@rozsy said:
@bloodsentinel: "Next go into your OS advanced power options. In Win10 I'm talking about right clicking on the windows button --->Power Options ---> Change Plan Settings ---> Change advanced power settings. In the new dialogue box scroll down to the PCI Express entry and expand it to revealLink State Power Management. Toggle this setting to off."
That exactly is this good for? And it happened during playing a game, i tried to actualize drivers but it didnt help. I tried render test for GPU and i had PCI-E 3.0 x16 @ x16 1.1 and during test it was PCI-E 3.0 x16 @ x16 3.0
The Link State Power Management is a windows power saving feature where the OS tries to detect when the GPU is idle and downclock it to save power. However, it's a pretty buggy power management code which does not always work and it can trigger an idle state even when the graphics cards is being used.
Let me see if I can dissect this into a way which makes sense so you know what all of this means:
PCI-Express controller version --->[PCI-E 3.0], [x16] <--- max PCI-E link width supported by motherboard, [@ x16]<--- current link width between GPU and motherboard, [1.1 - 3.0]<--- current PCI-E speed for the GPU.
When the GPU is idle it will be "3.0 x16 @ x16 1.1" and when the card starts to render something it will jump up to "3.0 x16 @ x16 3.0" so all is well. As long as it remains 3.0 x16 @ x16 then you're good.
The 1.1 jumping up to 3.0 is simply the graphics card's internal power saving feature turning off when the card is under load. This is normal.
You only want to be concerned with the second x16immediately following the "@" symbol. If this number fluctuates to lets say @ x8 or @ x4 or even @ x1 then you have a problem.
Here is the troubleshooting steps I would take to identify the nature of your problem:
**[1] Swap graphics cards in the PCI-e slot. If you have a secondary graphics card - a backup - somewhere in your house then uninstall the one currently in your system and delete all driver software for that device- use latest version of Display Driver Uninstaller. Then install the other graphics card into the PCI-e slot and install the latest drivers for that card. Try to game and see if it locks up while using a different GPU. This will help determine if the GPU is the cause or not for the problem.
**[2] Uninstall and remove graphics card to boot using motherboard's integrated video. Uninstall your graphics card and if your motherboard has an integrated video controller boot the system using it. Now try to game using the onboard video and see if the problem persists. If it does then your motherboard is bad and would be the cause of your problems.
[3] Ensure you are running the latest BIOS version for your motherboard. If not it's time to flash it.
[4] Remove any custom user settings in BIOS. Have you overclocked anything? If so then it's time to undo those changes. Upon startup go into your BIOS and load optimized defaults.
[5] Uninstall graphics driver and do a clean driver reinstallation. I talked about this in an earlier post - again use latest version of DDU.
[6] Get a temp reading during gaming. Find a good temp monitoring / reporting app that can record temps in a log file before a system failure. I can't think of one offhand, but they are out there. Run the app to monitor and record temps while gaming then when the problem happens go back to the log file the app created before the system lockup / crash. Look at temps for both your GPU and CPU since it may not be your graphics card overheating.
**Optional steps. Only do these if you have the means to accomplish them.
From everything you have told me the impression I'm getting is your problem is boiling down into one of three scenarios:
[1] Your system is overheating.
[2] Your motherboard is on its last leg.
[3] Your GPU is failing.
Now it's up to you to do extensive troubleshooting to isolate which is the cause.
Log in to comment