all the other bug checks seem to be 1001.
I don't understand how a hardware issue would only affect ED, and only on the first launch, but thanks for looking into this.
Requesting advice from computer wizards
- Thread starter SlickPBW
- Start date
- Status
all the other bug checks seem to be 1001.
I don't understand how a hardware issue would only affect ED, and only on the first launch, but thanks for looking into this.
That's not where I'm seeing the 116. You need to look in the description.
The hardware issue is always there. This game just happens to be loading the system in the right way to reveal it in an obvious manner.
ED is far more demanding that most people give it credit for. Indeed, since Path of Exile was patched, ED is the most GPU heavy game (quantifiable by how much power it will pull when allowed to run uncapped) I have played; and I play many newer and more visually appealing titles.
Those BugCheck restarts should never happen on a stable system. Every one of those bugchecks is likely evidence of an underlying hardware problem or a serious miscofiguration somewhere. Software shouldn't be able to cause them without underlying issues, and Elite: Dangerous (including Odyssey) is fully capable of running on a system with the same models of components you have without experiencing these issues.
As for why the crash would only be occurring on the first launch, transient loads are some of the worst and if the the work the game doing hasn't been done that session, it's not going to be cached and will be that much more likely to cause problems. The software is likely more or less stable, but the load it's putting on some components is beyond something's stable operating margins. That's why looking at the details of the errors is useful...with it you can narrow down the likely source of the problem, and then some testing that can more reliably induce problems with specific areas, to confirm or falsify the hypothesis.
For the record, this is what a 116 BugCheck is:
Bug Check 0x116 VIDEO_TDR_FAILURE - Windows drivers
The VIDEO_TDR_FAILURE bug check has a value of 0x00000116. This value indicates that an attempt to reset the display driver and recover from a timeout failed.
learn.microsoft.com
Multiple TDR failures in a short period of time will trigger that BSOD.
It's possible that it's a purely software/driver hold-up, but unlikely. You can increase the TDR window with a registry entry if you only care about addressing the symptoms and don't care to test for potential underlying issues that would cause the GPU driver to need to be reset.
Testing and Debugging TDR During Driver Development - Windows drivers
TDR testing and debugging information for developers
learn.microsoft.com
Increasing the delay to 8-10 seconds is the rational maximum. The error detection will still function, but the GPU driver will have more time to respond if it's simply an unusual spike in load causing issues, and not some underlying hardware problem. If you need more than a 10 second delay, something is probably broken and should probably be addressed before things get worse. However, if you're a "jesus take the wheel" type, just set the debug mode to TDR_DEBUG_MODE_IGNORE_TIMEOUT.
Last edited:
to ask another silly question is pc power set to maximum performance?
as being a pettie dabbler that plus the latest driver is what i would try to give the pc a kick up the jaksie
as all work an no play makes jack a dull boy
as being a pettie dabbler that plus the latest driver is what i would try to give the pc a kick up the jaksie
as all work an no play makes jack a dull boy
Last edited:
Roger. I looked through all the BugChecks: Total of 42 entries; there were 6x 119 and 1x 007, and the remainder were 116.
Thanks for the power explanation; I never considered that power would ever be an issue. It does make sense with other observations I have. Such as: I frequently get the crash when opening the gal map. If there's no crash, I still see a pretty big drop in FPS when opening either the galmap or system map, so there must be a pretty big load on the system with those. Now I'm wondering if there are any in-game options in the maps to lower the load. I'm currently defaulted to the Thargoid War view, and I would presume that has a lot more info required to load than some of the the other views, so I'm gonna change the default view and see if that helps.
Edit: Since these crashes are looking like an overpower demand situation, could this be causing damage?
I'd like to proceed to resolve this if possible, and if not, I'll follow your TDR advice above.
Good question. I don't remember setting that, but I checked and it's already set to "High Performance". The only other options are "Balanced" and "Power Saver" so I didn't change anything here.
Last edited:
You can open that DMP file with the windows debugger (link below from microsoft), but this would be quite technical
Install WinDbg - Windows drivers
Start here for an overview on the Windows debugger and installing WinDbg.
learn.microsoft.com
All of these are GPU related. Could be the driver, could be conflict with the Rift, could be power causing an issue with the Rift, could be a defective GPU. Could also be system memory, the memory controller, or the system agent/PCI-E controller as any data to or from the GPU would have to pass through these. Likewise, an issue with power delivery could underpin any of these.
A more complete list of the hardware in your system may help. Exact motherboard, GPU, memory, and power supply model and revision numbers especially. Precise age of the PSU too.
Run the sensor part of HWiNFO and check that the motherboard and GPU are reporting correct input voltages. Transient response problems are quick enough need an oscilloscope to observe, but static voltages will at least rule out the most obvious problems. Assuming your basic voltages appear normal (very close to the specified +3.3v, 5v, and 12v for the primary rails and very near 12v for the PCI-E input voltage on the GPU), you'll want to do some basic stress testing.
Get OCCT, and run the power stability test. This will hit the system extremely hard and should induce a crash if it's a clear power issue. If you can do 15-20 minutes of the power test without problems, run the CPU and then memory tests for the full hour each, if you can. If you encounter errors, that's a hardware problem, or incorrect firmware settings.
There are numerous ways you can force ED to use less power. However, these will only mask the issue, so should only be considered a temporary solution if a power problem is indicated. A frame rate cap will have the most profound impact, but if you're already capped and reaching that cap, reducing grapical detail will further reduce power demands. You can also use tools like NVIDIA-smi or MSI Afterburner to reduce the alloted power limit of the GPU.
I'd also disconnect your Occulus Rift and remove the software for it; you can reinstall it later, after you've sorted things.
As long as the specified limiters are in place there shouldn't be any practical way for power demand to be problematic, unless there is an issue with the supply. But yes, dirty power can cause excess wear on devices, or quickly kill them outright, in extreme cases. If you do identify a PSU problem, have a low quality PSU, or have a PSU that has simply worn out over time, you will want to replace it unless you want to run the risk of damaging anything and everything attached to it.
I just wanted to reiterate thanks to all those who have been helping me resolve this issue so far. It may seem like it's dragging out but I am juggling a bunch of other RL commitments right now including upcoming travel, so I will be scarce here for the next 2 weeks or so. After that, I'll pick up this troubleshooting again and hopefully find the root problem.
Probably not enough power from the psu. Graphics demands go up gpu draws more power, not enough and your system crashes and reboots.
You can try capping the fps to 60 either in-game or in nvidia control panel.
One way to see is installing msi afterburner and enabling the on screen display to show fps and power consumption and see what it says before the crash, I reckon power usage will start to spike before a crash dont screw around with other settings in afterburner
You can try capping the fps to 60 either in-game or in nvidia control panel.
One way to see is installing msi afterburner and enabling the on screen display to show fps and power consumption and see what it says before the crash, I reckon power usage will start to spike before a crash
dont screw around with other settings in afterburnerI know you might not see this for two weeks but I ain't reading all the history again in two weeks' time so I'll comment now.
My instinct reading your first comment, even before I read all the replies was it's probably power supply / voltage stability issues; so this is a vote to stick with that theory.
HOWEVER it's very intriguing that the situation improved when you first disabled some of the VR things, so somehow the presence or absence of the Rift is affecting things, although unfortunately not in a yes/no way.
Morbad already suggested uninstalling everything to do with the Rift. Before you do that use Device Manager to figure out which USB bus the Rift is connected to; physically disconnect every other device on that bus too; then disable that bus in Device Manager. This has worked for a lot of sim racers with VR problems.
If that removes the problem then you'll have to post back here with more details about how you have all the USB devices rigged and we can figure out a better combination of connectors and hubs to work around the problematic USB area.
Finally, and related to that, I haven't seen anyone mention motherboard firmware and there was a period a couple of years back where a lot of motherboards had voltage management problems when under stress, and also problems with the USB chipset, so it would be worth looking up your current motherboard and seeing what the known issues are and whether there are any updates with related fixes in the release notes.
My instinct reading your first comment, even before I read all the replies was it's probably power supply / voltage stability issues; so this is a vote to stick with that theory.
HOWEVER it's very intriguing that the situation improved when you first disabled some of the VR things, so somehow the presence or absence of the Rift is affecting things, although unfortunately not in a yes/no way.
Morbad already suggested uninstalling everything to do with the Rift. Before you do that use Device Manager to figure out which USB bus the Rift is connected to; physically disconnect every other device on that bus too; then disable that bus in Device Manager. This has worked for a lot of sim racers with VR problems.
If that removes the problem then you'll have to post back here with more details about how you have all the USB devices rigged and we can figure out a better combination of connectors and hubs to work around the problematic USB area.
Finally, and related to that, I haven't seen anyone mention motherboard firmware and there was a period a couple of years back where a lot of motherboards had voltage management problems when under stress, and also problems with the USB chipset, so it would be worth looking up your current motherboard and seeing what the known issues are and whether there are any updates with related fixes in the release notes.
A small but significant update.
I will continue to be scarce for at least another week due to being on travel but there were some developments before I left.
I continued to have crashes as previously described, but they started to increase in frequency. They have started happening in other graphically demanding games. Long story short, I have decided to replace my nearly 10 year old computer and the new one will be here in a couple of weeks. In the past I have always waited too long to replace computers and have suffered catastrophic hard drive failures before replacing, and this made restoring data from backups to the new computer harder than it needed to be so I’m not going to push my luck any further.
I have been researching the methods to transfer ED data to a new computer. There are several guides out there, some quite old, and some a little conflicting. I know how to get the launcher and run it on the new computer. Correct me if I’m wrong, but I also believe that all my important game information resides on FD servers with 4 exceptions: visited stars cache, journal files, key binds, and screenshots.
My questions are:
- after installing ED on the new computer, I am unsure if the sequence of running the newly installed game and transferring the game files requires a specific order. Specifically, should I copy the files over before or after launching the game for the first time? If not, I would expect ED to create these files as new files, so at that point would it just be a matter of copying the new files over and archiving the new ones? and if so, would that cause any problems with old/large files that have entries that predate the new installation? So, the bottom line question is: what is the best order of operations to install the new game and transfer files? (I realize that screenshot copying won’t make any difference, so this applies to the other 3)
Thanks in advance for any advice on this.
I will continue to be scarce for at least another week due to being on travel but there were some developments before I left.
I continued to have crashes as previously described, but they started to increase in frequency. They have started happening in other graphically demanding games. Long story short, I have decided to replace my nearly 10 year old computer and the new one will be here in a couple of weeks. In the past I have always waited too long to replace computers and have suffered catastrophic hard drive failures before replacing, and this made restoring data from backups to the new computer harder than it needed to be so I’m not going to push my luck any further.
I have been researching the methods to transfer ED data to a new computer. There are several guides out there, some quite old, and some a little conflicting. I know how to get the launcher and run it on the new computer. Correct me if I’m wrong, but I also believe that all my important game information resides on FD servers with 4 exceptions: visited stars cache, journal files, key binds, and screenshots.
My questions are:
- after installing ED on the new computer, I am unsure if the sequence of running the newly installed game and transferring the game files requires a specific order. Specifically, should I copy the files over before or after launching the game for the first time? If not, I would expect ED to create these files as new files, so at that point would it just be a matter of copying the new files over and archiving the new ones? and if so, would that cause any problems with old/large files that have entries that predate the new installation? So, the bottom line question is: what is the best order of operations to install the new game and transfer files? (I realize that screenshot copying won’t make any difference, so this applies to the other 3)
Thanks in advance for any advice on this.
I recommend having a backup of everything named Frontier* somewhere safe, in case it turns out that there's more things you want restored, or you screw up somehow.
It shouldn't make a difference. It might however be easier to start the game once and let it create stuff. That way you'll more easily see where exactly the files need to go.
Don't worry about archiving any new files. Just put the old ones in the correct places (overwriting new ones as necessary).
When I did it ages ago I let the installer create the required folders before cancelling then copying the files over. I dunno if it also refers to any registry keys to run ED but letting it make those would be the best move.
I had an issue with Vr/WMR using a HP headset that if Touch keyboard service wasnt running the memory usage would continue to rise and not release until the system became unresponsive and would crash. I could watch the memory usage rise in task manager until the system froze, turn on touch keyboard at any time and it would instantly release the memory and all was good, putting it out there it may help someone in the future.
Robert Maynard
Volunteer Moderator
Thread closed at the request of the OP.
- Status