• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

gtx660TI clock stuck at idle freq (324MHz)

dbartist

New Member
Joined
Apr 2, 2016
Messages
11 (0.00/day)
hi i have an nvidia gtx660ti and it will not perform as expected
this all started when my monitor broke, and i replaced the gpu with my old one (9500gt) because i needed the analog output. Now im using a digital/analog adapter for the monitor.
i reinstalled windows because it was simpler than cleaning up the driver mess, bf3 now plays at 15 fps instead of a steady 60.
Things iv tried: power limitation (windows and nvidia controll panel) is ok
totally cleared cmos, (in accordance with mobo manual)
clean OS installation (and obviously drivers)
overclocking the gpu, first using mobo auto-clock, then manually
the clock will just stay there even when the gpu load is 100% (running games/benchmarks)
gppu.jpg

i think the gpu "thinks" it's consuming 2 much power and wont clock the core or memory
but i have no idea how to fix it
the cooler performs fine, i heated the gpu at around 50C and it was back at 35 in like half a minute(i didnt time it).The heating was done by blowing warm air at the case intakes and slowly raising the temp for 3-5 mins, i am extremely carefull with those stuff so im sure i didnt damage the card in anyway.
The card is just 2 months out of warranty and was not very stressed
Any ideas ??

mobo- ASUS p8z77-v LX
i7 3770k
2x4GB ram
 
forgot to mention i have no access to digital monitor, another pc with similar gpu capabilities, and my psu +12V is fine (expexted voltage drop at full load)
 
Your core clock is way too high, the card is unstable, and the drivers are falling back to the fail-safe mode.

First, drop everything back to stock and test the card. Then overclock using something like MSI Afterburner. Don't do anything with the motherboard. And your GPU base clock should be around 1050-1150MHz with the stock voltage, not anywhere near 1400Mhz+.
 
Last edited:
pls read the whole post, i already did all that, the problem is not the core clock, and the test were done in default settings, this picture u saw was after trying to solve the problem by OC
 
Try using your system on a tv, via hdmi. Some adapters are wonky.
 
Use ASUS GPU Tweak software not Afterburner since your card is an ASUS brand. Try to make all settings at "default." You need to get the Core clock down to or below 1000MHz, not overclock it. The base clock for a reference GTX 660Ti is actually 915 MHz, I would start there.

http://www.asus.com/supportonly/ASUS GPU Tweak for Graphics cards/HelpDesk_Download

Your boost clock shows 980 MHz so you are correct there. It is also possible that the card is somehow messed up as well. Can you try it in a different computer to test? You say your PSU is fine, but what exact PSU is it? Power problems can be the root of a lot of frustrations. It could be your PSU. But my gut thinks the card is bad.

If you are brave, you could attempt to re-flash the cards BIOS. Check ASUS forums and Google and you'll get some info on that. But yes, that could brick your card. But it's doable -check out this video:

https://www.bing.com/videos/search?...0E014B7F143A5BEE20FF0E014B7F143A5&FORM=VRDGAR
 
Use ASUS GPU Tweak software not Afterburner since your card is an ASUS brand. Try to make all settings at "default." You need to get the Core clock down to or below 1000MHz, not overclock it. The base clock for a reference GTX 660Ti is actually 915 MHz, I would start there.

http://www.asus.com/supportonly/ASUS GPU Tweak for Graphics cards/HelpDesk_Download

Your boost clock shows 980 MHz so you are correct there. It is also possible that the card is somehow messed up as well. Can you try it in a different computer to test? You say your PSU is fine, but what exact PSU is it? Power problems can be the root of a lot of frustrations. It could be your PSU. But my gut thinks the card is bad.

If you are brave, you could attempt to re-flash the cards BIOS. Check ASUS forums and Google and you'll get some info on that. But yes, that could brick your card. But it's doable -check out this video:

https://www.bing.com/videos/search?...0E014B7F143A5BEE20FF0E014B7F143A5&FORM=VRDGAR
This setup has been working flawlessly for 3 years, i tried OC with ASUS tweak, After burner, nvidia inspector, nothing works the psu is sufficient (750w i dont remember brand).
reflashing soulds like a good idea but im not touching it without a ups
 
Maybe you did forget to put back the power connector to the video card?
 
Not sure what it means myself, could be a hardware problem, but your Power Consumption is at 327%.... that would make the card throttle constantly. Not sure why it's doing that, maybe someone with a kepler knows?
 
Azumay may be onto something as well. You are just inviting problems by using that adapter. Try your system on a monitor/TV using the HDMI connector only -no adapter.
And I still am uneasy about your PSU. Especially if you are unsure of the brand. Some people give little thought to their PSU, but I am of the school of thought that it is an extremely crucial component and brand matters a lot! I stick with SeaSonic or Corsair PSUs. Just because its 750w or worked before doesn't mean anything. It could be bad now. There are good quality brands of PSUs and crappy brands. If yours is not a good brand, then it could be wonky, causing ripples in the delivery of clean power and problems throughout your system.


1) Try card in different system - this eliminates card if it works in a different computer
2) Try a different (working) card in your system - this eliminates system as problem
3) Try connection to monitor/tv without using funky adapter
4) Verify that you have a good quality PSU- not just enough wattage
5) Double check using GPU Tweak that your Core clock is set below 1000MHz
6) Consider flashing BIOS of card if all else fails (and you don't need a UPS to flash your cards BIOS?)

Good luck.
 
Last edited:
Azumay may be onto something as well. You are just inviting problems by using that adapter. Try your system on a monitor/TV using the HDMI connector only -no adapter.
And I still am uneasy about your PSU. Especially if you are unsure of the brand. Some people give little thought to their PSU, but I am of the school of thought that it is an extremely crucial component and brand matters a lot! I stick with SeaSonic or Corsair PSUs. Just because its 750w or worked before doesn't mean anything. It could be bad now. There are good quality brands of PSUs and crappy brands. If yours is not a good brand, then it could be wonky, causing ripples in the delivery of clean power and problems throughout your system.


1) Try card in different system - this eliminates card if it works in a different computer
2) Try a different (working) card in your system - this eliminates system as problem
3) Try connection to monitor/tv without using funky adapter
4) Verify that you have a good quality PSU- not just enough wattage
5) Double check using GPU Tweak that your Core clock is set below 1000MHz
5) Consider flashing BIOS of card if all else fails (and you don't need a UPS to flash your cards BIOS?)

Good luck.

1,3)not possible in the near future
2)already done system is fine
4)psu is fine i am 100% sure because i checked it recently, also its a thermaltake smart m750w

i want to flash the bios but power-failures happen randomly and for like no reason at all, and thats why i need a ups...
btw that's how my monitor failed, from a power surge
dunno, maybe ill do it at 4 or 5 in the morning and take my chances if i dont find a ups to use

thanks for the help
 
Maybe you did forget to put back the power connector to the video card?
look at how long my post is, i already have double-checked the cables, also gpu has 2leds that turn from red to green when u connect power to it
 
look at how long my post is, i already have double-checked the cables, also gpu has 2leds that turn from red to green when u connect power to it
It doesn't matter how long winded your post is, even simple things get overlooked! If you are going to give attitude to people attempting to help, you will end up with no help.
 
@dbartist First of all, turn off any auto-overclocking features you might have.

Secondly, your perfcap reason is Power. If you've messed with any overclocking software, what is the power target that you've set? Have you set it very low by accident?


What the hell is with the BIOS flashing recommendation? When will people learn that the last thing you ever want to do is to flash a GPU's bios due to the high probability of something going wrong? Of course it is possible. That doesn't mean it should be done without exhausting every other possibility.

Back to OP, it should not be an issue with a 3D load not being recognized by the card (I get that all the time with this GTX 750 Ti and have not had it with R7 265 or 280X). If that was the case, it should be marked Utilization in Perfcap.


EDIT: I read your original post. How do you check power limit in Windows? I've never heard of such a thing...have you used Afterburner or PrecisionX in the past? How about you give those softwares a try and see what the reading for power limit is?

If you believe the adapter might be the culprit, get a hold of a new one. DVI-I adapters are not expensive. I have used 2 over the years without any such issues.

EDIT 2: Hold on a sec. Nowhere in your original post did you mention that you double checked the PCIe power cables. And you tell someone who gives you a polite, friendly reminder to do so to "look at how long [your] post is"? Does that make any sense to you?
 
Last edited:
I would like to help but this thread has me very confused. It reminds me of another new person here.
 
It doesn't matter how long winded your post is, even simple things get overlooked! If you are going to give attitude to people attempting to help, you will end up with no help.
EDIT 2: Hold on a sec. Nowhere in your original post did you mention that you double checked the PCIe power cables. And you tell someone who gives you a polite, friendly reminder to do so to "look at how long [your] post is"? Does that make any sense to you?
If a person answers without reading even the first post i dont consider this a polite answer because he has made no effort understanding the problem in the first place, i was very specific in my question and we had a small talk about power just above his post

EDIT: I read your original post. How do you check power limit in Windows? I've never heard of such a thing...have you used Afterburner or PrecisionX in the past? How about you give those softwares a try and see what the reading for power limit is?
by power limit in windows i meant (controll panel)->(power settings) and so on
messing with the power limitations or clocks by means of OC sofware has no effect in benchmark performance (always same fps) and the clock is still stuck at 324MHz

i think the gpu "thinks" it's consuming 2 much power and wont clock the core or memory
but i have no idea how to fix it
My TDP is 324% and this measurement is the same in all software, perfcap reason is power but i know i am feeding enough wattage to the card
tdp 3.24 means that with that kind of wattage the tempreture of the chip would asymptotically approach a tempreture 3.24 times higher than maximum allowed, so the card clocks down the clocks. i know for a fact that temprature is not rising, i dont know by what means the power is measured but my guess would be that something went wrong there

What the hell is with the BIOS flashing recommendation? When will people learn that the last thing you ever want to do is to flash a GPU's bios due to the high probability of something going wrong? Of course it is possible. That doesn't mean it should be done without exhausting every other possibility.
have not yet done that, for stated reasons
 
I have a simple question, are running that card at stock speeds? I know that a GTX 660 Ti can't do much more than 1200 on the core before the driver fails and resets to the 2D/default settings. Also the driver that you are using is bad, most people went back to 362.xx. I'm still running 358.xx and 356.xx on my rigs.

I can only OC my CPU and memory in my MB bios but I only have four Z77's and a P67 for reference. Don't look at my System Specs because I gave that rig away years ago.
 
stock speeds, stock clocks, stock volts/power limits.

get them back to where they should be, and see how it behaves. Looks like an overclock done badly, and the settings have stuck to me.
Just because it was stable in the past, doesnt mean its stable now.
 
I have a simple question, are running that card at stock speeds? I know that a GTX 660 Ti can't do much more than 1200 on the core before the driver fails and resets to the 2D/default settings. Also the driver that you are using is bad, most people went back to 362.xx. I'm still running 358.xx and 356.xx on my rigs.

I can only OC my CPU and memory in my MB bios but I only have four Z77's and a P67 for reference. Don't look at my System Specs because I gave that rig away years ago.

stock speeds, stock clocks, stock volts/power limits.

get them back to where they should be, and see how it behaves. Looks like an overclock done badly, and the settings have stuck to me.
Just because it was stable in the past, doesnt mean its stable now.

clocks were always at stock speed i never touched them, because the gpu was adequate to play multiplayer bf3 @ 60fps
i only messed with the card after it started underperforming in an effort to fix it
 
If you haven't already, try the card in another computer (hopefully that's something you can do). See how it reacts in another setup - if the issue follows the card, I'd say something broke on the card.
 
If you haven't already, try the card in another computer (hopefully that's something you can do). See how it reacts in another setup - if the issue follows the card, I'd say something broke on the card.
unfortunatelly i dont have access to another pc with compatible mobo... bummer
 
Once again, I'll point out the 327% TDP afterburner is reporting. Why? If the cards thinks its at 327% power usage, it will throttle to the lowest speed right? Perfcap reason is PWR.
Could be failed vrms.... Never seen more than 110% on my maxwell with limit set to 120.....
 
Hello everyone.
OP probably has already abandoned this thread, but in case anyone else stumbles upon this, I also experience exact same issues on one of my GPUs on GTX 690. Core clock won't go higher than 324 MHz, TDP is off the roof above 300%, afterburner won't allow modifying voltage or memory clock, insane throttling when SLI enabled (for obvious reasons). I've spent my whole day on this issue and it appears that it's just a defective GPU and there's nothing really we can do about it. Most people on the forums ended up with RMAing their cards. Sadly that's the reality and there's no way to fix this AFAIK. Thank god my second GPU on the card is working fine, so I'm essentially left with a GTX 680.
 
u running more than one monitor? if yes that could be the problem I have seen some cards don't downclock when there is more than 1 monitor connected.
 
Back
Top