GTX 1060 Power Limit

Joined
Nov 20, 2021
Messages
4
Hey guys!
I am overclocking a GTX 1060 and I noticed that power limit restrictions will trigger even though the card is performing below 70% TDP.
This is forcing gpu core to lower its frequency even though power should not be an issue.
Any idea how to remove this limitation?
On MSI Afterburner I have power limit set to max (108%) but it does nothing apparently.
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
hwinfo or gpu-z -- read all sensors "limits problems"

gpu-z example
https://forum-en.msi.com/index.php?attachments/bez-n%C3%A1zvu-png.147904/
but best is MSI KOMBUSTOR // AFTERBURNER
what remember my 1060 ,.,.2100mhz core max "100-150 offset only !" TDP - 108./. is OK
but send VBIOS name " I will verify the possibility of TDP "

or GPU-z - ADVANCED - here is TDP info :unsure:

best OC - voltage curve "under 1.05V or similar / 2100mhz max"
Vrel - limit "max voltage" - custom curve fix this
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
or heat - 90C max / 94C slowdown / 98 shutdown
but maybe it's information from mobile gpu "pascal"
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
best is any OSD in game ,,,for determine HW status / debug / setup for smooth FPS
RIVA STATISTIC SERVER + HWINFO or AFTERBURNER

OSD - overlay in 3d "real-time monitoring"
EXAMPLE for OSD
View attachment 153017

this try set ,.,.OSD in any 3D // custom OC in afterburner
verify your setup
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
fast way for undervolt or OC + undervolt

.........AFTERBURNER
example ,,you need to find ,,custom voltage limit + max usable OC

first is undervolt "voltage lock 0.78-0.8V"..,second is OC offset + voltage lock
https://forum-en.msi.com/index.php?attac...zvu-jpg.138583/

verification in 3D render
https://forum-en.msi.com/index.php?attac...vu2-jpg.138585/

and controls is CTRL + F - voltage editor
zone select - SHIFT + mouse

and then pull the selected zone down "under the original curve" - voltage lock on new point

EXAMPLE - your voltage - 1V or more "2100mhz from offset is limit for gtx1060"
and memory +500 - OK "9000 effective// 2250mhz you see .,.,gpuZ,., HWINFO"
 
Joined
Nov 20, 2021
Messages
4
hwinfo or gpu-z -- read all sensors "limits problems"

gpu-z example
but best is MSI KOMBUSTOR // AFTERBURNER
what remember my 1060 ,.,.2100mhz core max "100-150 offset only !" TDP - 108./. is OK
but send VBIOS name " I will verify the possibility of TDP "

or GPU-z - ADVANCED - here is TDP info :unsure:

best OC - voltage curve "under 1.05V or similar / 2100mhz max"
Vrel - limit "max voltage" - custom curve fix this
Here is some GPU-Z info. You can check the log and see that in this case Power limit activated during 23.4% TDP



or heat - 90C max / 94C slowdown / 98 shutdown
but maybe it's information from mobile gpu "pascal"
As you can see in the logs, temperature is not a problem, always far below 50ºC in this case
 

Attachments

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
will see
,,.but best is MSI KOMBUSTOR + HWINFO LOG
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
you using PSU cable for GPU? or only SLOT power ??
1x6 pin or 1x8pin - gtx1060

edit
this is OK ,.1x 8pin
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
teamviewer ? i try create new profile on AFTERBURNER - 1.05V "your is all time on MAX"
and try MSI KOMUSTOR + CPU burner - your PSU looks weak
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
similar problem is CPU "HWINFO"
when cpu get PROCHOT or WHEA - GPU drop
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
some like this
msi kombustor + hwinfo
MS-01
,.,.after time add CPU burner "half .,,.full load "
second GL -3.2GB here VRAM TEST + after time cpu

this zones you must read - CPU "RED" green is "GPU"
Bez názvu.png
in your situation + all POWER from cables - 12V / 3.3 /5V - in graph
verify drop under heavy usage

This stress is ideal for game stability

//edit
or LOG from HWINFO + this usage
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
or teamviewer ? faster than translation - send PM

I can't do the translation in the morning :coffee:
 
Joined
Nov 20, 2021
Messages
4
teamviewer ? i try create new profile on AFTERBURNER - 1.05V "your is all time on MAX"
and try MSI KOMUSTOR + CPU burner - your PSU looks weak
All time on MAX is intentional. I am locking the highest clock on MSI Afterburner for best performance and proper overclock testing.


similar problem is CPU "HWINFO"
when cpu get PROCHOT or WHEA - GPU drop
My CPU is not overclocked so I don't think that would be an issue.


I will report again after testing with Kombustor and HWINFO
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
All time on MAX is intentional. I am locking the highest clock on MSI Afterburner for best performance and proper overclock testing.




My CPU is not overclocked so I don't think that would be an issue.


I will report again after testing with Kombustor and HWINFO
you must test combination test - this stress PSU
and from this is possitible your problem - 12V pcie drop under stable zone / or cpu drop ,..,etc
 

next3r

Well-known member
MAJOR
Joined
Jun 22, 2018
Messages
4,485
All time on MAX is intentional. I am locking the highest clock on MSI Afterburner for best performance and proper overclock testing.
if you have weak PSU ,.,problem above

possitible repair is lower voltage + OC offset - identical performance "but more safe voltage range"for gpu or PSU
1.05 or 1V + 100 or 150 offset - identical performance / better results when have power limit problem

edit
try GL test for 3.2-5.2GB for your 6GB VRAM "full vram stress"
CORE is possitible OK .,,vram dead - possible problem

restore all setup in afterburner "or only try set written above - new "lower" voltage curve
 
Joined
Nov 20, 2021
Messages
4
I have done more tests and realized that, even without any overclock, HWiNFO will say GPU is being limited by both Power and "Reliability Voltage".

Edit:
1-More tests shows that using both CPU+GPU at default values at full load will not trigger Power limit if using 108% power limit in MSI Afterburner
2-GPU power limit without any overclock and 100% in MSI AB will trigger if close to 100% as it should.
This means the problem is in the overclock. I will run some more tests.

Edit2: The "Reliability Voltage" aparently just means the GPU won't clock itself higher because it reached the highest known stable state.
 
Last edited:
Top