Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Cosmitz]

Check last edit in previous post.

Rebooted, opposite fullscreen, sound off, and i am running Geforce 378.78.

No matter what i do, it crashes within 10-ish second of loading, and this is about the same timeframe as it would have happened when i got back from the battle when the worldsim resumed (when it crashed the first time). It doesn't crash if i don't unpause.

[Alex]

Thank you for trying this stuff. Really not much to go on here, unfortunately.

Oh - another thought occurs, can you try running the game using starsector.bat rather than the exe?

Isn't 7.0.800.15 32bit?

Same version numbers for 32/64 bit versions, pretty sure.

[Cosmitz]

Ok. Nailed it to something.

If i push shift, timespeed up, after i load, it doesn't freeze. (tested 3 times, both on and off) And come to think of it, the other occurence was related to me speeding up time. I mean, as an old player it's second nature that i just half my finger on shift almost always.

But upon loading the game after it became 'unstuck', to the 'stuck' save event, it almost instantly freezes upon loading.

[Alex]

Just double-checked to make sure I wasn't pressing shift without thinking about it - still no crash.

[oreganor]

The change to x64 JRE 7u80 does not prevent the crash in my case. I tried disabling sound (as the last lines of the log refer to some updates on the sound threads), CTD still happening.

I observed a few things (My card is an ATI HD7700):

- Windows Event Log assings the error purely to java engine stoping responding in all cases.

- The screen flicker happens in a special way (Game screen freezes with the translucent ship detail window open for a few moments, then it goes black, after a moment the desktop is briefly visible and then the screen goes black again, this time with the windows problem dialog appearing) that reminds me of when some games do not like ATI drivers and the driver recovers itself.

- If I have an open browser window in the background (I use chrome) the browser window stops to refresh itself after the CTD happens on Starsector (I have to close the browser window because nothing I do makes it redraw itself). Again, I have seen this behaviour on ATI drivers after a faulty application makes them crash. But on all this cases (I mean the possitive ones that are fixed at some point by a patch on the application/game itself) on the Event Log the one that stopped responding was the ATI driver itself.

I'm going to play with compatibility modes/graphical options to see if anything has an impact on the crash. Are there any extra vmparams that affect graphic display/process that I could test?

[oreganor]

Compatibility modes doesn't have an impact on CTD (as expected they make the error recovery behave in different ways... But I think that's not related to the main problem).

Meanwhile fullscreen option HAS AN IMPACT:

- Fullscreen mode, resolution to the max (1680x1050) -> CTD as reported

- NOT Fullscreen, resolution to the max (1680x1050) -> CTD as reported (Correct me if I'm wrong but SS behaviour looks like fullscreen if you select max resolution reported by the monitor, doesn't matter if I choose fullscreen or not?)

- NOT Fullscreen, resolution down to 1440x900 (So the game window is visible with a portion of the desktop visible around the borders) -> NOT CTD after at least 100 ship details windows opened.

- Fullscreen, resolution down to 1440x900 -> NOT CTD after at least 100 ship details windows opened.


It's not the 1st time I see a game behaving like this were fullscreen + max resolution = CTDs (Specially on some versions of Unity Engine), that later is corrected by a patch on the game itself.


EDIT: I haven't mentioned it specifically, but the rest of games, window itself, dxdiag, etc, etc are working without problems and not reporting any issues at all at max resolution.

[Cosmitz]

Ok, went full rambo on this. The issue was consistent but only on one install on my machine. My save on another install. Fine. Moved my install to another machine as is, fine. Moved the install to another local disk. Fine. Did a chkdsk for 8 hours on the original drive. While it came out 'clean' "Windows has made corrections to the file system."

Reran original install in original location.

Solved. For once, the issue /was/ on my end.

---


Oreganor, try running a chkdsk /f on the drive SS is on. If that fixes it for you, we can call this closed.

[oreganor]

I will do it, but I think the issues are different... Mine is strictly related to an UI action, which is opening the ship detail window... And the change on resolution fixing it, starts to smell to something related to vsync...

...Yours was timed driven as it did happen at a given moment with independence of what you were doing. I can play for hours for as long as I don't abuse ship details too much (in 1680x1050) and nothing bad happens for as long as I don't raise from 1440x900...

...Definitively mine looks like one of those horrible glitches hard to reproduce linked most likely to a particular combo of card/OS/related hardware. Will try a chkdsk /f anyways, that will not harm in either case , but unless some1 else with a similar hardware helps, I think it's just bad luck on my side and the Dev has enough on his dish to actually waste time in fixing a bug that (so far) can be avoided (1440x900 is not preatty... But at least I can play without crossing fingers each time I check a ship's stats before buying it ) and affects a single person (or at least I haven't seen other reports like mine so far).

[oreganor]

As I was suspecting has to do with vsync. Luckily I could create a performance profile in my Radeon Software linked to the local java.exe that SS uses, to do some forced vsync tests:

- At fullscreen 1680x1050 with vsync forcefully active in my Radeon profile for this game. More than 100 ship detail windows without problems.

- At fullscreen 1680x1050 with vsync forcefully inactive in my Radeon profile for this game. It crashed at the 3rd window opened. This time, as I got the Radeon profiler on, on top of SS crashing, I got also a proper driver soft failure reported by windows.

So at least in my system I need vsync on to prevent this CTD from happening. Hope that helps.

[Alex]

@Cosmitz: whew, glad you figured it out. Hadn't thought about chkdsk, but in retrospect that makes sense - will be good to know to suggest in the future, in case someone has a similar issue.

So at least in my system I need vsync on to prevent this CTD from happening. Hope that helps.

Glad you figured it out as well Will help in case someone with AMD has a similar issue, certainly - what's weird is that the game has vsync turned on by default, so forcing it on shouldn't have any effect. (This is almost definitely a driver bug; could possibly be worked around by making the game do something different, but with these things it's hard to do that w/o being able to reproduce in the dev environment.)

(Correct me if I'm wrong but SS behaviour looks like fullscreen if you select max resolution reported by the monitor, doesn't matter if I choose fullscreen or not?)

What it does in this case is create an "undecorated" window - a window without the caption bar or borders, which would make a full-resolution window not fit on the screen. It's different from fullscreen in that it can be switched to and from more quickly, though I believe it's also handled by the OS/graphics drivers differently than a regular window. Kind of an in-between case.

[oreganor]

Radeon default preset for Vsync is a bit confusing because it comes in 4 states:

- Always on

- Always off

- Always on unless application says off.

- Always off unless application says on <--- Default State

If I have to bet... My AMD driver is not understanding the setting you are using and defaulting to vsync off. In fact having 4 options here is totally pointless, with off, on and application controlled (like NVIDIA defines Vsync states) is more than enough but...

...AMD drivers are... well, we all know how they are .

[Alex]

Interesting! I wonder if that means the game is running with vsync off on a substantial number of AMD cards. Ah, well.

...AMD drivers are... well, we all know how they are .

Hah, yes

[DIEHARD25]

Hello all

Have same (or similar) issue. Game suddenly freezes, Java became unresponded.

My specs:
Win 7 x64
FX 8350
MSI GTX 1060
8 Gb RAM
SSD 60 Gb (game installed on SSD)

Game version - 0.8a (RC19)

Video drivers are up-to-date (from stable branch).

Already try to forced VSynC On - no effect. Also, checkdisk dont give me any effect.
Already try with different JRE versions - both 7.79 and 7.8, also try out 8 (with disabling warnings in settings.json)
Already try launching wia .bat, with and without sound, different fullscreen combinations (on and off) and different resoultions (from 1920x1080 downgrade 4 or 5 others res).
Log don't have any suspicious info - seems to be, problem somwere in native Java, not in game files.

Almost forgot - vmparams modified for use 2 to 4 Gb of RAM.
Already try both clean and modded versions (SS+)

Also, if this important - freeze become on same in-game date (13-th Dec 206).

upd. i will try to clearly re-install the game and try out videodrivers from Beta branch today at evening (in 8 hours at my local) - will post results after

upd2. sorry for bad English

[Schwartz]

You could try this little tool for a stack trace when SS hangs itself. Note that you need to run it *while* the hang occurs, so you may be a bit pressed for time. For 64-bit Java you need the 64-bit executable.

http://www.nirsoft.net/utils/what_is_hang.html

[DIEHARD25]

yeah, thanks for idea

will try it at evening today.