After I took the Windows update 1903, the native method GDI.SwapBuffers from Windows cause JVM to crash.
The native method GDI.SwapBuffers it's called through glCanvas.swapBuffers() method. Does any of you encounter this problem? Do you know an alternative to swapBuffers method? In other words how can i draw the objects without calling swapBuffers at all? |
Administrator
|
Hello
Actually, it crashes on this call but the problem is probably elsewhere in the driver. Please provide much more information: https://jogamp.org/wiki/index.php?title=Jogl_FAQ#Bugreports_.26_Testing What you're looking for is impossible, you must call swapBuffers(). Either use another graphics card with a driver still working reliably with Microsoft Windows 10 update 1903 or try to use another version of the current driver or report this bug against the GPU designer (Nvidia, AMD, Intel, ...) or use another operating system with a working OpenGL driver for your current hardware (any GNU Linux distribution) or maybe use Mesa: https://fdossena.com/?p=mesa/index.frag
Julien Gouesse | Personal blog | Website
|
After i uninstall the Windows version 1903, and restore the previous version, everything was working again.
It is a problem with that Windows version and needs to be fixed by Microsoft, but i was thinking if there is a workaround to make swapBuffers method to work if that Windows version is installed. |
Administrator
|
Thank you for the feedback. Please can you indicate which hardware is concerned? It could help someone having the same problem.
It needs to be fixed by the GPU manufacturer. There are some render quirks to work around some driver bugs in JOGL but I still consider that the corporations providing the drivers should fix their bugs themselves. Have you checked whether your bug has already been reported?
Julien Gouesse | Personal blog | Website
|
I have NVIDIA GeForce MX150 From: gouessej [via jogamp] <ml+[hidden email]>
Thank you for the feedback. Please can you indicate which hardware is concerned? It could help someone having the same problem.
Julien Gouesse |
Personal blog |
Website If you reply to this email, your message will be added to the discussion below: http://forum.jogamp.org/GDI-SwapBuffers-is-crashing-after-Windows-Update-tp4040010p4040014.html
To unsubscribe from GDI.SwapBuffers is crashing after Windows Update,
click here. |
Administrator
|
In reply to this post by gouessej
On 8/30/19 1:24 PM, gouessej [via jogamp] wrote:
> Hello > > Actually, it crashes on this call but the problem is probably elsewhere in the > driver. Please provide much more information: > https://jogamp.org/wiki/index.php?title=Jogl_FAQ#Bugreports_.26_Testing I posted the my latest Windows unit tests, but indeed the machine wasn't updated to this Windows 10 update 1903 (May 2019). Doing this errand today, maybe you are "lucky" and I reproduce the issue :) > > What you're looking for is impossible, you must call swapBuffers(). Either use > another graphics card with a driver still working reliably with Microsoft > Windows 10 update 1903 or try to use another version of the current driver or > report this bug against the GPU designer (Nvidia, AMD, Intel, ...) or *use > another operating system with a working OpenGL driver for your current > hardware (any GNU Linux distribution)* or maybe use Mesa: > https://fdossena.com/?p=mesa/index.frag It is a PIA, Julien - yes. However, I have to say one good thing about the API stability. Our Windows regression tests on latest Java11 tip showed the least regressions. *duck and hide* Maybe Windows just wanted to catch up with the regression count, pre this Windows update we only had 4 unit failures :) More on Mesa3D hardware driver (in)stability later, I could only use the software driver. ~Sven |
Administrator
|
Updated Windows machine.
Windows 10 - 1903 Update NVidia GeForce GTX-460 - Driver 391.35 - Display Server 23.21.13.9135 - Driver 388.13 - Display Server 23.21.13.8813 I have tested 2 unit tests ad-hoc and no issues whatsoever Platform: WINDOWS / Windows 10 10.0 (10.0.0), amd64 (X86_64, GENERIC_ABI), 6 cores, littleEndian true MachineDataInfo: runtimeValidated true, 32Bit false, primitive size / alignment: int8 1 / 1, int16 2 / 2 int 4 / 4, long 4 / 4 int32 4 / 4, int64 8 / 8 float 4 / 4, double 8 / 8, ldouble 16 / 16 pointer 8 / 8, page 4096 Platform: Java Version: 11.0.4 (11.0.4u0), VM: OpenJDK 64-Bit Server VM, Runtime: OpenJDK Runtime Environment Platform: Java Vendor: AdoptOpenJDK, https://adoptopenjdk.net/, JavaSE: true, Java9: true, Java6: true, dynamicLib: true, AWT enabled: true GL Profile GLProfile[GL4bc/GL4bc.hw] GL Version 4.6 (Compat profile, arb, compat[ES2, ES3, ES31, ES32], FBO, hardware) - 4.6.0 NVIDIA 391.35 [GL 4.6.0, vendor 391.35.0 (NVIDIA 391.35)] Quirks [NoDoubleBufferedBitmap, NoSurfacelessCtx] Impl. class jogamp.opengl.gl4.GL4bcImpl GL_VENDOR NVIDIA Corporation GL_RENDERER GeForce GTX 460/PCIe/SSE2 GL_VERSION 4.6.0 NVIDIA 391.35 GLSL true, has-compiler-func: true, version: 4.60 NVIDIA / 4.60.0 GL FBO: basic true, full true GL_EXTENSIONS 325 GLX_EXTENSIONS 30 ~Sven On 9/2/19 3:19 PM, Sven Gothel [via jogamp] wrote: > On 8/30/19 1:24 PM, gouessej [via jogamp] wrote: >> Hello >> >> Actually, it crashes on this call but the problem is probably elsewhere in the >> driver. Please provide much more information: >> https://jogamp.org/wiki/index.php?title=Jogl_FAQ#Bugreports_.26_Testing > > I posted the my latest Windows unit tests, but indeed the machine wasn't > updated to this Windows 10 update 1903 (May 2019). > > Doing this errand today, maybe you are "lucky" and I reproduce the issue :) > >> >> What you're looking for is impossible, you must call swapBuffers(). Either use >> another graphics card with a driver still working reliably with Microsoft >> Windows 10 update 1903 or try to use another version of the current driver or >> report this bug against the GPU designer (Nvidia, AMD, Intel, ...) or *use >> another operating system with a working OpenGL driver for your current >> hardware (any GNU Linux distribution)* or maybe use Mesa: >> https://fdossena.com/?p=mesa/index.frag > > It is a PIA, Julien - yes. > However, I have to say one good thing about the API stability. > Our Windows regression tests on latest Java11 tip showed the least > regressions. *duck and hide* > > Maybe Windows just wanted to catch up with the regression count, > pre this Windows update we only had 4 unit failures :) > > More on Mesa3D hardware driver (in)stability later, > I could only use the software driver. > > ~Sven > > |
Administrator
|
Note that he's probably using it on a laptop, maybe Optimus chooses another chipset. Forcing "high performance" might help as we can't use GPU affinity in JOGL.
Julien Gouesse | Personal blog | Website
|
In reply to this post by Sven Gothel
This is the full specifications of my machine:
GL Profile GLProfile[GL4bc/GL4bc.hw] GL Version 4.6 (Compat profile, arb, compat[ES2, ES3, ES31, ES32], FBO, hardware) - 4.6.0 NVIDIA 398.75 [GL 4.6.0, vendor 398.75.0 (NVIDIA 398.75)] Quirks [NoDoubleBufferedBitmap, NoSurfacelessCtx] Impl. class jogamp.opengl.gl4.GL4bcImpl GL_VENDOR NVIDIA Corporation GL_RENDERER GeForce MX150/PCIe/SSE2 GL_VERSION 4.6.0 NVIDIA 398.75 GLSL true, has-compiler-func: true, version: 4.60 NVIDIA / 4.60.0 GL FBO: basic true, full true GL_EXTENSIONS 376 GLX_EXTENSIONS 30 |
Administrator
|
Have you modified the power mode to force the use of the dedicated video card?
Julien Gouesse | Personal blog | Website
|
Free forum by Nabble | Edit this page |