#jogamp @ irc.freenode.net - 20140918 05:05:43 (UTC)


20140918 05:05:43 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20140917050543.html
20140918 05:05:43 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20140918050543.html
20140918 05:06:26 * [Mike] (~Mike]@anon) Quit ()
20140918 06:37:02 * monsieur_max (~maxime@anon) has joined #jogamp
20140918 06:58:29 * monsieur_max (~maxime@anon) Quit (Ping timeout: 255 seconds)
20140918 07:01:02 * monsieur_max (~maxime@anon) has joined #jogamp
20140918 07:05:20 * monsieur_max (~maxime@anon) Quit (Ping timeout: 258 seconds)
20140918 07:20:57 * jvanek (jvanek@anon) has joined #jogamp
20140918 07:21:25 * monsieur_max (~maxime@anon) has joined #jogamp
20140918 07:47:44 * monsieur_max (~maxime@anon) Quit (Ping timeout: 260 seconds)
20140918 07:48:57 * monsieur_max (~maxime@anon) has joined #jogamp
20140918 07:51:00 <monsieur_max> sorry for the connect/disconnect, got pretty bad rain here
20140918 07:52:54 * rmk0-alt_ (~rmk0@anon) Quit (Ping timeout: 246 seconds)
20140918 07:53:12 * rmk0-alt (~rmk0@anon) has joined #jogamp
20140918 10:03:51 * hija (~hija@anon) has joined #jogamp
20140918 11:21:35 * zzuegg is now known as zzuegg|afk
20140918 12:09:00 * hija (~hija@anon) Quit (Quit: hija)
20140918 12:09:20 * hija (~hija@anon) has joined #jogamp
20140918 12:24:13 * hija (~hija@anon) Quit (Quit: hija)
20140918 12:27:18 * hija (~hija@anon) has joined #jogamp
20140918 14:36:29 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20140918 14:51:33 * [Mike] (~Mike]@anon) has joined #jogamp
20140918 15:41:31 <rmk0> quiet in here
20140918 16:00:19 <sgothel> Hi Mark :)
20140918 16:00:47 <rmk0> lo
20140918 16:00:53 <sgothel> hot fall .. and just digging through bugs
20140918 16:01:17 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=1068 <- thought about this optimization .. will try and see
20140918 16:02:09 <rmk0> hm, right
20140918 16:02:34 <rmk0> those are bits of opengl i've never had to go anywhere near...
20140918 16:02:39 <sgothel> then I prepared the graph/font bugs .. will see
20140918 16:02:41 <rmk0> some other library was always handling it for me
20140918 16:03:20 <sgothel> hehe
20140918 16:03:48 * rmk0 buries head in sand
20140918 16:04:11 <sgothel> na, you will be one of the valuable testers :)
20140918 16:05:52 <rmk0> was hoping to grab harvey
20140918 16:05:58 <rmk0> not seen him for ages
20140918 16:06:49 <sgothel> hope just vacation .. not me having pissed them all off in Vancouver .. me never can shut up :)
20140918 16:06:56 <rmk0> hehe
20140918 16:07:55 <sgothel> Usually Qun soothes the situation, my best part .. but now, it was only my loudness :)
20140918 16:08:24 <sgothel> was also waiting for Xerxes .. hmm
20140918 16:08:46 <sgothel> Oh .. Harvey's boss gave the green light to merge parts of JaamSim (dual license)
20140918 16:08:55 <sgothel> i.e. math stuff .. will see
20140918 16:09:17 <sgothel> Anything I can help with ?
20140918 16:09:34 <rmk0> was hoping to see what they'd done with the blender "API"
20140918 16:10:13 <sgothel> what I can remember was they used the importer from that C* open format .. duh, forgot
20140918 16:10:39 <rmk0> he said something about writing a python plugin
20140918 16:10:44 <sgothel> and blender API .. import, he said, means to traverse through 'it' .. using their code
20140918 16:10:55 <sgothel> right
20140918 16:10:57 <rmk0> my main complaint was that what they call an API is just a load of internal data structures exposed via python
20140918 16:11:07 <rmk0> so you get to rewrite your code every couple of months when something changes
20140918 16:11:07 <sgothel> yes .. makes sense
20140918 16:11:12 <sgothel> yup
20140918 16:11:16 <rmk0> but... he said they've avoided that somehow
20140918 16:11:28 <sgothel> .. letting that python thing doing it :)
20140918 16:11:40 <sgothel> or using blender -> *other format I forgot*
20140918 16:12:05 <sgothel> that hopeful format .. open source .. etc .. tsts senior moments
20140918 16:12:20 <rmk0> the only open format blender exports to is collada
20140918 16:12:25 <sgothel> which now supports curves as well .. yup :)
20140918 16:12:30 <rmk0> and they basically miss most of the data when exporting
20140918 16:12:46 <sgothel> yeah, blender for graph - the final vision ..
20140918 16:24:31 * noxuma (~markt@anon) has joined #jogamp
20140918 16:25:05 <noxuma> Hey guys, I seem to crash horribly when running with NEWT on an AMD card? Has this issue been addressed in the last year or so?
20140918 16:25:41 <noxuma> EXCEPTION_ACCESS_VIOLATION (0xc0000005)
20140918 16:25:42 <noxuma> j jogamp.nativewindow.windows.GDI.SwapBuffers(J)Z+0
20140918 16:25:42 <noxuma> j jogamp.opengl.windows.wgl.WGLUtil.SwapBuffers(J)Z+12
20140918 16:25:42 <noxuma> j jogamp.opengl.windows.wgl.WindowsWGLDrawable.swapBuffersImpl(Z)V+23
20140918 16:25:42 <noxuma> j jogamp.opengl.GLDrawableImpl.swapBuffers()V+67
20140918 16:25:42 <noxuma> j jogamp.opengl.GLAutoDrawableBase.defaultSwapBuffers()V+23
20140918 16:25:42 <noxuma> j com.jogamp.newt.opengl.GLWindow.swapBuffers()V+1
20140918 16:26:13 <sgothel> pls check: our bugreports, latest JOGL and latest AMD-Windows driver
20140918 16:26:33 <sgothel> not familiar w/ this crash ..
20140918 16:27:23 <noxuma> It seems to happen after a few seconds of running, nothing is changing on my screen
20140918 16:27:35 <noxuma> the most similar bug i can see is
20140918 16:27:35 <noxuma> http://forum.jogamp.org/Native-lib-crash-td4025369.html
20140918 16:27:45 <noxuma> but that was with some NV driver
20140918 16:28:07 <noxuma> C [atio6axx.dll+0xa5f53]
20140918 16:28:07 <noxuma> C [OpenGL32.dll+0x3abee]
20140918 16:28:07 <noxuma> C [GDI32.dll+0x39bd3]
20140918 16:28:07 <noxuma> C [nativewindow_win32.dll+0x256d]
20140918 16:28:43 <sgothel> "Has this issue been addressed in the last year or so?" - some AMD driver issues has been addressed, check bugzilla
20140918 16:29:03 <sgothel> you must test w/ latest JOGL yourself .. and if that doesn't help, retest also w/ latest AMD Windows driver
20140918 16:29:10 <sgothel> if this then still fails, pls file a bug report
20140918 16:30:09 <sgothel> pls use one of our unit tests to reproduce, which you then can reference ..
20140918 16:30:30 <noxuma> Ok thanks for ur time
20140918 16:30:52 <sgothel> note: all our unit tests pass on a windows+AMD machine ..
20140918 16:32:03 <noxuma> must be something we're doing here. all our amd machines crash
20140918 16:32:10 <noxuma> nvidia and intel work fine
20140918 16:33:20 <sgothel> 4.2.12430 Compatibility Profile Context 13.152.1.8000 <- amd driver
20140918 16:33:53 <sgothel> hence me stating to use our unit tests, and if that works .. cont. producing a good unit test exposing the issue
20140918 16:36:05 <noxuma> i have 14.20.1004 <- amd driver
20140918 16:36:13 <noxuma> should i go back to an older driver?
20140918 16:36:26 <sgothel> jogl version ?
20140918 16:36:35 <noxuma> I know that in the version of jogl we embed the newer nvidia drivers were broken
20140918 16:36:48 <sgothel> does our def. unit test work ?
20140918 16:36:56 <sgothel> i.e. wiki/FAQ/Bbugreport ..
20140918 16:37:04 <sgothel> then try others ..
20140918 16:37:16 <sgothel> if our tests fail .. (i.e. not your fault) pls file bugreport
20140918 16:37:16 <noxuma> somewhere around 2.1.5 i think
20140918 16:37:25 <sgothel> always test w/ latest!
20140918 16:38:06 <noxuma> I know, i would love to use the latest, but I am on embedded hardware, that isn't always an option
20140918 16:39:08 <sgothel> I am talking about triaging the issue here, Sir - not your final product
20140918 16:39:17 <sgothel> i.e. has it been already fixed ?
20140918 16:39:25 <noxuma> Yup, going to try that
20140918 16:40:24 <sgothel> then you can either backport [yourself] .. or .. maybe hire one of us for consulting (special support)
20140918 16:40:46 <sgothel> if issue exist in latest version .. our fault, we will fix it, no problem
20140918 16:50:28 <noxuma> Well its not crashing! =D But it is running at 2 fps instead of 60 now
20140918 16:50:57 <noxuma> Thanks for the help, I think its something on my end now
20140918 16:51:07 <sgothel> 'instead of 60', how is that value known if crashed before ?
20140918 16:51:57 <noxuma> my nvidia version works :)
20140918 16:52:18 <sgothel> well, so no regression - AMD 2fps, NV 60fps
20140918 16:52:20 <noxuma> and it runs at 60, this card is of equal power, it too should run at 60
20140918 16:52:35 <noxuma> i've had it running on AMD long ago ( on jogl 1.1) and it was 60 then too
20140918 16:52:43 <sgothel> try w/ DebugGL enabled, i.e. -Djogl.debug.DebugGL
20140918 16:53:06 <sgothel> 'woulda shoulda coulda' .. no assumptions, shader can have huge impact, as well as state changes
20140918 16:54:27 <noxuma> its 60 on an intel 4000 too ;) trust me, its a safe assmption that it will run 60 here
20140918 16:55:07 <sgothel> let me know what the culprit was, curious
20140918 16:55:41 <noxuma> Yup yup, back to 60
20140918 16:55:59 <noxuma> I was reading gl.glGetShaderInfoLog(newHandle, results[0], results, 0, data, 0);
20140918 16:56:05 <noxuma> but result size was 0
20140918 16:56:16 <noxuma> so it was printing a lot of errors and that killed my framerate
20140918 16:56:24 <noxuma> all is good, looks solid
20140918 16:56:32 <sgothel> great
20140918 16:57:39 <noxuma> Cool so all i have to do is upgrade our pipeline to the latest jogl and it seems all is good
20140918 17:11:42 * hija (~hija@anon) Quit (Quit: hija)
20140918 18:02:03 <monsieur_max> Hi there, do we have an example of code to "record" frames somewhere , for a screencast for example ?
20140918 18:02:13 <monsieur_max> sgothel: maybe you have an idea
20140918 18:03:58 <bbbruce> I recently made something lazy that just opens an ffmpeg process and pipes in raw rgb images
20140918 18:04:38 <monsieur_max> ah, nice idea :)
20140918 18:06:37 <monsieur_max> assuming all my rendering is already made in a FBO, this could be "easily"
20140918 18:06:41 <monsieur_max> done
20140918 18:08:00 <bbbruce> https://pastee.org/esjsp
20140918 18:08:45 <bbbruce> it is part of my own framework but should translate easily to whatever it is you are using
20140918 18:14:45 <monsieur_max> bbbruce: thanks a lot, I'll read that :)
20140918 19:34:11 <noxuma> 32bit jogl seems to give me this problem: not sure why. anyone seen it before: dumpStackID = (*env)->GetStaticMethodID(env, gdiClazz, "dumpStack", "()V");
20140918 19:34:12 <noxuma> NativewindowCommon_FatalError(env, "FatalError jogamp_nativewindow_windows_GDIUtil: can't get method dumpStack");
20140918 19:34:21 <noxuma> 64bit works ok
20140918 19:34:33 <noxuma> (latest jogl)
20140918 21:01:55 * zzuegg|afk (~zzuegg@anon) Quit (Quit: Nettalk6 - www.ntalk.de)
20140918 21:08:14 * noxuma (~markt@anon) has left #jogamp
20140918 21:16:19 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140918 21:23:47 * hija (~hija@anon) has joined #jogamp
20140918 21:27:52 * rmk0-alt (~rmk0@anon) Quit (Ping timeout: 240 seconds)
20140918 21:28:10 * rmk0-alt (~rmk0@anon) has joined #jogamp
20140918 21:28:10 * rmk0-alt (~rmk0@anon) Quit (Changing host)
20140918 21:28:10 * rmk0-alt (~rmk0@anon) has joined #jogamp
20140918 22:03:14 * phao (~phao@anon) has joined #jogamp
20140918 22:09:10 * hija (~hija@anon) Quit (Quit: hija)
20140918 22:15:28 * hija (~hija@anon) has joined #jogamp
20140918 23:13:50 * hija (~hija@anon) Quit (Quit: hija)
20140918 23:14:56 * hija (~hija@anon) has joined #jogamp
20140919 00:45:35 * hija (~hija@anon) Quit (Quit: hija)
20140919 03:48:27 * phao (~phao@anon) Quit (Quit: Leaving)
20140919 05:05:44 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20140919050544.html