Periodic framerate drops corresponding to massive drops in GPU load
Specs: MSI gt683R with Intel Core i7 2630QM clocked at 2.0 ghz, 12gb of DDr3 memory, 2 500gb hard drives and a GeForce GTX 560M with 1.5gb internal graphics ram. The GPU usually runs around 80-90 Celsius. I first noticed this problem with Mechwarrior Online. I assumed it was a problem with that game, but none of the fixes mentioned made any difference for very long. I recently (last couple weeks) noticed the problem in other games as well. Problem is this: I will start with fully playable FPS (usually in the 40-60 range) then, with some regularity, my framerate will drop down to the 5-15 range, which in FPS games like MWO, makes the game pretty much unplayable. My FPS will stay low for about 15-30 seconds, then pop back up to playable levels again for a short time. This phenomena has started occuring in every game I play with a 3D component, from Arkham City to Minecraft. I used a monitor and noticed that there was a correlation between these FPS drops and my GPU load plummeting from nearly 100%, which I would expect with a full screen game with detailed graphics on a slightly older card, to as little as 0% in the logs. I don't know why the load drops like this, but the load drop coincides with the card temp going over 90 celsius. I can only assume there is a temperature related throttle somewhere. Any ideas on how to fix this? I only have the laptop to game on right now and no chance of buying anything better for at least a year.
Specs: MSI gt683R with Intel Core i7 2630QM clocked at 2.0 ghz, 12gb of DDr3 memory, 2 500gb hard drives and a GeForce GTX 560M with 1.5gb internal graphics ram. The GPU usually runs around 80-90 Celsius.

I first noticed this problem with Mechwarrior Online. I assumed it was a problem with that game, but none of the fixes mentioned made any difference for very long.

I recently (last couple weeks) noticed the problem in other games as well.

Problem is this: I will start with fully playable FPS (usually in the 40-60 range) then, with some regularity, my framerate will drop down to the 5-15 range, which in FPS games like MWO, makes the game pretty much unplayable. My FPS will stay low for about 15-30 seconds, then pop back up to playable levels again for a short time. This phenomena has started occuring in every game I play with a 3D component, from Arkham City to Minecraft.

I used a monitor and noticed that there was a correlation between these FPS drops and my GPU load plummeting from nearly 100%, which I would expect with a full screen game with detailed graphics on a slightly older card, to as little as 0% in the logs. I don't know why the load drops like this, but the load drop coincides with the card temp going over 90 celsius. I can only assume there is a temperature related throttle somewhere.

Any ideas on how to fix this? I only have the laptop to game on right now and no chance of buying anything better for at least a year.

#1
Posted 11/18/2013 02:41 AM   
having the same thing here with my GTX 675m. Temps are fine, still the system shows close to 0% GPU load in irregular intervals, like a throttle. I just posted about it in the Feedback-Thread for the newest driver. Let me paste that for you: ManuelG's response to someone else's post: ManuelG said: huyhung411991 said: Installed and tested. GPU Boost not working with some notebook GPUs (1392369) is not fixed. It is mismatching thermal policies which we are investigating. Not fixed in this driver. My response: I can confirm exactly this - since (if I remember correctly) every driver past 314.22 my laptop, running a GTX 675m, throttles down because of the GPU in irregular intervals. Temperatures of CPU and GPU are monitored and always fine. CPU never goes past ~75° Celsius, GPU never past 80° Celsius, at which point my laptop's fan spins up to it's maximum and brings the GPU down to ~72° again. Working fine, no temperature alerts of any kind. I fiddled around a bit and used the ThrottleStop utility to create log files while my laptop does these throttles. I found that it is the "BD Prochot" function of my Intel chip that triggers the throttle pretty much as soon as NVidia Optimus switches from my Intel HD4000 to my GTX675m. Deactivating BD Prochot via ThrottleStop gets rid of the problem, but of course I then have to be careful and always monitor the temperatures. Also, I could solve the problem by doing a clean install of the 314.22 driver (again, if I remember correctly), the throttling stopped immediately. But since Battlefield 4 refuses to run stable on my laptop with that driver installed, I now run the newest driver with BD Prochot deactivated... please solve this so I can turn this function on again for the sake of my laptop's safety!
having the same thing here with my GTX 675m. Temps are fine, still the system shows close to 0% GPU load in irregular intervals, like a throttle. I just posted about it in the Feedback-Thread for the newest driver. Let me paste that for you:

ManuelG's response to someone else's post:

ManuelG said:
huyhung411991 said:
Installed and tested. GPU Boost not working with some notebook GPUs (1392369) is not fixed.


It is mismatching thermal policies which we are investigating. Not fixed in this driver.

My response:

I can confirm exactly this - since (if I remember correctly) every driver past 314.22 my laptop, running a GTX 675m, throttles down because of the GPU in irregular intervals. Temperatures of CPU and GPU are monitored and always fine. CPU never goes past ~75° Celsius, GPU never past 80° Celsius, at which point my laptop's fan spins up to it's maximum and brings the GPU down to ~72° again. Working fine, no temperature alerts of any kind.

I fiddled around a bit and used the ThrottleStop utility to create log files while my laptop does these throttles. I found that it is the "BD Prochot" function of my Intel chip that triggers the throttle pretty much as soon as NVidia Optimus switches from my Intel HD4000 to my GTX675m. Deactivating BD Prochot via ThrottleStop gets rid of the problem, but of course I then have to be careful and always monitor the temperatures. Also, I could solve the problem by doing a clean install of the 314.22 driver (again, if I remember correctly), the throttling stopped immediately. But since Battlefield 4 refuses to run stable on my laptop with that driver installed, I now run the newest driver with BD Prochot deactivated... please solve this so I can turn this function on again for the sake of my laptop's safety!

#2
Posted 11/19/2013 08:55 PM   
I have the exact same problem with my 650m gt. Also stuttering problem when vsync is on still persists as well as other random stutters. Reverted back to Asus driver 307.07 with gpu boost. After the last beta driver everything is broken.
I have the exact same problem with my 650m gt. Also stuttering problem when vsync is on still persists as well as other random stutters. Reverted back to Asus driver 307.07 with gpu boost. After the last beta driver everything is broken.

#3
Posted 11/20/2013 11:14 AM   
One thing NVIDIA should also check is the following issue - as soon as I turn off vsync in game the throttling stops i checked with 331.82 and previous driver. So many users report vsync related issues on the 600 series and no one from NVIDIA is taking action.
One thing NVIDIA should also check is the following issue - as soon as I turn off vsync in game the throttling stops i checked with 331.82 and previous driver. So many users report vsync related issues on the 600 series and no one from NVIDIA is taking action.

#4
Posted 11/20/2013 11:27 AM   
Update: I'm pretty sure that my GPU is throttling due to heat, it gets up to 93 degrees and my framerates drop. I have cranked up the laptop's fan and set it on a cooling plate and dropped the GPU temp to 59-60 Celsius while idle, which kind of concerns me. I think my laptop is either not able to cool enough, or there is something wrong inside the case that's blocking heat transfer. Ideas?
Update: I'm pretty sure that my GPU is throttling due to heat, it gets up to 93 degrees and my framerates drop. I have cranked up the laptop's fan and set it on a cooling plate and dropped the GPU temp to 59-60 Celsius while idle, which kind of concerns me. I think my laptop is either not able to cool enough, or there is something wrong inside the case that's blocking heat transfer. Ideas?

#5
Posted 11/29/2013 01:23 AM   
I only have framerate drops in some games most notably Bioshock Infinite with v-sync enabled, disabling v-sync fixes the issues for me. It's probably just a bug with newer drivers. I would just use an older driver or disable v-sync for the time being.
I only have framerate drops in some games most notably Bioshock Infinite with v-sync enabled, disabling v-sync fixes the issues for me. It's probably just a bug with newer drivers. I would just use an older driver or disable v-sync for the time being.

--Windows 7 64-bit | Windows 8.1 64-bit | Asus Z97-WS | i7 4770k @ 4.5ghz | 3x Evga GTX 680 4GB | 16GB GSkill @1866mhz 9-10-9-28 2T | Thermaltake Tough Power 1200w
--Windows 7 64-bit | Windows 8.1 64-bit | Gigabyte Z97x Gaming GT | i7 4790k @ 4.4ghz | 3x Evga GTX 970 4GB | 16GB GSkill @1600mhz 8-8-8-24 2T | Thermaltake 1000w
--Windows 8.1 | Asus Z87-A | i5-4670k @ 4.5ghz | 2x Sapphire R9 280x 3GB| 8GB GSkill @1600mhz 8-8-8-24 2T | Thermaltake 1000w
--Monitor: Samsung ToC 27" 1920x1080 60hz
--Windows 7 64-bit | Windows 8.1 64-bit| Clevo P170EM 17" 1920x1080 60hz | i7-3610QM | GTX 675M | 8GB 1333mhz
Nvidia Shield Portable & Tablet

#6
Posted 12/07/2013 04:07 PM   
Hey Guys! I turned off the BD Prochot feature using ThrottleStop 6.0, but my card still stucks at a ceratin clock rate. I tried to downgrade my driver, but it doesn't seem to work, even with maximum performance settings. I see that you have the same kind of problem. Is there a way, to let my GPU run at full clock rate? This situation pretty much sucks, I hope that someone knows a solution for this, because i've been messing around with this for 3 hours now. *Update: I installed the 295.55 driver, still no changes. What is the purpose of having a dedicated videocard, if my system doesn't even use it? It runs only at 270Mhz (gpu), and it won't change. When a game shows pre-rendered cutscennes it goes at 710Mhz, but when it comes to rendering 3D it just slows down. The temp is not going above 75 °C, i guess the cooling is good. Something is holding back. Do you know any trick about disabling this underclokcing feature?
Hey Guys!

I turned off the BD Prochot feature using ThrottleStop 6.0, but my card still stucks at a ceratin clock rate.
I tried to downgrade my driver, but it doesn't seem to work, even with maximum performance settings.

I see that you have the same kind of problem.

Is there a way, to let my GPU run at full clock rate?

This situation pretty much sucks, I hope that someone knows a solution for this, because i've been messing around with this for 3 hours now.

*Update: I installed the 295.55 driver, still no changes. What is the purpose of having a dedicated videocard, if my system doesn't even use it? It runs only at 270Mhz (gpu), and it won't change.
When a game shows pre-rendered cutscennes it goes at 710Mhz, but when it comes to rendering 3D it just slows down. The temp is not going above 75 °C, i guess the cooling is good. Something is holding back.
Do you know any trick about disabling this underclokcing feature?

#7
Posted 12/22/2013 12:15 AM   
I have found this in the PDF: nvidia-smi.1.pdf "Clocks Throttle Reasons Retrievesinformation about factors that are reducing the frequencyofclocks. Only on supported T esla devices from Kepler family. If all throttle reasons are returned as "Not Active"itmeans that clocks are running as high as possible. Idle - Nothing is running on the GPU and the clocks are dropping to Idle state. This limiter may be removed in a later release. Application Clocks Setting GPU clocks are limited by applications clocks setting. E.g. can be changed using nvidia−smi −−applications−clocks= SW Power Cap - SW Power Scaling algorithm is reducing the clocks belowrequested clocks because the GPU is consuming too much power.E.g. SW power cap limit can be changed with nvidia−smi −−power−limit= HW Slowdown - HW Slowdown (reducing the core clocks by a factor of 2 or more) is engaged. This is an indicator of: *temperature being too high *External Power BrakeAssertion is triggered (e.g. by the system power supply) *Power draw is too high and Fast Trigger protection is reducing the clocks *May be also reported during PState or clock change ** This behavior may be removed in a later release Unknown - Some other unspecified factor is reducing the clocks." The interesting thing is the HW Slowdown, which is doing the throttle. I'm curious about the too high temperature in numbers. How much °C is that "too high temperature"? Should I buy a notebook cooler? Or is there an unknown reason?
I have found this in the PDF: nvidia-smi.1.pdf

"Clocks Throttle Reasons
Retrievesinformation about factors that are reducing the frequencyofclocks. Only on supported T esla
devices from Kepler family.
If all throttle reasons are returned as "Not Active"itmeans that clocks are running as high as possible.
Idle - Nothing is running on the GPU and the clocks are dropping to Idle state. This limiter
may be removed in a later release.

Application Clocks Setting
GPU clocks are limited by applications clocks setting. E.g. can be changed using
nvidia−smi −−applications−clocks=

SW Power Cap - SW Power Scaling algorithm is reducing the clocks belowrequested clocks because the
GPU is consuming too much power.E.g. SW power cap limit can be changed with
nvidia−smi −−power−limit=

HW Slowdown - HW Slowdown (reducing the core clocks by a factor of 2 or more) is engaged.
This is an indicator of:
*temperature being too high
*External Power BrakeAssertion is triggered (e.g. by the system power supply)
*Power draw is too high and Fast Trigger protection is reducing the clocks
*May be also reported during PState or clock change
** This behavior may be removed in a later release

Unknown - Some other unspecified factor is reducing the clocks."

The interesting thing is the HW Slowdown, which is doing the throttle.
I'm curious about the too high temperature in numbers. How much °C is that "too high temperature"?
Should I buy a notebook cooler? Or is there an unknown reason?

#8
Posted 12/22/2013 12:10 PM   
i have the same problem with my gtx460m there is any fix for that?i have instat frame drop for 1 sec and gpu load drops and it does this way over and over after 10 15 seconds of normal gamplay and normal fps my temperature are not hitting 85c LE:i attach a print with my gpu throttling in action
i have the same problem with my gtx460m
there is any fix for that?i have instat frame drop for 1 sec and gpu load drops and it does this way over and over after 10 15 seconds of normal gamplay and normal fps
my temperature are not hitting 85c

LE:i attach a print with my gpu throttling in action

#9
Posted 01/29/2014 06:24 PM   
The same problém with my GT750 here including stuttering in games, even in low resolutions and minimal details. With v-sync enabled, games are unplayble. When I disable v-sync, games look horrible due to tearing. So NVIDIA, what should we do if we can not change our cards in laptops? We have been waiting too long...
The same problém with my GT750 here including stuttering in games, even in low resolutions and minimal details. With v-sync enabled, games are unplayble. When I disable v-sync, games look horrible due to tearing. So NVIDIA, what should we do if we can not change our cards in laptops? We have been waiting too long...

#10
Posted 01/31/2014 03:22 PM   
[quote="Nazgulgnome"]Specs: MSI gt683R with Intel Core i7 2630QM clocked at 2.0 ghz, 12gb of DDr3 memory, 2 500gb hard drives and a GeForce GTX 560M with 1.5gb internal graphics ram. The GPU usually runs around 80-90 Celsius. I first noticed this problem with Mechwarrior Online. I assumed it was a problem with that game, but none of the fixes mentioned made any difference for very long. I recently (last couple weeks) noticed the problem in other games as well. Problem is this: I will start with fully playable FPS (usually in the 40-60 range) then, with some regularity, my framerate will drop down to the 5-15 range, which in FPS games like MWO, makes the game pretty much unplayable. My FPS will stay low for about 15-30 seconds, then pop back up to playable levels again for a short time. This phenomena has started occuring in every game I play with a 3D component, from Arkham City to Minecraft. I used a monitor and noticed that there was a correlation between these FPS drops and my GPU load plummeting from nearly 100%, which I would expect with a full screen game with detailed graphics on a slightly older card, to as little as 0% in the logs. I don't know why the load drops like this, but the load drop coincides with the card temp going over 90 celsius. I can only assume there is a temperature related throttle somewhere. Any ideas on how to fix this? I only have the laptop to game on right now and no chance of buying anything better for at least a year.[/quote] Same specs, same problem here
Nazgulgnome said:Specs: MSI gt683R with Intel Core i7 2630QM clocked at 2.0 ghz, 12gb of DDr3 memory, 2 500gb hard drives and a GeForce GTX 560M with 1.5gb internal graphics ram. The GPU usually runs around 80-90 Celsius.

I first noticed this problem with Mechwarrior Online. I assumed it was a problem with that game, but none of the fixes mentioned made any difference for very long.

I recently (last couple weeks) noticed the problem in other games as well.

Problem is this: I will start with fully playable FPS (usually in the 40-60 range) then, with some regularity, my framerate will drop down to the 5-15 range, which in FPS games like MWO, makes the game pretty much unplayable. My FPS will stay low for about 15-30 seconds, then pop back up to playable levels again for a short time. This phenomena has started occuring in every game I play with a 3D component, from Arkham City to Minecraft.

I used a monitor and noticed that there was a correlation between these FPS drops and my GPU load plummeting from nearly 100%, which I would expect with a full screen game with detailed graphics on a slightly older card, to as little as 0% in the logs. I don't know why the load drops like this, but the load drop coincides with the card temp going over 90 celsius. I can only assume there is a temperature related throttle somewhere.

Any ideas on how to fix this? I only have the laptop to game on right now and no chance of buying anything better for at least a year.


Same specs, same problem here

#11
Posted 03/11/2014 10:09 PM   
Specs: RAM: 8.00 GB PROCESSOR: AMD FX- 8350 EIGHT CORE PROCESSOR GRAPHICS CARD: NVIDIA GEFORCE GTX 670 SYSTEM: 64 BIT WINDOWS 7 Im new to this and dont know what else to put, sorry. I seem to have the same problem, running all kinds of games, not just one. I am just like you, 40-60 frames at start and after 5-15min. it drops to 2-15 frames. Maybe a heat problem, but I do not think my temps are too high, 60c-70c on almost every component. Maybe something in the BIOS? I do realize this is an old topic, but was wondering if any of yall had an update.
Specs:
RAM: 8.00 GB
PROCESSOR: AMD FX- 8350 EIGHT CORE PROCESSOR
GRAPHICS CARD: NVIDIA GEFORCE GTX 670
SYSTEM: 64 BIT WINDOWS 7
Im new to this and dont know what else to put, sorry. I seem to have the same problem, running all kinds of games, not just one. I am just like you, 40-60 frames at start and after 5-15min. it drops to 2-15 frames. Maybe a heat problem, but I do not think my temps are too high, 60c-70c on almost every component. Maybe something in the BIOS? I do realize this is an old topic, but was wondering if any of yall had an update.

#12
Posted 06/21/2014 12:49 AM   
[quote="Nazgulgnome"]Update: I'm pretty sure that my GPU is throttling due to heat, it gets up to 93 degrees and my framerates drop. I have cranked up the laptop's fan and set it on a cooling plate and dropped the GPU temp to 59-60 Celsius while idle, which kind of concerns me. I think my laptop is either not able to cool enough, or there is something wrong inside the case that's blocking heat transfer. Ideas?[/quote] You should really check your fan(s) and try to get the dust out. I do this once every couple months or so and the temps are great when I finish. I use a can of compressed air that's made for dusting electronics, however be careful to not blow air at your fan too fast. I usually do short bursts until all dust is gone. This is for my old notebook btw, with GT 650m which has a single fan/copper pipe for both cpu and gpu. 59-60 degrees Celsius is really high for idle. A couple other things you might want to try after cleaning all dust, like propping the back of your laptop up an inch or so (it allows better air flow for some notebooks), also you may want to think about re-applying thermal paste to your gpu. A tube of good paste should run you 10 bucks or so, and you should notice a BIG difference in temps after that. Try youtube for a tutorial on your specific notebook.
Nazgulgnome said:Update: I'm pretty sure that my GPU is throttling due to heat, it gets up to 93 degrees and my framerates drop. I have cranked up the laptop's fan and set it on a cooling plate and dropped the GPU temp to 59-60 Celsius while idle, which kind of concerns me. I think my laptop is either not able to cool enough, or there is something wrong inside the case that's blocking heat transfer. Ideas?


You should really check your fan(s) and try to get the dust out. I do this once every couple months or so and the temps are great when I finish. I use a can of compressed air that's made for dusting electronics, however be careful to not blow air at your fan too fast. I usually do short bursts until all dust is gone. This is for my old notebook btw, with GT 650m which has a single fan/copper pipe for both cpu and gpu.

59-60 degrees Celsius is really high for idle. A couple other things you might want to try after cleaning all dust, like propping the back of your laptop up an inch or so (it allows better air flow for some notebooks), also you may want to think about re-applying thermal paste to your gpu. A tube of good paste should run you 10 bucks or so, and you should notice a BIG difference in temps after that. Try youtube for a tutorial on your specific notebook.

Asus ROG G750JM | Win 8.1 Pro x.64 | Core i7 4700HQ | GTX 860m 2gb gddr5 (Maxwell) | 16gb DDR3L Dual Channel 1600 | 256gb SanDisk Ultra Sata III SSD | 5400rpm 1TB HDD |

Dell Inspiron 17R SE (7720) | Win 8.1 Pro x.64 | Core i7 3630QM | GT 650m 2gb gddr5 | 8gb DDR3 1600 | 128gb Mushkin Atlas Sata 3 mSata | 1TB 5400rpm HDD | 500gb 7200rpm HDD |

Desktop | Gigabyte 970FX | AMD FX-8150 4ghz Watercooled | 16gb ram 1600mhz | EVGA Dual Superclocked GTX 770 4gb gddr5 | 850w EVGA Gold Plus PSU | 2TB HDD | 256gb SSD | (Upgrading: Work in Progress)
My Nvidia GPUs

#13
Posted 06/22/2014 04:36 AM   
I wanted to come back and offer some more help towards getting your notebook's temperatures under control. I did a quick search and I found a good Youtube vid with instructions on how to re-apply thermal compound. [u][b]Please remove your Battery before you begin[/b][/u], as the guy in the video did not lol. [center][url=https://www.youtube.com/watch?v=3sXAB7DI_NA]Please check it out HERE[/url][/center] [i](Just an FYI if your wondering: The MSI GT780 and your GT683 are almost identical, from what I've seen, in terms of dis-assembly)[/i] *Also you use Isopropyl Alcohol (rubbing alcohol) and cue-tips to clean off the old paste, from the tops of both the cpu and gpu, [i][b]AS WELL AS[/b][/i] the bottoms of the heatsinks. You want to get rid of [u]ALL[/u] of the old paste.(Spotless) The guy in the video wasn't very clear on that. **One last thing, the guy in the video used the 'spread' method to apply the new thermal paste on the cpu and gpu which is not the best way. Personally, I believe that just placing a pea sized dot of the paste on the top / middle of both the cpu and gpu, and then letting the heat-sinks spread it automatically when you screw them down, to be the correct way of applying thermal paste. Good Luck :)
I wanted to come back and offer some more help towards getting your notebook's temperatures under control. I did a quick search and I found a good Youtube vid with instructions on how to re-apply thermal compound. Please remove your Battery before you begin, as the guy in the video did not lol.




(Just an FYI if your wondering: The MSI GT780 and your GT683 are almost identical, from what I've seen, in terms of dis-assembly)

*Also you use Isopropyl Alcohol (rubbing alcohol) and cue-tips to clean off the old paste, from the tops of both the cpu and gpu, AS WELL AS the bottoms of the heatsinks. You want to get rid of ALL of the old paste.(Spotless) The guy in the video wasn't very clear on that.

**One last thing, the guy in the video used the 'spread' method to apply the new thermal paste on the cpu and gpu which is not the best way. Personally, I believe that just placing a pea sized dot of the paste on the top / middle of both the cpu and gpu, and then letting the heat-sinks spread it automatically when you screw them down, to be the correct way of applying thermal paste.

Good Luck :)

Asus ROG G750JM | Win 8.1 Pro x.64 | Core i7 4700HQ | GTX 860m 2gb gddr5 (Maxwell) | 16gb DDR3L Dual Channel 1600 | 256gb SanDisk Ultra Sata III SSD | 5400rpm 1TB HDD |

Dell Inspiron 17R SE (7720) | Win 8.1 Pro x.64 | Core i7 3630QM | GT 650m 2gb gddr5 | 8gb DDR3 1600 | 128gb Mushkin Atlas Sata 3 mSata | 1TB 5400rpm HDD | 500gb 7200rpm HDD |

Desktop | Gigabyte 970FX | AMD FX-8150 4ghz Watercooled | 16gb ram 1600mhz | EVGA Dual Superclocked GTX 770 4gb gddr5 | 850w EVGA Gold Plus PSU | 2TB HDD | 256gb SSD | (Upgrading: Work in Progress)
My Nvidia GPUs

#14
Posted 06/23/2014 03:50 AM   
Scroll To Top