@WhiskeyTourettes wrote:Following up. After extensive testing I can confirm nothing to do with software whether its the OS, DirectX, drivers, or repairing the game matters.
On jedha you will get device hung errors at some point most likely.
As suspected in my last post. An increase of SOC voltage directly effects the frequency of my crashes on Jedha running 4k epic with Ray tracing. I am now running 1.30v SOC. The max I can on my 7950x3d. 8 hours on Jedha 1 crash in cutscene. At 1.25 SOC crashes every 5-15 minutes. My RAM is doing nothing special. Standard 6000mhz CL30 everyone runs. Not really overclocking my CPU. You can't do much of that anyway with an x3D. You can only change the voltage curve.
Temps are all fine. Rock solid stability in everything else. Yet there is an improvement when I up SOC and a drastic one at that. So, who is to blame here? AMD and Intel for releasing unstable hardware that can't really run pcie express 4.0 GPUs at safe SOC or the developer? To me quality assurance is bad right now on PC hardware. AMD let their chips be released at 1.4 SOC set as default on motherboards then chips starting combusting and now they just pull 1.3SOC out of their *? Who really knows if that is safe either but I need to run 1.30 SOC to get through Jedha and even then I experienced this error at least once so far. I will try increasing my VDDIO again up to 1.4v Nothing I can do now about the SOC unless I want blow my chip up.
Jedha is the ultimate gaming rig stress test for some reason. Whether its really respawns fault or not it would be nice if somehow their was an update to make this level run more friendly with RT on. There are seemingly hundreds of people on forums with this issue at a minimum. Not to mention unreported cases.
Well I got to fight a surprising main boss 15 times. Not because the boss was too hard but because the game kept crashing halfway through the fight. I lowered my RAM down to base JEDEC voltage timings and frequency. In other words the standardized spec that should basically always be stable. Still getting this error. I cant increase my SOC voltage to stabilize the PCI express signaling anymore. I dont know what to do at this point. Either game is broken or certain hardware can't run this game stable. I see alot of people on the latest gen hardware both Intel and amd complaining.
I turned Ray tracing off and I was able to complete this boss on Jedha. Im not saying the name of the boss so I dont spoil anything but im sure the tech support could take a guess here and get it right.
Bottom line this is the display driver crashing. You will find corresponding nvddm errors and windows reporting errors. All survivor is doing is reporting the display adapter was removed. This really need to be fixed by Nvidia.
Now survivor could more gracefully handle this error by pausing game and waiting for driver to reappear.
i think most of the fixes on this and other threads for this error are false fixes, it seems that rebooting the computer is effective for me as any other fix. This could indicate a time base / cache issue. But unclear.
@scyto wrote:Bottom line this is the display driver crashing. You will find corresponding nvddm errors and windows reporting errors. All survivor is doing is reporting the display adapter was removed. This really need to be fixed by Nvidia.
Now survivor could more gracefully handle this error by pausing game and waiting for driver to reappear.i think most of the fixes on this and other threads for this error are false fixes, it seems that rebooting the computer is effective for me as any other fix. This could indicate a time base / cache issue. But unclear.
For many a reboot does not fix it. The problem comes right back after a reboot quickly. The reality IMO is that modern hardware is not working correctly for this game for whatever reason. Its not a bad graphics card or RAM. It is related to the SOC voltage as the problem is clearly reduced for me when pumping more SOC voltage. Why only this game seems to overload my rig I have no idea. I can run the adaptive OCCT graphics test all day long and it not crash at lower than max SOC voltage. The OCCT adaptive stress test is specifically designed as a simulation for Unreal Engine 4 which this game uses. I can run RAM stress tests for days and never have an issue. CPU stress tests the same. This has to do with an error with pcie express bus signal integrity which relates to SOC. Even running my system completely stock with the only thing being touched is SOC voltage. This game crashes on Jedha with this device hung crap. Turning off RT solves it. Respawn needs to take a look at specifically ray tracing on Jedha IMO. I think most people with this issue are using RT. For those that are not I suspect all they need to do is increase their SOC voltage and they will likely be OK. I am an X3D owner and I can't tune my SOC any higher than 1.30v. I tried the whole TDRdelay and level regedit thing. Doesnt make a difference. No software solution makes a difference. I've done it all. I get no crashes anywhere but Jedha.
