Early performance notes

Hello, I seem to be having a problem with the latest distro.

When I run the Mythruna.sh, everything works up until i get to 19-23% on loading single player, then I get a crash, and the log stops recording before it can tell me what the error is. In Multiplayer it crashes on connect. The funky thing is, there is no problem under Windows.

Graphics Card: Mobile Intel G45

Sound Card: I think it’s a SoundBlaster…

OS: Linux Ubuntu 10.10 Maverick Meercat 64Bit dual booted with Win7 64bit.

Wherever you were running from, it should have create some kind of error log with the exception. Some big number -error.log. Basically in the same directory it created the mythruna.db.



If it didn’t create one of these then it’s a more serious crash and may have happened somewhere in native code that I can’t control.



If it works on windows then it’s possible that it’s the Linux support for your graphics card that’s at play here. :frowning:

OH, I thought those irritating logs were some kinda openjre error, but i just remembered i’m running sun-java

here it is!

http://pastebin.com/J1GpH0uD

Yeah… the JVM crashed hard in native code… somewhere in the OGL driver it looks like. Momoko_fan is much better at deciphering these things than I am…



…but I suspect there’s not much I can do to make Mythruna work in that case. :frowning:

No idea, everything should work fine. It might be useful to know what arguments to glDrawRangeElements made it fail though …

if p speed is correct , that would be EXTREMELY weird, since all the other OGL apps I run (including Minecraft Beta, lol) faster under ubuntu than under windows. I tried to run the windows version on wine, but to no avail. It’s strange that it can load the menu though…

It is weird… especially since as you say, it runs find on windows on the same box, right? Which to me means that the hardware can handle it.



I do some things that Minecraft probably doesn’t do… and maybe those things cause problems for the linux drivers? Don’t know what exactly they’d be, though.

Hello again, I think I may have found out something, I tried to run a test build that had no menu, it ran all the way up to the point that everything on the HUD was rendering, tried to start generation, and crashed right then. Maybe Linux doesn’t like the fractal noise you use? It shouldn’t be a block rendering thing since the preview block renders, but I haven’t seen your code so I don’t know…

If it’s in native code (as indicated by your log) then it’s definitely something to do with the graphics. And the fractal stuff is all Java.



Since the preview block starts out as just dirt… it could be any of the other materials causing issues, water, grass, etc… Unfortunately, it seems like the linux drivers don’t handle it well enough to tell use what the issue is.



If you run from the console do you get any error output prior to the crash? Maybe a shader had issues earlier or something that causes problems later.

Given when this hits… I’m wondering if it is related to Bloom since one or two other users have also had odd issues where bloom crashes. And I don’t initialize bloom until the frame right after the HUD stuff appears. (because if I do it at the same time other bad things happen… so maybe it’s all related.)

Core i5 650 with iGPU, Ubuntu 11.04 64bit (new installation) at 1280x720x24bit windowed: ~ 10-12 fps

for Mythruna versions 20110309 to 20110430

(AntiAliasing is not supported by driver or iGPU, Vsync not enabled)

Note: you can shorten the clip with F12 (well, the first hit will make it bigger so hit it twice). Maybe that will give you some more FPS.

Hi,



Intel core 2 duo P8400 2.26GHz

geforce 9600M GT

4 (3) Gb RAM

win 7 32bit

game build 20110430



1280x720 24bpp windowed (no vsync) - avg 75fps; BUT with 2xaa or better sometimes loading is very slow and it crashes during load



1680x1050 32bpp fullscreen (no vsync) - avg 60fps; works fine even with 2xaa or better

_____________________________________________________________________________________________________

when i hit F5 game immediately crashes



this is log:



Uncaught exception from thread:Thread[LWJGL Renderer Thread,5,main]

java.lang.RuntimeException: Error calling method:createBlueprint

at mythruna.KeyMethodAction.callMethod(KeyMethodAction.java:70)

at mythruna.KeyMethodAction.onAction(KeyMethodAction.java:79)

at com.jme3.input.InputManager.invokeActions(InputManager.java:143)

at com.jme3.input.InputManager.onKeyEventQueued(InputManager.java:384)

at com.jme3.input.InputManager.processQueue(InputManager.java:563)

at com.jme3.input.InputManager.update(InputManager.java:600)

at com.jme3.app.Application.update(Application.java:463)

at com.jme3.app.SimpleApplication.update(SimpleApplication.java:228)

at com.jme3.system.lwjgl.LwjglAbstractDisplay.runLoop(LwjglAbstractDisplay.java:137)

at com.jme3.system.lwjgl.LwjglDisplay.runLoop(LwjglDisplay.java:161)

at com.jme3.system.lwjgl.LwjglAbstractDisplay.run(LwjglAbstractDisplay.java:206)

at java.lang.Thread.run(Unknown Source)

Caused by: java.lang.reflect.InvocationTargetException

at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)

at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)

at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)

at java.lang.reflect.Method.invoke(Unknown Source)

at mythruna.KeyMethodAction.callMethod(KeyMethodAction.java:66)

… 11 more

Caused by: java.lang.NegativeArraySizeException

at mythruna.TestObjectListener.createBlueprint(TestObjectListener.java:94)

at mythruna.GameAppState.createBlueprint(GameAppState.java:859)

… 16 more

pspeed said:
Note: you can shorten the clip with F12 (well, the first hit will make it bigger so hit it twice). Maybe that will give you some more FPS.

Thanks, this increases performance (to ~18-21 fps at 1280x720x24bit windowed mode) with clip set to 64

@pspeed: is there a documentation of key bindings?



F1: toggle postprocessing

F2: save screenshot

F3: toggle HUD (heads-up display) visibility

F4: none

F5: crashes the game, see #33

F6, F7, F8, F9, F10: none

F11: toggle stats/FPS display

F12: change clip (64/96/128/160/192) (clip = rendering distance)



Enter: open console, send message, close console

WASD: walk

Space: jump

Esc: toggle menu

C: select focused block to build it

Arrows: turn camera



Mouse left: remove focused block or (if it is e.g. a chair) move it. max. distance: 11 blocks

Mouse right: build selected block; max. distance: 10 blocks

Mouse scroll: change block selection

Its right below the link you used to download mythruna :wink:

ragan said:
Hi,
Intel core 2 duo P8400 2.26GHz
geforce 9600M GT
4 (3) Gb RAM
win 7 32bit
game build 20110430
1280x720 24bpp windowed (no vsync) - avg 75fps; BUT with 2xaa or better sometimes loading is very slow and it crashes during load
1680x1050 32bpp fullscreen (no vsync) - avg 60fps; works fine even with 2xaa or better
_____________________________________________________________________________________________________
when i hit F5 game immediately crashes
this is log:
Uncaught exception from thread:Thread[LWJGL Renderer Thread,5,main]
java.lang.RuntimeException: Error calling method:createBlueprint
at mythruna.KeyMethodAction.callMethod(KeyMethodAction.java:70)
at mythruna.KeyMethodAction.onAction(KeyMethodAction.java:79)
at com.jme3.input.InputManager.invokeActions(InputManager.java:143)
at com.jme3.input.InputManager.onKeyEventQueued(InputManager.java:384)
at com.jme3.input.InputManager.processQueue(InputManager.java:563)
at com.jme3.input.InputManager.update(InputManager.java:600)
at com.jme3.app.Application.update(Application.java:463)
at com.jme3.app.SimpleApplication.update(SimpleApplication.java:228)
at com.jme3.system.lwjgl.LwjglAbstractDisplay.runLoop(LwjglAbstractDisplay.java:137)
at com.jme3.system.lwjgl.LwjglDisplay.runLoop(LwjglDisplay.java:161)
at com.jme3.system.lwjgl.LwjglAbstractDisplay.run(LwjglAbstractDisplay.java:206)
at java.lang.Thread.run(Unknown Source)
Caused by: java.lang.reflect.InvocationTargetException
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)
at java.lang.reflect.Method.invoke(Unknown Source)
at mythruna.KeyMethodAction.callMethod(KeyMethodAction.java:66)
... 11 more
Caused by: java.lang.NegativeArraySizeException
at mythruna.TestObjectListener.createBlueprint(TestObjectListener.java:94)
at mythruna.GameAppState.createBlueprint(GameAppState.java:859)
... 16 more


Edit: misread the original post...

Is that when hitting f5 in multiplayer?
ragan said:
1280x720 24bpp windowed (no vsync) - avg 75fps; BUT with 2xaa or better sometimes loading is very slow and it crashes during load


Regarding this, without a crash log I can't be sure but given the circumstances I'm wondering if bloom and/or the post-processing filters are involved.

Starting them in update had some weird problems for me and I added a work around but I'm wondering if certain cases make it still fail.
normen said:
Its right below the link you used to download mythruna ;)

too easy, too logical to see :( - sorry.... *confused*
anouser said:
too easy, too logical to see :( - sorry.... *confused*

I am familiar with that problem, you wont believe how many people ask for documentation for jME3 using a page that contains the link to it right at the top ;)