Space Engine 0.9.7.1 freezing on load, using ATI drivers

[Nigholith]

Trying to run Space engine 0.9.7.1 through wine 1.7.13 on Fedora 20, using the ATI proprietary 13.35 drivers on an R9 270x GPU; using KDE 4.12.2 because the ATI proprietary drivers don't like GNOME.

The main menu works fine, but totally freezes when trying to start the engine proper. Space Engine has Platinum rating on the appDB, all be it running version 0.96 of SE. I don't get any log output.

I'm new to Wine and Linux generally, so I've not tried it with winetricks or anything yet. Other 3D applications, like Banished, run out of the box on wine without issues; so there's no systemic issue.

Any help would be appreciated. Space Engine is distributed for free on the developer website, so if somebody wouldn't mind testing the latest version and seeing whether the fault reproduces, that'd be grand. Be warned that launching the .desktop shortcut actually froze my whole system, so you might wanna save stuff before testing.

[Bob Wya]

Hi Nigholith,

I'm not familiar with the game you are trying to run... (Well never heard of it before actually) I see the tumbleweed is blowing the AppDB page for it as well...

The important thing with Wine is to read the FAQ. So you get familiar with 32-bit/64-bit Wineprefixes, etc. how these work.

The best way to troubleshoot applications, being run with Wine, is to launch them from the commandline. Make sure your $WINEPREFIX env variable is correctly exported in the shell before running your application. You need to run with the $PWD set to directory where the Windows executable resides. E.g.

Create a 32-bit Wineprefix (still prefered for the "ultimate" compatibility):

Code: Select all

mkdir ~/wineprefix
export WINEARCH=win32
export WINEPREFIX=~/wineprefix/space_engine32
winecfg # directory ~/wineprefix/space_engine32 is created first time WINEPREFIX is accessed - running winecfg will setup the WINEPREFIX environment

Launch installer:

Code: Select all

cd ~/Downloads
wine SE-0971-setup.exe

Launch game:

Code: Select all

cd ~/wineprefix/space_engine32/drive_c/Program\ Files\SpaceEngine 
wine SE.exe &> log.txt
tar -cf log.txt.tar log.txt

You can attach the log.txt.tar file to your post or if it's not too long then post the contents of log.txt inline - with CODE brackets. Why tar it up? Well because the WineHQ forums are bit picky about what file attachments it will allow

You don't want to get all "leet" like most Ubuntu users and install DX11, .net 5.0, etc., etc. with winetricks until you've seen what errors you are getting in the console log. This won't helped you none...

I'll download the game and take a look if it'll run on my laptop. However I don't have Fedora installed (could never get on with it) and the different hardware (older AMD GPU with radeon driver) means the comparison won't be apples-to-apples.

Bob

[Bob Wya]

@Nighollith,

OK I downloaded, installed (in a 32-bit Wineprefix) and ran the game.

Mostly it runs - fullscreen or windowed, sound is OK, some rendering glitching in the game (contents of the picture in the Stellar object info window is garbled for instance). I don't think the planetary bodies are being rendered correctly when seen close up. Not really sure what the game should look like under Windows (and can't really be bothered testing it). It is kinda of neat being able to travel around at 100 Million Parsecs/second though!

The console log is spammed with...

Code: Select all

EE r600_shader.c:157 r600_pipe_shader_create - translation from TGSI failed !
EE r600_state_common.c:745 r600_shader_select - Failed to build shader variant (type=1) -1

But the r600 radeon driver is still under quite heavy development. Thankfully it's nothing like the buggy trainwreck that is FGLRX/Catalyst.

I'm using Wine 1.7.16 on Arch 64-bit btw.

So in summary the game should run (on vanilla Wine) - but whether Catalyst will play ball with it is anyone's guess

Bob

[Nigholith]

Thanks for walking me through prefixes and testing it Bob Wya, I should have really gone through the FAQ before posting.

The log file cites a RtlpWaitForCriticalSection call in wgl.c, and is way above my limited experience grade to interpret; the full, all be it short, log is attached. I'll try some Googlefu tomorrow.

Ugh, the issues I've had with Catalyst, it wouldn't surprise me at all if its some obscure driver fault; sadly it's the only driver that seems to be able to leverage anything close to the full power of the card, the open drivers seem to give me half the performance under ideal conditions. The whole ATI situation on linux seems like a cluster frag at the moment; I naïvely hoped that because AMD was publishing official Linux drivers, it would have near the same stability and support as nVidia provide. Turns out even installing the damned Catalyst drivers ate up half my afternoon, dealing with various incompatibilities and depreciations.