bios screen therefore not start up.

From troubleshooting it only happens if I have 2 GPUs plugged into the X16 slots, if I only have one it works fine.

I have changed the BIOS to the recommend from the thread here for the z270-a and set it all to GEN2.

Any other suggestions?

There were 4 Gen2 settings I needed to change, did you change them all? Could be an issue running 2 on the board with 7 or more, i've only tested with 1 on the board +6 risers or all risers. Check for faulty riser cables too, that slowed up a friend of mine getting 7 going on this board.

This was my thought as well; WarwickNZ.

I'm not sure if you can use 2x GPUs direct on the mobo + 5x with risers with this mobo.

When using a GPU directly more pcie lanes / pcie bandwidth is used. This may result in insufficient bandwidth on one of the cards to enable normal operation.

While using a card directly I would change the pcie bandwidth to gen2 for all cards. Then when you get your risers, I would change it back to auto.

As WarwickNZ indicated; there are multiple pcie bandwidth related settings in the bios. There may also be an SLI setting that needs to be disabled as well.

I should probably rephrase. I have all 6 GPUs on risers, the only time it does not post is when I have 2 of the x16 slots with the risers in them. I will double check and make sure they are all the same gen type. I had it on auto first but when I did, it wouldn't even boot past one card. I'm to the point I may just reset the MB bios via cmos and start fresh.

Have u managed to get it work?

Brucelats

sr. member

Activity: 326

Merit: 250

Guys! I use Asus Z270-P motherboard. And i am having problems when i add 5th gpu. I am using ubuntu desktop 16.04. I use 5x gtx 1070.

Please, can anyone give me some info how to install my fifth card and then eventualy add 6th or 7th. How to setup Bios, what version etc. I guess nvOC wont work properly also before i deal.with this.

Please.help me Sad

spent last day trying to figure it all out.

When i add 5th gpu its either stuck. It doesnt let me enter bios even. I just see black screen. Or if i change bios options, i get into ubuntu but system sees only 4 gpus. 5th is invisible. Gpu 0 1 2 3 are shown 4 is not there.

Maxximus007

full member

Activity: 153

Merit: 100

Quote from: weareken on June 19, 2017, 09:09:12 PM

A tip and a question:

Tip: If you're tuning your GPUs to reduce power, nvidia-smi can be quite helpful. I open it in a second ssh session so I can watch the miner output in the main window and the GPU in the second.

You can run it with this (refreshes every 0.1 seconds):

Code:

watch -n0.1 nvidia-smi

Question:

My rig is using about 50W more than it should as compared to same Eth mh/s under windows. I've got cclock set to -35 and mclock set to 820. pl is set to 75. I get ~23mh/s, which is what I'm shooting for. from the nvidia-smi output it looks like the problem is that three of the cards (this rig is 3x PNY and 1xEVGA) are not honoring the powerlimit. Still new to tuning nvidia cards until linux. Any ideas?

Code:

Update:
Apparently quite a few people are experiencing this issue with GTX 1060's https://devtalk.nvidia.com/default/topic/1011804/linux/nvidia-smi-not-fully-supported-on-gtx-1060/. In this case, mine is a PNY 3GB version which someone else is having issues with as well. Unclear as to whether this is a driver bug or somehow related to particular motherboards (a Gigabyte z97 SOC Force in my case).

Thanks for the tip!
Looking at the table posted: The cap is in place (75W), but the card is not honouring it. There might be a firmware upgrade for the card?
Out of curiosity: What's the avg Sol/s you get with these cards, or do you just mine Ethash with it? Do you mine single or dual mode?

lbrasi

newbie

Activity: 26

Merit: 0

Hi All,

I have decided to give nvOC a go, I am using an older Gigabyte GA-EX58-UD4P (Socket LGA1366) board and an i7 920 CPU. I have two EVGA GTX 1080ti on risers in both PCI-E x16 slots.
I am able to boot fine but things just are not stable, I can run for about 15 hours but later GPU0 drops to about 150-170w and SOLs are suffering, I can reboot and get it working but it seems to happen again.

I have no idea if any BIOS changes need to be made to get this running stable or not, does anyone have any idea what I can try?

Also if I kill the mining process and make changes to the onebash config it doesn't seem to accept the changes, more specially the fan speed. I attempt to launch the miner again and get the below error but the process does start;

Failed to connect to Mir: Failed to connect to server socket: No such file or directory
Unable to init server: Could not connect: Connection refused

ERROR: The control display is undefined; please run `nvidia-settings --help`
for usage information.

weareken

newbie

Activity: 27

Merit: 0

A tip and a question:

Tip: If you're tuning your GPUs to reduce power, nvidia-smi can be quite helpful. I open it in a second ssh session so I can watch the miner output in the main window and the GPU in the second.

You can run it with this (refreshes every 0.1 seconds):

Code:

watch -n0.1 nvidia-smi

Question:

My rig is using about 50W more than it should as compared to same Eth mh/s under windows. I've got cclock set to -35 and mclock set to 820. pl is set to 75. I get ~23mh/s, which is what I'm shooting for. from the nvidia-smi output it looks like the problem is that three of the cards (this rig is 3x PNY and 1xEVGA) are not honoring the powerlimit. Still new to tuning nvidia cards until linux. Any ideas?

Code:

Update:
Apparently quite a few people are experiencing this issue with GTX 1060's https://devtalk.nvidia.com/default/topic/1011804/linux/nvidia-smi-not-fully-supported-on-gtx-1060/. In this case, mine is a PNY 3GB version which someone else is having issues with as well. Unclear as to whether this is a driver bug or somehow related to particular motherboards (a Gigabyte z97 SOC Force in my case).

Topic: [ mining os ] nvoc - page 324. (Read 418549 times)