After the hardware upgrade, the track crashes every time after loading.

Discussion in 'Bug Reports' started by ddooewq, Jun 19, 2022.

Tags:
  1. DJCruicky

    DJCruicky Registered

    Joined:
    Jan 25, 2012
    Messages:
    1,592
    Likes Received:
    1,121
    The hang at 79% is normal on a track you are loading for the first time. It's generating a collision file found in ...\UserData\Log\CBash.
    If you getting a pause EVERY time then that is NOT normal.
    After that the cars are loaded into memory.
     
  2. Stefan Donath

    Stefan Donath Registered

    Joined:
    Jul 7, 2021
    Messages:
    50
    Likes Received:
    23
    Within singleplayer i got this every time and the function is given every time.
    Within multiplayer i got it only after track is reinstalled i think. And within multiplayer, the function is only given the first time after track reinstallation. I'm not sure, but when i'm not totally wrong, there was no cbash file generated. Because firstly when this ctd occours i deleted cbash file and shader cache when rF2 crashed to desktop because sometimes it helps when those files are deleted. The next time i would delete the file and it wasn't there.
    But i have to test this again.
    Maybe there is really something with cbash generation/reading/saving.
    The strange thing is the fact that singleplayer seems to generate cbash every time but multiplayer don't do so (i think only once).
    In singleplayer i can load to track. In multiplayer only once after reinstallation. Even when session jumps to next track, rF2 crashed to desktop. And this after i drove for about 2.5h on the server before it goes to next track.
    At the moment my workaround is functional for me, costs only a little time but this is okay for me.
    But maybe they can find something.
     
    DJCruicky likes this.
  3. Alex Sawczuk

    Alex Sawczuk Administrator Staff Member

    Joined:
    Oct 4, 2010
    Messages:
    851
    Likes Received:
    971
    The 79% pause is a bug in RC, it's been fixed and will be in the next update.

    Thanks for the dmp file.
     
  4. Alex Sawczuk

    Alex Sawczuk Administrator Staff Member

    Joined:
    Oct 4, 2010
    Messages:
    851
    Likes Received:
    971
    If you set PostFX to off does it still crash?
     
  5. Stefan Donath

    Stefan Donath Registered

    Joined:
    Jul 7, 2021
    Messages:
    50
    Likes Received:
    23
    Still crashing in multiplayer when postfx is set to off
    Singleplayer is fine, no crash.
    Tested with PostFx Ultra/Medium/Off
     
  6. Coutie

    Coutie Moderator Staff Member

    Joined:
    Oct 5, 2010
    Messages:
    3,779
    Likes Received:
    2,241
    Seems like it's an issue with AMD drivers. Which version are you running?
     
  7. DJCruicky

    DJCruicky Registered

    Joined:
    Jan 25, 2012
    Messages:
    1,592
    Likes Received:
    1,121
    See post 58.
     
  8. Stefan Donath

    Stefan Donath Registered

    Joined:
    Jul 7, 2021
    Messages:
    50
    Likes Received:
    23
    So i've updated to 22.7.1 and deleted Cbash and Shadercache and at the moment it don't ctd while join to server.
    I've connected several times to server and also restarted rF2 and join to server.
    I will inform you if this behavior occours again.
     
    SmellySkidmark and davehenrie like this.
  9. Stefan Donath

    Stefan Donath Registered

    Joined:
    Jul 7, 2021
    Messages:
    50
    Likes Received:
    23
    So i've tested a little more. 22.7.1 doesn't solve the issue.

    I can join the server which i joined first directly after deleting cbash and shadercache in both modes (drive/spectate) even after a new start of rF2.
    When i join another server in drive mode get a ctd.
    I can join those other server in spectate mode without ctd. In this case the preview window is black (maybe this is related due to an empty server).
    Then i delete the shadercache once and can join another server. But the first server which is functional before gives me a ctd.
    Then deleting shadercache once again and i can rejoin the first server (some changes on dynamic.cache are done).
    And then after restart of rF2 i can also join second server (also some changes on dynamic.cache are done).
    Then tested a third server within the same rF2 session (no restart) and get a ctd (no changes on dynamic.cache (same track as first server)).
    Restarted rF2, tested 3rd server and get another ctd (no changes on dynamic.cache (same track as first server)).
    Within Singleplayer i can join all tracks [edit server -> tracks] without problem.

    This behavior is really strange.
     
    Last edited: Jul 27, 2022
  10. davehenrie

    davehenrie Registered

    Joined:
    Jul 6, 2016
    Messages:
    7,454
    Likes Received:
    4,369
    There are some new features in the AMD drivers. I wonder if trying a more vanilla group of settings may help or hinder this investigation?
    First thing to turn off, would be the noise reduction, then items like Advanced Sync or Video Upscaling.
     
  11. Stefan Donath

    Stefan Donath Registered

    Joined:
    Jul 7, 2021
    Messages:
    50
    Likes Received:
    23
    Noise reduction is firstly introduced with 22.7.1 i think. So maybe there is a problem with 22.7.1 but not with the older versions.
    I've read something about AMD optimized shader cache. Maybe i deactivate this and give it a try.
    But its already late here. I will check some things tomorrow. Maybe we can find a pattern.
     
  12. ddooewq

    ddooewq Registered

    Joined:
    Jun 17, 2022
    Messages:
    33
    Likes Received:
    2
    Yes, I found it, I tried many ways, but I didn't try to install the old version of the graphics card driver.Until I saw the upstairs that there may be a problem with the AMD graphics card driver. . .o_Oo_Oo_Oo_O
    As I wrote upstairs, the driver can be run until the 22.5 version, but it will not work if you upgrade it upwards.
    Thank you very much for your help! !
     
  13. Stefan Donath

    Stefan Donath Registered

    Joined:
    Jul 7, 2021
    Messages:
    50
    Likes Received:
    23
    Okay, i've tested a lot more things.
    Unfortunately the tests didn't show consistency in every case.

    At first, the hotfix update of the release candidate fixed the functionality within single player. So now single player and multiplayer show the same behavior. So i can test now within single player, so the possibilities to test different things are wider than before.

    I've monitored all rrshd and cbash files and the dynamic.cache file during those tests to check if changes have a impact on the behavior.

    1. Graphics settings: set all graphics settings to the lowest possible value to check if this has any impact on the behavior. -> no change
    2. Fullscreen mode: changed mode from borderless window to fullscreen. -> this brings a big step forward (now it's possible to load any tracks and even one track multiple times in a row, but this depends on a special precondition), i will give more information later to this point.
    3. Borderless Window: changed mode back to borderless and i can join different tracks. On sebring i can join every layout in a row (here it seems there is no rrshd and cbash file generated). On other tracks like spa, i can join one layout and when change the layout or load the same layout the second time i got a ctd. After a restart of rF2 i can sometimes join a track once (not every start), but no later than the second try i got a ctd.
    4. Fullscreen Mode within session started in borderless window mode: after a successfully loading to track in borderless mode, i go back to main menu and changed mode to fullscreen mode. Than i attempt to go on the track which was previously (within same session) loaded within borderless mode and got a ctd.
    5. Fullscreen mode (started new rF2 session in fullscreen mode): i can load every track i tested and even the same track multiple times in a row.

    while testing i checked the change date of the mentioned shader files (rrshd, cbash, dynamic.cache) but it seems the last change time of those times has no correlation to the "functionality" of successfully joining a track.

    The only files that changed were those within "rFactor 2\UserData\player\LocalStorage folder". Unfortunately i couldn't delete files within this folder while rF2 is running because this are the temp files and under exlusive access by rF2 while running.
    Deleting the files while rF2 is closed aren't giving me additional information due to the fact that mostly the first start of a track is performed successfully after rF2 is started.

    So it seems that there is a problem with caching while borderless window mode is active. Maybe it's something in relation with the AMD adrenalin 22.5.2 (22.10 preview may) introduced dx11 performance fix. But i don't know if this will be fixed within the driver sometime in the future or if there is a fix possible within rF2 code.
    I've researched a lot if driver higher than 22.5.1 are showing problems within other games/apps and there are multiple things.
    Maybe S397 as developer could contact AMD to this problem. I think the chance that this behavior is checked and possibly solved by AMD is higher than when a user contact them.
    Maybe this behavior would be "inadvertently" solved by solving an other issue.
     
  14. nolive721

    nolive721 Registered

    Joined:
    Jan 2, 2016
    Messages:
    379
    Likes Received:
    188
    yes, problems for me with AMD drivers higher than 22.5.1. tracks like Laguna Seca, Hockenheim would load to 100% and kicks me then to Windows desktop
    so I am also thinking its related to this DX11 improvement they brought after that May release, RF2 being a DX11 title might not work well with these optimization that AMD implemented
     
  15. davehenrie

    davehenrie Registered

    Joined:
    Jul 6, 2016
    Messages:
    7,454
    Likes Received:
    4,369
    "Welcome to AMD troubleshooting, how may we help you?"
    uh, hi, I'm the guy who replaced Marcel, do you know him? No? ok, anyhow, we seem to have an issue with your new drivers, they are causing some of our cars & tracks to crash.
    "Oh that is indeed a problem...how many users are we talking about?"
    oh about 700 to 800 per day on Steam....
    "700?? Are all of these AMD users?"
    uh, no...probably only about 10% or less
    "I see....OH LOOK AT THE TIME!!! Gotta run, have a nice day."
    but.... but...
     
  16. Stefan Donath

    Stefan Donath Registered

    Joined:
    Jul 7, 2021
    Messages:
    50
    Likes Received:
    23
    Don't know what's ur problem. Why do i say they should contact the support.. because i know nothing about the engine and the used code fragments and maybe they can enclose the problem with the usage of memory dump and log files.
    Even if there are only some users using a amd card and having such a problem (i myself know some personally and give them help with the "workaround") it's worth a try. And at the moment amd is on the way back winning market share. Possibly they can win the next generation, even if not at the pure performance the chance is there that they win the performance/watt race.
    Additional to this rF2 is on a good way to possibly get more users with those new additions to physics, online system, content, ...
    So possibly over the time the part buying rF2 and running/buying a new amd card will grow. And the first thing the owner of a new card will do, is installing the newest driver (and maybe they have no other chance if the new generation needs at least a driver version where this big is showing and can't go back to an older version because the card than will run no more).
    Those possible new rF2 user will get ctds and maybe the will search for help within the forum. On the other side some of them would give back rF2 within steam and will downvote the sim within steam/internet for crashing all the time.
    While watching amd forum and reddit offical amd part than amd is listening to bug reports. So it's worth a try.
    I for myself check every bug report getting by a user even if this is only one out of million usecases because it can hint to a problem that could get a big one impacting much more programs over the time.
    I invest time for checking multiple things, try to analyze things and than there are users talking everything down. Yeah, maybe no one would analyse this on developer side who knows. But possibly users will decide to go with another sim because there are no crashes and they couldn't see the relation between driver versions and crashes.
    So at the moment the problem here is not the developer side, no the problem are some user talking down everything.
     
  17. davehenrie

    davehenrie Registered

    Joined:
    Jul 6, 2016
    Messages:
    7,454
    Likes Received:
    4,369
    yes it is worth a try, but rF2 is a niche title in a very niche market. That is all I'm saying. Sim Racing as a whole is a zit on an Elephant. I was merely trying to manage expectations. There has been a lot of user research concerning drivers here and that is fantastic. But looking at Stream charts, the number one game has half a million users on at 5:30am on the West Coast of the USA. rF2 had 379. Not 379k, but 379. According to Steam charts, rFactor2 is ranked 682nd most played game on Steam. Is a project manager of the driver development team at Nividia or AMD gonna open a file on a title that doesn't even reach 1% of the user base of the most popular game?
    I use rF2 daily, I have an AMD video card as I think it does not suffer the stutters that seem to plague the markedly faster Nvidia products which dominate the user base in rF2. I'm under no illusion that the small fraction of AMD users of an incredibly small rF2 user base will get noticed by anyone who has to control budgets and allocate personal in companies like AMD & Nvidia. Sorry for being a grouch. I'll try and do better.
     
  18. Stefan Donath

    Stefan Donath Registered

    Joined:
    Jul 7, 2021
    Messages:
    50
    Likes Received:
    23
    I'm with you that those people with problems using AMD cards within rF2 is really small. But maybe this problem is one connected to other problems within other games. In my case it seems that GTA 5 is also affected by those new drivers (getting a d3d error even when checked all game files. But not tested with 22.7.1). Within world of warships there are also problems were developer contacted amd already. Tomb raider 2013 als ctd with 22.5.2. And there are a lot more out there i think.
    Maybe these problems aren't related to eachother maybe they are. I think when each developer/user create a bug report, such a problem rises in the todo list.
    You say that rF2 is 682nd most played game (maybe at the specific time, maybe alltime but this isn't important) so compared to the amount of games available on steam (8528) this is within top 8%, so not bad. And realistically i thought there are more games in front of rF2.
    Even if there are some problems with actual drivers (22.5.2 and above) i play a lot DX11 games without any error.
    So maybe there are specific instructions which cause such issues.
    My 6800XT runs cool and quiet (with uv) and i think it's a really good card and with the latest optimizations (dx11, opengl) and features (noise cancelling, FSR 1.0 and FSR 2.0) it get even better.
    So hopefully AMD fix the problem.
     
    davehenrie and nolive721 like this.
  19. davehenrie

    davehenrie Registered

    Joined:
    Jul 6, 2016
    Messages:
    7,454
    Likes Received:
    4,369
    I will admit my error in NOT considering other games also having problems. Too focus'd on rF2 I guess. Good points.
     
    ddooewq and Stefan Donath like this.
  20. ddooewq

    ddooewq Registered

    Joined:
    Jun 17, 2022
    Messages:
    33
    Likes Received:
    2
    I'm confused right now, should I upgrade the graphics card driver, because there are a lot of new things added to improve other games. . Although I would like to experience this. But I love rf2, so I haven't updated the driver. This confuses me. . .:rolleyes::rolleyes::rolleyes::rolleyes:
     

Share This Page