GT80 Titan... "PCI-E 3.0 x16 @ x8 3.0"... and game freezes

Hi all,

When I run GPU-Z, it is telling me that both of the GTX 980M cards are running at x8. It reports this even under heavy load.

Is it correct that the 980M should be communicating with the MS-1812 motherboard using only 8 PCI-E lanes each?

I ask because I am trying to diagnose an issue that is driving me crazy! Every single game I play... whether it is GTA V, or Rayman Origins... every now and again, the game video will freeze for about 5 seconds, but the audio will continue to play. The video will resume where it left off.

I am not performing any overclocking of the CPU or GPU.

Thanks.

 

Could be a read/write issues causing that, espeically if this is happening when changing map areas. Is the game installed on the HDD or the SSD?

 

Thanks for the reply.

It happens to games on both SSD and HDD... and it happens to all my games (and I play quite a few). For example, it has just been messing up in Castlevania: Lords of Shadow. Not a heavy game, on either the HDD or the GPU.

My guess is that it is something to do with the way the GPU hits the utilisation limit... or maybe the CPU is throttling or something? But I have no idea where to start in diagnosing this.

I thought it might have been the monitoring of Afterburner... but I tried it just now, and it still does it.

It never does it at the start of a game - its always once I have been in the game about 10-15 minutes.

It is impossible trying to play racing sims when it is like this...

Any other ideas will be greatly appreciated.

Thanks.

 

I should also note it isn't when changing map areas. Sometimes, I will be turning a bend... in the middle of a fight... or just walking through a forest. There really is nothing that sticks out as being the cause.

 

Have monitoring software running while gaming so that we can get an idea of what is causing it. We need to monitor CPU, GPU, HDD all with current usage, temps, speeds etc. As well as which antivirus you are running.

8 xPCIe 3.0 is normal because there are only 16 lanes available for PCIe 3.0, and they are using half and half for each GPU. It doesn't diminish your performance so its not that. I have run many games, demanding and not, and they run flawlessly.

Also, keep in mind it has been known that the Killer Networks application causes issues with games. I experiences stuttering and performance degradation at times because of it. Try gaming by first closing the killer software (your network drivers still work, just no program running in the background doing any bandwidth control).

The issue you mention seems very similar to HDD spinning down. I had this issue with my old GT60 because the HDD would spin down to sleep when idle, and when it resumed operation, it resulted in those random freezes. the trick was to keep the HDD alive, and as a test i kept HW monitor running, monitoring the temp of the HDD, and my freezes went away.

 

Thanks for the reply.

I am running Avast Antivirus. Glad to hear that the PCI-E thing is doing the right thing though.

I exited all apps in my system tray (inclusive of the Killer app, but exclusive of the antivirus)... and I still got the problem.

I always game with the "Sport" mode on... and I read something before about this stuff can happen when the PC is drawing more power than the PSU can give it. Could this be a factor? I haven't been able to find an application that monitors power consumption... and the MSI Dragon app only shows you what it is doing at that time (I need to know what is happening when I am gaming).

I usually have AfterBurner running all the time - but I stopped using it to test this out (it wasn't AfterBurner at fault).

I also removed the "Get Windows 10" taskbar icon by uninstalling the Windows Update that put it on my machine... this did nothing either. I removed this because people had reported performance issues as a result of that utility.

Can you recommend the best utility to use to monitor all of this stuff?

Thanks.

 

I have reset the power options to all my power plans back to the defaults, and I reset the NVIDIA "global" settings back to defaults also. So I should now be using it in the "Sports, with High Performance Plan" mode exactly as per Windows and NVIDIA settings. I still have the problem.

 

As suggested you should post up pictures of graph logs of it happening so we can look at the data ourselves.

 

Graph logs from which software? I'm not too well up on this side of things...

 

Along with MSI Afterburner, you can use HWmonitor, Intel XTU to cover GPU and CPU, but I don't have much experience with software that can monitor HDD activity.

Have you tried a complete driver reinstall of your GPUs? A clean install with the latest drivers to ensure everything is up to speed. Maybe someone else could chime in to confirm. I don't think it is the PSU, but I could try testing with games and disconnecting my PSU to see how it behaves, just to simulate power loss. I haven't had problems when gaming except for the ones caused by the Killer software.

It could be some sort of hardware issue too. If we can't find the answer maybe the best will be to setup RMA.

 

Intel XTU and MSI afterburner are good starts, use the spanner at the top right of the monitoring on XTU to add in throttle reasons and power usage levels etc.

 

Thanks. I will look into those. In the meantime, I have recorded a log file in Afterburner, of all the things that it can log, with a 250ms interval. Is it OK to post that file here? Until I can get the full thing to you guys, at the point it happens, the following happens:

a) usage, FB usage, bus usage... for both GPUs... drop to 0%
b) frametime spikes up to 2805ms
c) framerate drops to 0fps, then rockets up to 110fps (strange, since vsync is turned on, at 60Hz)

The max CPU cores temp recorded was 83C.
The max GPU temp recorded was 45C.
The max CPU core usage was 89%.
The max GPU usage was 48%.

All other stats (SLI sync limit, GPU utilisation limit, etc) stay constant throughout the incident).

This is baffling.

 

Hmmm temps are well under any limit. Have you tried a clean driver install for your GPUs?

 

I will try it now and see what happens. Thanks.

 

I used the DDU v15.1.0.3 to completely remove the driver. I then tried to install v353.06 NVIDIA driver. However, both times I have tried, it is failing to install ShadowPlay. I have had a dig around, and I get to a page that says that this may be happening due to a corrupted DirectX installation! If this is the case... it may also have been contributing to my issue. I will follow that guide to correct my DirectX, hopefully get the drivers to install properly, and then report back.

Apologies for being verbose... but I am hoping that if we get to the bottom of this, it may help someone else someday.

 

I couldn't get ShadowPlay installed successfully (according to the NVIDIA installer)... but turning on ShadowPlay seemed to work OK. Don't really care about ShadowPlay anyway.

Regarding my problem... it still exists. Frame time was 2784ms. It is darned close to the one from last time!

I will look into getting those other logs together, in case they can provide any further information.

Thanks.

 

Hmmm, it could be hardware related, and you might need to setup RMA.

 

I took pics of all the different things that the Intel XTU monitored. However, I am not sure exactly what they are showing. The glitch in the game happened close to the end... but I think the couple of dips may be due to alt-tabbing and pausing the monitoring.

 

I think I may have identified what the cause of the problem is... but I am not sure what is occurring to make it freeze the game for 3 seconds.

Usually, I am lucky if I can do 4 laps on Project Cars before it stops for 3 seconds. Today, I managed 27 laps and it didn't freeze... and late last night, I managed a few full races (including practice session, qualifiers, and the actual race).

The thing that I think is causing it: my external hard drive. I unplugged it, and everything seemed to be OK.

Have you guys and gals heard of anything happening before regarding external hard drives? I don't really want to be having to unplug my external hard drive to do my gaming, as I have a lot of games installed on there (games that don't rely heavily on loading lots of assets).

Your comments and thoughts appreciated, as always.

Thanks.

 

It is, as we mentioned before, most likely cause when the hdd, be it external or internal, goes to a stand by/sleep mode and then awakes to be used again for any reason, resulting in freezes.

If its indeed that, then perhaps something can be tweaked for the hdd to be always on? Maybe confirm the integrity of the hdd as well.

Good that you found the cause

Sent from my Nexus 9 using Tapatalk

 

Yes there are:

http://www.ghacks.net/2009/05/18/prevent-power-saving-mode-of-external-hard-drives/