GPU usage drops to 0% for 5-10 seconds

[icecoldmilf]

Hello, I currently have a issue with gaming and somethings even playing youtube videos on 1080p/60 fps.

When I'm playing for instance League of Legends I'm getting around 150 fps, everything is going fine until my gpu usage drops to 0%, my fps will be like 10-40ish. The cpu usage is around 35% and will go up to 50% when my gpu is dropping to 0% during that time. I heard about bottleneck and I think my gpu needs a upgrade, but I would like a confirmation first before I buy a new gpu.

Specs:
Motherboard - ASRock 970 Extreme4
CPU - AMD FX-8350 @ 4.00 Ghz
GPU - XFX Radeon HD 6870
RAM - 8GB DDR3-1600
SSD - Samsung SSD 840 EVO 250GB ATA Device

This issue is really frustrating me, any help is appreciated.

 

[sz0ty0l4]

can be lot of issues😛ci-e connection issue, driver corruption, temperature /heat issue, psu issue, gpu stability issue( check clocks and voltages) and hmm can also be a mix of these.

To start troubleshooting i'd advice to download hwinfo64 and check windows event logs and also check dxdiag.
PLease list the logs you found can be linked to the issue( check log times etc).
please gather values of: 12v rail from hwinfo, gpu clock speeds, gpu core voltage, pci-e speed. best would be to keep hwinfo running when the issue happens, because it automatically does a sequential measurement and saves the data continually. then you can list min/max values.

 

[icecoldmilf]

can be lot of issues😛ci-e connection issue, driver corruption, temperature /heat issue, psu issue, gpu stability issue( check clocks and voltages) and hmm can also be a mix of these.

To start troubleshooting i'd advice to download hwinfo64 and check windows event logs and also check dxdiag.
PLease list the logs you found can be linked to the issue( check log times etc).
please gather values of: 12v rail from hwinfo, gpu clock speeds, gpu core voltage, pci-e speed. best would be to keep hwinfo running when the issue happens, because it automatically does a sequential measurement and saves the data continually. then you can list min/max values.

Thank you for the quick reply, I really appreciate it!

I've gathered some data playing a match on league on legends.

12v rail
- min 11.774 V
- max 12.038 V

gpu clock speed
- min 100 MHz
- max 900 MHz

gpu core voltage
- min 1.175 V
- max 1.175 V

pci-e speed
- min 5.000 Gbps
- max 5.000 Gbps

During this match I encountered the problem only once but normally when I play for a longer time I get this issue more often than once.

Also the GPU Thermal Diode reached 76 ℃, is that fine or should I be worried about this temperature?

I hope the data I provided is the correct, because this is the first time I use the program

 

[sz0ty0l4]

Yes but the problem is it doesn't actually point to anything. are you sure these are the values what got monitored when the issue happend?
76C is fine by the way.
For me it seems like both the gpu and the psu may be fine.

Can you check the windows event logger and look for suspicious events connected to the time when the issue happens? there should be something to start with. also look for events related to memory.

I suppose you have 2x 4gb memory stick right? You should try running the game with only 1x 4gb memory stick at a time and see if the issue comes up again. check for both sticks.
You should also run memtest on each stick.

Let me know what you found in the event logger and also how did it work out with the memory testing.

Ohh and one more thing. what's actually happening to the cpu when the issue occures? I mean you wrote it starts working harder, like 50%usage from 35%. what is the min max temperature, min max voltage, core clock on the cpu?

 

[icecoldmilf]

Yes but the problem is it doesn't actually point to anything. are you sure these are the values what got monitored when the issue happend?
76C is fine by the way.
For me it seems like both the gpu and the psu may be fine.

Can you check the windows event logger and look for suspicious events connected to the time when the issue happens? there should be something to start with. also look for events related to memory.

I suppose you have 2x 4gb memory stick right? You should try running the game with only 1x 4gb memory stick at a time and see if the issue comes up again. check for both sticks.
You should also run memtest on each stick.

Let me know what you found in the event logger and also how did it work out with the memory testing.

Ohh and one more thing. what's actually happening to the cpu when the issue occures? I mean you wrote it starts working harder, like 50%usage from 35%. what is the min max temperature, min max voltage, core clock on the cpu?

Alright im gonna try these things for today, I've been playing for around 2 hours today and the issue has still not come up.

I was checking the windows event logs and I saw this: The Autodesk Content Service service terminated unexpectedly. It has done this 1757 time(s). The following corrective action will be taken in 10000 milliseconds: Restart the service.

This error came up every 10 seconds and was counting up, so I uninstalled autocad and so far It's going great.

I will continue to play and keep hwinfo up and also checking the event logs, I hope autocad was the problem but I am not sure.

*I've been playing for around 3,5 hours now and I didnt encounter the issue yet! Hopefully it stays this way.

**2 Days past by and the gpu usage issue didn't come up. I think the problem was the AutoCAD service trying to restart every 10 seconds.

Thank you very much without you I would never knew about the windows event log !