POE 2 Complete PC Freeze while loading screen
|
After countless hours and money spent on trying to figure out why my game was crashing I finally determined the root cause.
For those of you with an i9 14900k go to your bios settings and sync all pcore ratios to 55. This completely fixed the game for me and I'm even able to use directx12 with everything on high settings And the newest version of windows 24h2 I hope this helps someone Good luck out there exiles Zuletzt bearbeitet von glitche#2433 um 18.02.2025, 23:57:27
|
|
|
Even after the new patch. still hard crashing. this sucks.
|
|
|
Reminder to use the "--ex-crash-report" to send crash report data to GGG whenever you crash.
(freeze or crash, both works) For Steam: Goto Library->Path of Exile 2 properties by right-clicking the icon->General->Launch Option-> Add --ex-crash-report to the box.  For Standalone client: Right the desktop shortcut->Properties->In the target field add --ex-crash-report  Zuletzt bearbeitet von Climet#7103 um 18.02.2025, 06:37:18
|
|
" Something tells me the two of you would have solved this issue in 1 day if you where given permissions to the actual code. Would be so sad but hilarious if the actual issue is because someone commented out code somewhere in the engine that no-one dares to touch anymore. Or if it were their own doing even: like "This piece of code don't seem to do anything .. load-times will likely be a bit faster if we just remove it" We will likely never know, but lets hope that the info might help them solve the issue. Im actually rooting for them. I believe no fans of the genre actually wants to see GGG fail. This old crappy engine they are using is really staring to become a huge problem though. Zuletzt bearbeitet von snurrfint#5780 um 18.02.2025, 09:03:25
|
|
|
Because I have some time I tried to investigate the hang some more. But then I saw that there was an update and I tried to generate the hang "burning" some low tier maps. Entering the map and then go back to hideout and after like 10+ maps I had 0 crashes :). Idk what happened and idk if my last analysis helped GGG but at first glance it seems to be fixed
// LE: Nvm. Just got a crash Zuletzt bearbeitet von IceCool10#6669 um 18.02.2025, 13:21:56
|
|
|
can't play since I "upgraded" to windows 11...
online up to 5-10mins. hard crash, freeze... this is rubbish! fix it please. DxDiag has been sent. |
|
|
I gave up and uninstalled the game.
Good luck. |
|
|
Just uninstall the game at this point and move on with your life, it will not change any time soon or for the better. This trend has persisted for a while by GGG.
Call them out on every single social media platform, don't let them get away with this. Like so may others I paid to play this game and I can not even do that. There will be a fix in a few years... maybe... sigh Someone should start a class action lawsuit to refund the early access purchase price. They delivered not even half of what they promised in EA and the game is actually unplayable. Zuletzt bearbeitet von Braric#4417 um 18.02.2025, 18:28:06
|
|
|
After some hours I've managed to trigger another crash again. And I kinda saw something. I can't say that I have some good info but who knows, maybe it helps. Also, my debugger bugged and it prints the same thing twice ( sorry about that ) but I didn't have that much time to make the output look nice and I didn't want to try to restart the debugger and try to generate another crash.
So I was printing the stack trace from all threads when the exception occurs: " That is just a snippet but the thing is : Thread 15 has the same lpStartAddress as the others ( bold text ) and yes, that is a return address not the start address of the thread but based on the assembly code the start address is the same. But for some reason thread 15 generates an exception. " And this is the code that creates those threads: " And those other threads that don't generate an exception are waiting and here is the code: " And this code is called from here ( from sub_140C55D10 ): " Then I searched for a call to this function (sub_140C55D10) to try to understand what might cause the exception but there are no calls. At least not in the main executable. All I could found was an array that contains this function: " Then I've searched for that array in the code: " Ok, so this looks like a function that prepares an EXCEPTION_RECORD struct or something like that. Now this code seems to be called only if cmp rbx, rsi is NOT true ( RBX != RSI ) and RBX and RSI are two QWORDs that are sequentially in memory RBX = *RSI and then RSI = *(RSI+8). Now I was trying to understand what are those and when RBX != RSI but unfortunately I don't have too much time now. All I've seen was a .text:0000000140C53EFC 48 8D B7 F8 00 00 00 lea rsi, [rdi+0F8h] and .text:0000000140C53CF8 49 8B F0 mov rsi, r8 So RSI comes initially from RDI+0xF8 or from the third parameter ?? ( although it's a variadic function ) ( probably first case -> RDI+0xF8 ). And while I am writing this, Im pretty sure it's from [rdi+0xf8]. Seeing that 0xf8 I was like "WAIT A MINUTE!". I know that value. It's the offset of RIP from the _CONTEXT structure :). " And every thread has such a structure. So I am pretty sure rdi holds the _CONTEXT structure. Idk how much this helps ( I hope it does ) but I guess at least I can come back to it when I have some more time. |
|
|
I used to crashed to my desktop every 5m. After I lower my pcore to 46 I haven't crashed since. Thank you!
|
|
