[ mining os ] nvoc - page 105. | Bitcointalksearch.org

papampi

full member

Activity: 686

Merit: 140

Linux FOREVER! Resistance is futile!!!

Quote from: WaveFront on December 18, 2017, 04:46:02 AM

Hello,
in the PAPAMPI WTM_AUTO_SWITCH settings, what is the exact role of the setting:

WTM_AUTO_SWITCH_SYNC_INTERVAL="3"

I suppose the units here are minutes.

Cheers

That's the time in minutes between syncing and getting coins profitability data from whattomine.com

WaveFront

member

Activity: 126

Merit: 10

Hello,
in the PAPAMPI WTM_AUTO_SWITCH settings, what is the exact role of the setting:

WTM_AUTO_SWITCH_SYNC_INTERVAL="3"

I suppose the units here are minutes.

Cheers

thaelin

newbie

Activity: 64

Merit: 0

I did find out what my restart problem was. I have a GPU that is falling out now and then. It seems to have a different style of ram in it than the others and is a bit slower. So the WatchDog says you had a GPU fall out and I am re-starting the 3main which brings it back for whatever time until it does it again. Limits my down time on the tester and so is working great.

thanks again thay

Hope my comments are helpful and not a hinderence. I do seem to come across the wrong way some times.

thaelin

newbie

Activity: 64

Merit: 0

@cryptatome

OK papmp, got the time stamp going but had to put the change into the WTM check and not auto switch as it is off. I am only wanting at this time to see what coins are bobbing to the top and for how long and what kind of frequency that they will have. Years ago I did an AI program for the stock market and it did quite well for a long time until they found out I had it running and changed the game. Then "08 came and wiped out all I did.
I see patterns and that is what I am looking for here as well. Once I have a sufficient data set to work with, I will be able to predict which will be the best to run... I hope..

Crypto as soon as I get the thing soused out I will be glad to pass on to you. It will be a simple form of which you will extract the data comma delimited and import to excel like pgm.

You will have to extend the size of the log file for What to min check so you can capture a large amount of data

papampi

full member

Activity: 686

Merit: 140

Linux FOREVER! Resistance is futile!!!

Quote from: martyroz on December 17, 2017, 09:10:08 PM

Quote from: crazydane on December 17, 2017, 08:15:50 PM

Rolled 19-2 out to all my rigs and no issues whatsoever.

Is there a standalone download for 19_2.0? or are you typing some command to update?

Right now I have two rigs on 19_1.4 mining Equihash, so not in a great hurry to upgrade.

For now update script
Run:

Code:

bash nvOC-19-2-update

Image install will be ready soon

papampi

full member

Activity: 686

Merit: 140

Linux FOREVER! Resistance is futile!!!

Quote from: monck on December 17, 2017, 03:54:53 PM

Quote from: papampi on December 17, 2017, 05:40:03 AM

Quote from: papampi on December 17, 2017, 02:24:38 AM

Quote from: monck on December 16, 2017, 09:47:02 PM

Yes it should be terminated and start again
WTM Auto Switch Logic is :
if ALGO_SPECIFIC_OC="NO", it only restart miner and watchdog (to prevent wdog from catching low util and restart again)
if ALGO_SPECIFIC_OC="YES", it checks coins algo
if algo changed, it restart watchdog, temp control and 3main so specific algo takes effect
if algo not changed, it only restart watchdog and miner

You will see a miner terminated in gnome-terminal when algo not changed, if you check wtm tab you see it started with new command line and you can watch miner output log with:

Code:

~/nvOC miner-log

I have some ideas for next version to show the miner info in a new tab when miner terminated.

Thanks for explanation - the problem is that nothing is restarting after termination neither wtm nor watchdog and miner.
This behavior is on two independant rigs.
At the time I am running without Auto Switch
As I'am a Linux virgin - is there a possibility to give you more information for check?

I made a small change so the terminal keeps showing miner output even after switch.

papampi,
many thanks for your support - functionality is great and miner stays in terminal after coin switch.
But is this also possible for the Watchdog terminal - otherwise don't see whether the Watchdog is running or not?
Watchdog is terminated and not rectivating after coin change.

It is running in the background
In a guake tab run

Code:

~/nvOC wdog-log

Tigel

newbie

Activity: 66

Merit: 0

Quote from: joshuajones02 on December 18, 2017, 12:44:50 AM

what can I type in the terminal to modify GPU settings while miner is running for each GPU individually? i.e Power limit, oc core, & fan speed

For GPU1:

nvidia-smi -i 1 -pl 75

export DISPLAY=:0
nvidia-settings -a [gpu:1]/GPUGraphicsClockOffset[3]=200
nvidia-settings -a [gpu:1]/GPUMemoryTransferRateOffset[3]=1000
nvidia-settings -a [gpu:1]/GPUFanControlState=1
nvidia-settings -a [fan:1]/GPUTargetFanSpeed=8

--
Tigel

joshuajones02

full member

Activity: 210

Merit: 100

what can I type in the terminal to modify GPU settings while miner is running for each GPU individually? i.e Power limit, oc core, & fan speed

Tigel

newbie

Activity: 66

Merit: 0

Would it be complicated to add bminer?

https://bitcointalksearch.org/topic/annbminer-a-fast-equihashethashcuckaroo29z-miner-for-amdnvidia-gpus-1649-2519271
https://www.bminer.me/

--
Tigel

wi$em@n

newbie

Activity: 46

Merit: 0

Quote from: crazydane on December 17, 2017, 10:16:57 PM

The power sensing circuitry in the 1050Ti is faulty so nvidia disabled pulling these false readings in the recent drivers.

https://devtalk.nvidia.com/default/topic/1024744/nvidia-387-12-breaks-power-reading-in-nvidia-smi-/

OK, that's the answer! The bad thing is that it does not only fails to read, but what is more important to set power limit. On older driver it did both correctly, I checked that with wattmeter.

crazydane

full member

Activity: 558

Merit: 194

The power sensing circuitry in the 1050Ti is faulty so nvidia disabled pulling these false readings in the recent drivers.

https://devtalk.nvidia.com/default/topic/1024744/nvidia-387-12-breaks-power-reading-in-nvidia-smi-/

wi$em@n

newbie

Activity: 46

Merit: 0

Quote from: Stubo on December 17, 2017, 02:22:51 PM

Quote from: wi$em@n on December 17, 2017, 01:56:47 PM

Quote from: Stubo on December 17, 2017, 01:38:03 PM

Quote from: wi$em@n on December 17, 2017, 12:47:48 PM

HI,
after applying the latest version of nvOC-19-2-update I've got this in the MINER_TEMP_CONTROL:

GPU 0, Target temp: 65, Current: 60, Diff: 5, Fan: 50, Power: [Not Supported]
GPU 1, Target temp: 65, Current: 59, Diff: 6, Fan: 50, Power: [Not Supported]
GPU 2, Target temp: 65, Current: 57, Diff: 8, Fan: 50, Power: [Not Supported]
GPU 3, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU 4, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU 5, Target temp: 65, Current: 64, Diff: 1, Fan: 50, Power: [Not Supported]
GPU 6, Target temp: 65, Current: 61, Diff: 4, Fan: 50, Power: [Not Supported]
GPU 7, Target temp: 65, Current: 59, Diff: 6, Fan: 50, Power: [Not Supported]
GPU 8, Target temp: 65, Current: 63, Diff: 2, Fan: 55, Power: [Not Supported]
GPU 9, Target temp: 65, Current: 64, Diff: 1, Fan: 50, Power: [Not Supported]
GPU 10, Target temp: 65, Current: 60, Diff: 5, Fan: 50, Power: [Not Supported]
GPU 11, Target temp: 65, Current: 62, Diff: 3, Fan: 50, Power: [Not Supported]
GPU 12, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
Sun Dec 17 10:43:54 MST 2017 - All good, will check again in 13.5 seconds

Why power has become "Not Supported"?

Ah. Are you are using the 1050 GPU? For that GPU, Nvidia's linux API does not support reading the power. Hence, reading it returns [Not Supported]. It is the only GPU (so far) that we have seen like that. As far as I know, it has little to no impact on the temp control for 19-2.0 but leenoox can correct me if I am wrong.

It is 1050ti. On nvOC-19-1.4 and early versions of nvOC-19-2 it was like this:

GPU 0, Target temp: 75, Current: 73, Diff: 2, Fan: 50, Power: 70
GPU 1, Target temp: 75, Current: 73, Diff: 2, Fan: 50, Power: 70
GPU 2, Target temp: 75, Current: 75, Diff: 0, Fan: 50, Power: 70
GPU 3, Target temp: 75, Current: 68, Diff: 7, Fan: 50, Power: 70
GPU 4, Target temp: 75, Current: 68, Diff: 7, Fan: 50, Power: 70
GPU 5, Target temp: 75, Current: 66, Diff: 9, Fan: 50, Power: 70

Sun Dec 17 11:52:06 MST 2017 - All good, will check again soon

Hmm. Try the this on your rig and see what it returns:

Code:

nvidia-smi --query-gpu=power.draw --format=csv

It returns the same:
m1@m1-desktop:~$ GPU 0, Target temp: 65, Current: 60, Diff: 5, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 1, Target temp: 65, Current: 59, Diff: 6, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 2, Target temp: 65, Current: 57, Diff: 8, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 3, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 4, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 5, Target temp: 65, Current: 64, Diff: 1, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 6, Target temp: 65, Current: 61, Diff: 4, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 7, Target temp: 65, Current: 59, Diff: 6, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 8, Target temp: 65, Current: 63, Diff: 2, Fan: 55, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 9, Target temp: 65, Current: 64, Diff: 1, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 10, Target temp: 65, Current: 60, Diff: 5, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 11, Target temp: 65, Current: 62, Diff: 3, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ GPU 12, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU: command not found
m1@m1-desktop:~$
m1@m1-desktop:~$ Sun Dec 17 10:43:54 MST 2017 - All good, will check again in 13.5 seconds

crazydane

full member

Activity: 558

Merit: 194

I used the upgrade script and had modified versions of 1bash ready to push to reach miner after the update script completed.

I believe papampi is working on a 19_2.0 image.

martyroz

full member

Activity: 325

Merit: 110

Quote from: crazydane on December 17, 2017, 08:15:50 PM

Rolled 19-2 out to all my rigs and no issues whatsoever.

Is there a standalone download for 19_2.0? or are you typing some command to update?

Right now I have two rigs on 19_1.4 mining Equihash, so not in a great hurry to upgrade.

crazydane

full member

Activity: 558

Merit: 194

Rolled 19-2 out to all my rigs and no issues whatsoever.

I have rigs with 1050Tis, 1060's, 1070's, 1070Ti's and 1080Tis and each rig has between 3 and 9 GPUs, so a decent mix of GPU models and rig configurations.

http://www.cstone.net/~dk/19-2rollout.JPG

In a future version, I'd love to see a "summary" feature of the rig stats and I soon won't be able to fit them in a single browser window on a 4k monitor.

Stubo

member

Activity: 224

Merit: 13

Quote from: killeriq on December 17, 2017, 03:36:29 PM

Quote from: papampi on December 17, 2017, 12:32:19 PM

Quote from: killeriq on December 17, 2017, 11:28:44 AM

Hello,

Please could someone explain me or point me to right direction?

Im bit confused about OC settings:

there are multiple OC parts:

1. This one i understand is the MAIN setting for all card/algo/hash

I either enable/disable PowerLimit, but the CORE and MEM stays

Code:

# OC (overclock) and Power Limit settings #
# #
########################################################

POWERLIMIT="NO" # YES or NO # Choose YES to set Global Power Limit for all GPU's

POWERLIMIT_WATTS=75 # Global Power Limit in watts for all GPU's. If enabled, INDIVIDUAL settings will override this.

__CORE_OVERCLOCK=100 # Global Core overclock for all GPU's if not using INDIVIDUAL settings

MEMORY_OVERCLOCK=100 # Global Memory overclock for all GPU's if not using INDIVIDUAL settings

2. Set this as NO - dont want to be ALGO (coin) specific

Code:

# Default values set at the end of 1bash are for 1060, remember to set yours accordingly.
ALGO_SPECIFIC_OC="NO" # YES or NO # See the description above

3. This one i thought its hash (equihash) specific OC setting, but it wont take it - ignores
So what is it for? not sure what relation it has with FANs

Code:

# ENSURE YOU USE MINER_TEMP_CONTROL for fanspeed

equihash_POWERLIMIT_WATTS=115
__equihash_CORE_OVERCLOCK=110
equihash_MEMORY_OVERCLOCK=500

4. Those are per card in case i have different models in the rig

Code:

########################################################
# #
# INDIVIDUAL Power Limit, Temp and Clock settings #
# #
########################################################

# Settings below are for rigs using INDIVIDUAL powerlimit / target temps / clocks only
# If set to YES, the INDIVIDUAL settings will override the GLOBAL settings

INDIVIDUAL_POWERLIMIT="NO" # YES or NO

# Set individual powerlimits here if INDIVIDUAL_POWERLIMIT="YES"
INDIVIDUAL_POWERLIMIT_0=100
INDIVIDUAL_POWERLIMIT_1=100

5. Not sure why are also HASH specific OC settings and also ALGO/COIN specific - bit confused Huh

Code:

# ALGO_SPECIFIC_OC settings #
# #
########################################################

# ALGO_SPECIFIC_OC settings
# OC settings you can use based on COIN or ALGO (how ever you want to categorize them)
# Deafault values are for 1060 check http://krypto-mining.blogspot.co.uk/ for other OC values

if [ $ALGO_SPECIFIC_OC == "YES" ]
then

# Equihash
if [ $COIN == "BTG" -o $COIN == "ZEC" -o $COIN == "ZEN" -o $COIN == "HUSH" -o $COIN == "ZCL" -o $COIN == "KMD" -o $COIN == "ZPOOL_EQUIHASH" -o $COIN == "MPH_EQUIHASH" -o $COIN == "NICE_EQUIHASH" -o $COIN == "VOTE" ]
then
ALGORITHM="EQUIHASH"
POWERLIMIT_WATTS=130
__CORE_OVERCLOCK=150
MEMORY_OVERCLOCK=800
fi

Your part 3 is for salfter auto switch... Don't know if there is a better way to show it

Code:

########################################################
# #
# SALFTER_NICEHASH_PROFIT_SWITCHING & #
# SALFTER_MPH_PROFIT_SWITCHING Settings #
# #
########################################################

CURRENCY=USD

POWER_COST=0.20

MINIMUM_PROFIT=0.0

MPH_USERNAME=replace_with_your_MPH_username

PAYMENT_ADDRESS="$BTC_ADDRESS"

WORKER_NAME="$WORKERNAME"

PROFIT_CHECK_TIMEOUT=600

# ENSURE YOU USE MINER_TEMP_CONTROL for fanspeed

daggerhashimoto_POWERLIMIT_WATTS=120
__daggerhashimoto_CORE_OVERCLOCK=100
daggerhashimoto_MEMORY_OVERCLOCK=100

If you want individual OC its the last part of 1bash your part 5

Ah the no.3 its only setting when im using the SALFTER_MPH_PROFIT_SWITCHING Settings ?
If i dont use that switcher (this setting WTM_AUTO_SWITCH="NO" ) it ignores the OC values. Did i get it right?

No.1

Code:

# OC (overclock) and Power Limit settings #
# #
########################################################

POWERLIMIT="NO" # YES or NO # Choose YES to set Global Power Limit for all GPU's

POWERLIMIT_WATTS=75 # Global Power Limit in watts for all GPU's. If enabled, INDIVIDUAL settings will override this.

__CORE_OVERCLOCK=100 # Global Core overclock for all GPU's if not using INDIVIDUAL settings

MEMORY_OVERCLOCK=100 # Global Memory overclock for all GPU's if not using INDIVIDUAL settings

Those are OC setting in TOP of other ?

POWERLIMIT="NO" - will be able to take Watts from values bellow ALGO OC (No.5) and Individual (No.4)

POWERLIMIT="YES" - if i set POWERLIMIT_WATTS=75 other No.4/5 wont go above 75?

How are those 2 managed:
__CORE_OVERCLOCK=100 # Global Core overclock for all GPU's if not using INDIVIDUAL settings
MEMORY_OVERCLOCK=100

- does it overwrite also ALGO OC? (No.5)?

Thanks

Here is the order of precedence for OC:

1) If INDIVIDUAL_CLOCKS=YES, those will override any other OC settings
2) If damNmad_ALGO_SPECIFIC_OC=YES, settings in that section (bottom of 1bash) override __CORE_OVERCLOCK and MEMORY_OVERCLOCK
3) __CORE_OVERCLOCK and MEMORY_OVERCLOCK determine OC settings

Hope this helps.

monck

newbie

Activity: 15

Merit: 0

Quote from: papampi on December 17, 2017, 05:40:03 AM

Quote from: papampi on December 17, 2017, 02:24:38 AM

Quote from: monck on December 16, 2017, 09:47:02 PM

Yes it should be terminated and start again
WTM Auto Switch Logic is :
if ALGO_SPECIFIC_OC="NO", it only restart miner and watchdog (to prevent wdog from catching low util and restart again)
if ALGO_SPECIFIC_OC="YES", it checks coins algo
if algo changed, it restart watchdog, temp control and 3main so specific algo takes effect
if algo not changed, it only restart watchdog and miner

You will see a miner terminated in gnome-terminal when algo not changed, if you check wtm tab you see it started with new command line and you can watch miner output log with:

Code:

~/nvOC miner-log

I have some ideas for next version to show the miner info in a new tab when miner terminated.

Thanks for explanation - the problem is that nothing is restarting after termination neither wtm nor watchdog and miner.
This behavior is on two independant rigs.
At the time I am running without Auto Switch
As I'am a Linux virgin - is there a possibility to give you more information for check?

I made a small change so the terminal keeps showing miner output even after switch.

papampi,
many thanks for your support - functionality is great and miner stays in terminal after coin switch.
But is this also possible for the Watchdog terminal - otherwise don't see whether the Watchdog is running or not?
Watchdog is terminated and not rectivating after coin change.

killeriq

newbie

Activity: 22

Merit: 0

Quote from: papampi on December 17, 2017, 12:32:19 PM

Quote from: killeriq on December 17, 2017, 11:28:44 AM

Hello,

Please could someone explain me or point me to right direction?

Im bit confused about OC settings:

there are multiple OC parts:

1. This one i understand is the MAIN setting for all card/algo/hash

I either enable/disable PowerLimit, but the CORE and MEM stays

Code:

# OC (overclock) and Power Limit settings #
# #
########################################################

POWERLIMIT="NO" # YES or NO # Choose YES to set Global Power Limit for all GPU's

POWERLIMIT_WATTS=75 # Global Power Limit in watts for all GPU's. If enabled, INDIVIDUAL settings will override this.

__CORE_OVERCLOCK=100 # Global Core overclock for all GPU's if not using INDIVIDUAL settings

MEMORY_OVERCLOCK=100 # Global Memory overclock for all GPU's if not using INDIVIDUAL settings

2. Set this as NO - dont want to be ALGO (coin) specific

Code:

# Default values set at the end of 1bash are for 1060, remember to set yours accordingly.
ALGO_SPECIFIC_OC="NO" # YES or NO # See the description above

3. This one i thought its hash (equihash) specific OC setting, but it wont take it - ignores
So what is it for? not sure what relation it has with FANs

Code:

# ENSURE YOU USE MINER_TEMP_CONTROL for fanspeed

equihash_POWERLIMIT_WATTS=115
__equihash_CORE_OVERCLOCK=110
equihash_MEMORY_OVERCLOCK=500

4. Those are per card in case i have different models in the rig

Code:

########################################################
# #
# INDIVIDUAL Power Limit, Temp and Clock settings #
# #
########################################################

# Settings below are for rigs using INDIVIDUAL powerlimit / target temps / clocks only
# If set to YES, the INDIVIDUAL settings will override the GLOBAL settings

INDIVIDUAL_POWERLIMIT="NO" # YES or NO

# Set individual powerlimits here if INDIVIDUAL_POWERLIMIT="YES"
INDIVIDUAL_POWERLIMIT_0=100
INDIVIDUAL_POWERLIMIT_1=100

5. Not sure why are also HASH specific OC settings and also ALGO/COIN specific - bit confused Huh

Code:

# ALGO_SPECIFIC_OC settings #
# #
########################################################

# ALGO_SPECIFIC_OC settings
# OC settings you can use based on COIN or ALGO (how ever you want to categorize them)
# Deafault values are for 1060 check http://krypto-mining.blogspot.co.uk/ for other OC values

if [ $ALGO_SPECIFIC_OC == "YES" ]
then

# Equihash
if [ $COIN == "BTG" -o $COIN == "ZEC" -o $COIN == "ZEN" -o $COIN == "HUSH" -o $COIN == "ZCL" -o $COIN == "KMD" -o $COIN == "ZPOOL_EQUIHASH" -o $COIN == "MPH_EQUIHASH" -o $COIN == "NICE_EQUIHASH" -o $COIN == "VOTE" ]
then
ALGORITHM="EQUIHASH"
POWERLIMIT_WATTS=130
__CORE_OVERCLOCK=150
MEMORY_OVERCLOCK=800
fi

Your part 3 is for salfter auto switch... Don't know if there is a better way to show it

Code:

########################################################
# #
# SALFTER_NICEHASH_PROFIT_SWITCHING & #
# SALFTER_MPH_PROFIT_SWITCHING Settings #
# #
########################################################

CURRENCY=USD

POWER_COST=0.20

MINIMUM_PROFIT=0.0

MPH_USERNAME=replace_with_your_MPH_username

PAYMENT_ADDRESS="$BTC_ADDRESS"

WORKER_NAME="$WORKERNAME"

PROFIT_CHECK_TIMEOUT=600

# ENSURE YOU USE MINER_TEMP_CONTROL for fanspeed

daggerhashimoto_POWERLIMIT_WATTS=120
__daggerhashimoto_CORE_OVERCLOCK=100
daggerhashimoto_MEMORY_OVERCLOCK=100

If you want individual OC its the last part of 1bash your part 5

Ah the no.3 its only setting when im using the SALFTER_MPH_PROFIT_SWITCHING Settings ?
If i dont use that switcher (this setting WTM_AUTO_SWITCH="NO" ) it ignores the OC values. Did i get it right?

No.1

Code:

# OC (overclock) and Power Limit settings #
# #
########################################################

POWERLIMIT="NO" # YES or NO # Choose YES to set Global Power Limit for all GPU's

POWERLIMIT_WATTS=75 # Global Power Limit in watts for all GPU's. If enabled, INDIVIDUAL settings will override this.

__CORE_OVERCLOCK=100 # Global Core overclock for all GPU's if not using INDIVIDUAL settings

MEMORY_OVERCLOCK=100 # Global Memory overclock for all GPU's if not using INDIVIDUAL settings

Those are OC setting in TOP of other ?

POWERLIMIT="NO" - will be able to take Watts from values bellow ALGO OC (No.5) and Individual (No.4)

POWERLIMIT="YES" - if i set POWERLIMIT_WATTS=75 other No.4/5 wont go above 75?

How are those 2 managed:
__CORE_OVERCLOCK=100 # Global Core overclock for all GPU's if not using INDIVIDUAL settings
MEMORY_OVERCLOCK=100

- does it overwrite also ALGO OC? (No.5)?

Thanks

Stubo

member

Activity: 224

Merit: 13

Quote from: wi$em@n on December 17, 2017, 01:56:47 PM

Quote from: Stubo on December 17, 2017, 01:38:03 PM

Quote from: wi$em@n on December 17, 2017, 12:47:48 PM

HI,
after applying the latest version of nvOC-19-2-update I've got this in the MINER_TEMP_CONTROL:

GPU 0, Target temp: 65, Current: 60, Diff: 5, Fan: 50, Power: [Not Supported]
GPU 1, Target temp: 65, Current: 59, Diff: 6, Fan: 50, Power: [Not Supported]
GPU 2, Target temp: 65, Current: 57, Diff: 8, Fan: 50, Power: [Not Supported]
GPU 3, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU 4, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU 5, Target temp: 65, Current: 64, Diff: 1, Fan: 50, Power: [Not Supported]
GPU 6, Target temp: 65, Current: 61, Diff: 4, Fan: 50, Power: [Not Supported]
GPU 7, Target temp: 65, Current: 59, Diff: 6, Fan: 50, Power: [Not Supported]
GPU 8, Target temp: 65, Current: 63, Diff: 2, Fan: 55, Power: [Not Supported]
GPU 9, Target temp: 65, Current: 64, Diff: 1, Fan: 50, Power: [Not Supported]
GPU 10, Target temp: 65, Current: 60, Diff: 5, Fan: 50, Power: [Not Supported]
GPU 11, Target temp: 65, Current: 62, Diff: 3, Fan: 50, Power: [Not Supported]
GPU 12, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
Sun Dec 17 10:43:54 MST 2017 - All good, will check again in 13.5 seconds

Why power has become "Not Supported"?

Ah. Are you are using the 1050 GPU? For that GPU, Nvidia's linux API does not support reading the power. Hence, reading it returns [Not Supported]. It is the only GPU (so far) that we have seen like that. As far as I know, it has little to no impact on the temp control for 19-2.0 but leenoox can correct me if I am wrong.

It is 1050ti. On nvOC-19-1.4 and early versions of nvOC-19-2 it was like this:

GPU 0, Target temp: 75, Current: 73, Diff: 2, Fan: 50, Power: 70
GPU 1, Target temp: 75, Current: 73, Diff: 2, Fan: 50, Power: 70
GPU 2, Target temp: 75, Current: 75, Diff: 0, Fan: 50, Power: 70
GPU 3, Target temp: 75, Current: 68, Diff: 7, Fan: 50, Power: 70
GPU 4, Target temp: 75, Current: 68, Diff: 7, Fan: 50, Power: 70
GPU 5, Target temp: 75, Current: 66, Diff: 9, Fan: 50, Power: 70

Sun Dec 17 11:52:06 MST 2017 - All good, will check again soon

Hmm. Try the this on your rig and see what it returns:

Code:

nvidia-smi --query-gpu=power.draw --format=csv

wi$em@n

newbie

Activity: 46

Merit: 0

Quote from: Stubo on December 17, 2017, 01:38:03 PM

Quote from: wi$em@n on December 17, 2017, 12:47:48 PM

HI,
after applying the latest version of nvOC-19-2-update I've got this in the MINER_TEMP_CONTROL:

GPU 0, Target temp: 65, Current: 60, Diff: 5, Fan: 50, Power: [Not Supported]
GPU 1, Target temp: 65, Current: 59, Diff: 6, Fan: 50, Power: [Not Supported]
GPU 2, Target temp: 65, Current: 57, Diff: 8, Fan: 50, Power: [Not Supported]
GPU 3, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU 4, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
GPU 5, Target temp: 65, Current: 64, Diff: 1, Fan: 50, Power: [Not Supported]
GPU 6, Target temp: 65, Current: 61, Diff: 4, Fan: 50, Power: [Not Supported]
GPU 7, Target temp: 65, Current: 59, Diff: 6, Fan: 50, Power: [Not Supported]
GPU 8, Target temp: 65, Current: 63, Diff: 2, Fan: 55, Power: [Not Supported]
GPU 9, Target temp: 65, Current: 64, Diff: 1, Fan: 50, Power: [Not Supported]
GPU 10, Target temp: 65, Current: 60, Diff: 5, Fan: 50, Power: [Not Supported]
GPU 11, Target temp: 65, Current: 62, Diff: 3, Fan: 50, Power: [Not Supported]
GPU 12, Target temp: 65, Current: 63, Diff: 2, Fan: 50, Power: [Not Supported]
Sun Dec 17 10:43:54 MST 2017 - All good, will check again in 13.5 seconds

Why power has become "Not Supported"?

Ah. Are you are using the 1050 GPU? For that GPU, Nvidia's linux API does not support reading the power. Hence, reading it returns [Not Supported]. It is the only GPU (so far) that we have seen like that. As far as I know, it has little to no impact on the temp control for 19-2.0 but leenoox can correct me if I am wrong.

It is 1050ti. On nvOC-19-1.4 and early versions of nvOC-19-2 it was like this:

GPU 0, Target temp: 75, Current: 73, Diff: 2, Fan: 50, Power: 70
GPU 1, Target temp: 75, Current: 73, Diff: 2, Fan: 50, Power: 70
GPU 2, Target temp: 75, Current: 75, Diff: 0, Fan: 50, Power: 70
GPU 3, Target temp: 75, Current: 68, Diff: 7, Fan: 50, Power: 70
GPU 4, Target temp: 75, Current: 68, Diff: 7, Fan: 50, Power: 70
GPU 5, Target temp: 75, Current: 66, Diff: 9, Fan: 50, Power: 70

Sun Dec 17 11:52:06 MST 2017 - All good, will check again soon

Topic: [ mining os ] nvoc - page 105. (Read 418542 times)