# OpenGL Why so many fence/sync types?

## Recommended Posts

Okay, my engine is specifically designed to handle multi-API and cross platform portability.  So far it works pretty well, but there's one thing that's been a big of confusion for me as of late, and that is the need for so many different vendors for fence/sync objects.

For starters, I know what a fence/sync object is and what it does.  Back in the day, it was simple to use either GL_NV_fence or GL_ARB_sync if you were running windows.  For MacOSX, I had GL_APPLE_fence which I used before I learned to use core OpenGL.  Now I am constantly discovering new fence/sync extensions, even a EGL_ANDROID_native_fence_sync.  Is it better to just stick to one?  What's the difference between all of these and the ones I tagged?

Right now, my engine supports the following 4 extensions:

- GL_NV_fence (Windows and Android)

- GL_APPLE_fence (Mac OSX)

- GL_ARB_sync (all platforms supporting core OpenGL)

- GL_APPLE_sync (iOS)

I have not ported my engine to Android yet (due to the sheer insanity of getting anything working in a reasonable amount of time) but when I do, I'm not sure whether I should be using using (or should I say favouring) the native EGL fence or NV's fence.  Instinctively, I would go for GL_NV_fence since EGL is not always something I'd be accessing with Android (since I normally used Java to initialize EGL in the past; Microsoft's EGL initialization code is pretty nice but a pain in the are to debug like most Android dev solutions).

The option to auto detect the appropriate extension is there, and I'll be adding an option to let the user select the preferred one if desired.  My Macbook's battery is getting low so I have to wrap this up quickly.

When I get a chance I'll show the code that handles this.  Why can't it be more simple?  Direct3D11 for Windows (unlike 12) does not have a fence interface to use directly (Xbox does) so I make use of queries which works fine.

Be right back.  I'll explain more in a moment.

Shogun.

##### Share on other sites

It looks like all four extensions give the same functionality except GL_APPLE_sync can only be used in OpenGL ES, GL_ARB_sync and GL_ARB_fence are only for OpenGL, and GL_NV_fence can be used in GL and GLES both (remember that GL and GLES are effectively different APIs, worked on by two separate Khronos working groups, even they have converged in the past few years).

EGL_ANDROID_native_fence_sync works with either GL or GLES as long as you use EGL as your platform layer and effectively does the same job as the other 4 extensions with the addition of being able to synchronize across contexts as well.

##### Share on other sites

OpenGL has core and extensions.

Extensions start by being proposed by a vendor. GL_NV_sync is such example (it was proposed and implemented by NVIDIA, although other vendors can also implement it if they want).

When an extension becomes really useful/widespread but needs some tweaking (i.e. a different behavior in edge cases, a different interface to accomodate for certain hardware) it may be promoted to ARB (IIRC ARB stands for Architecture Review Board, but don't quote me on that). Which is what happened when GL_NV_sync became GL_ARB_sync.

In some cases, an extension is vendor agnostic but it's not popular/stable enough to be ARB, so its name will say EXT.

Once it becomes really useful it can get into core. GL_ARB_sync became into core in OpenGL 3.2; which means it's guaranteed/mandatory to be present starting from OpenGL 3.2

GL_APPLE_sync (GL ES) & GL_APPLE_fence (GL) are Apple's way of doing it. However since Apple already supports GL 3.2; that means you can use GL_ARB_sync.

Long story short you only need to aim for three:

1. GL_ARB_sync (Windows, Linux, OSX, all vendors; unless you don't target GL core >= 3.2)
2. GL_APPLE_sync (iOS)
3. EGL_ANDROID_native_fence_sync (Android)

The rest are just historic relics that vendors must support so old games can still run.

Cheers

Edited by Matias Goldberg

##### Share on other sites

Thanks for the responses.  I'm rather busy today, and I haven't had much of a chance to finish this.

This is my code:

https://github.com/blueshogun96/KunaiEngine/blob/master/source/KeOpenGL/KeOpenGLFence.cpp

/*
* OpenGL fencing/syncronization functionality
* Notes: The renderer needs to determine the appropriate functionality to use based on the
*		  OS and it's available extensions.  Choose wisely.
*
* Legacy and Core OpenGL:
* https://www.opengl.org/registry/specs/NV/fence.txt
* https://www.opengl.org/registry/specs/APPLE/fence.txt
* https://www.opengl.org/registry/specs/ARB/sync.txt
*
* OpenGL ES:
* https://www.khronos.org/registry/vg/extensions/KHR/EGL_KHR_fence_sync.txt
* https://www.khronos.org/registry/gles/extensions/APPLE/APPLE_sync.txt
*
*/

#include "Ke.h"
#include "KeRenderDevice.h"
#include "KeOpenGLRenderDevice.h"
#include "KeOpenGLFence.h"

/*
* Debugging macros
*/
#define DISPDBG_R( a, b ) { DISPDBG( a, b ); return; }
#define DISPDBG_RB( a, b ) { DISPDBG( a, b ); return false; }
#define OGL_DISPDBG( a, b ) error = glGetError(); if(error) { DISPDBG( a, b << "\nError code: (" << error << ")" ); }
#define OGL_DISPDBG_R( a, b ) error = glGetError(); if(error) { DISPDBG( a, b << "\nError code: (" << error << ")" ); return; }
#define OGL_DISPDBG_RB( a, b ) error = glGetError(); if(error) { DISPDBG( a, b << "\nError code: (" << error << ")" ); return false; }

/*
* NVIDIA fencing functions (GL_NV_fence)
*/
bool KeOpenGLCreateFenceNV( IKeOpenGLFence** fence )
{
#if GL_NV_fence
GLenum error = glGetError();

/* Generate a new fence */
glGenFencesNV( 1, &(*fence)->fence );
OGL_DISPDBG_RB( KE_ERROR, "Error generating new fence!" );

return true;
#else
DISPDBG_RB( KE_ERROR, "GL_NV_fence not supported!" );
#endif
}

bool KeOpenGLInsertFenceNV( IKeOpenGLFence** fence )
{
#if GL_NV_fence
GLenum error = glGetError();

/* Set the fence */
glSetFenceNV( (*fence)->fence, GL_ALL_COMPLETED_NV );
OGL_DISPDBG_RB( KE_ERROR, "Error setting new fence!" );

return true;
#else
DISPDBG_RB( KE_ERROR, "GL_NV_fence not supported!" );
#endif
}

bool KeOpenGLTestFenceNV( IKeOpenGLFence* fence )
{
#if GL_NV_fence
if( glTestFenceNV( fence->fence ) )
return true;

return false;
#else
DISPDBG_RB( KE_ERROR, "GL_NV_fence not supported!" );
#endif
}

void KeOpenGLBlockOnFenceNV( IKeOpenGLFence* fence )
{
#if GL_NV_fence
glFinishFenceNV( fence->fence );
#else
DISPDBG( KE_ERROR, "GL_NV_fence not supported!" );
#endif
}

void KeOpenGLDeleteFenceNV( IKeOpenGLFence* fence )
{
#if GL_NV_fence
glDeleteFencesNV( 1, &fence->fence );
#else
DISPDBG( KE_ERROR, "GL_NV_fence not supported!" );
#endif
}

bool KeOpenGLIsFenceNV( IKeOpenGLFence* fence )
{
#if GL_NV_fence
if( glTestFenceNV( fence->fence ) )
return true;

return false;
#else
DISPDBG_RB( KE_ERROR, "GL_NV_fence not supported!" );
#endif
}

/*
* APPLE fencing functions (GL_APPLE_fence)
*/
bool KeOpenGLCreateFenceAPPLE( IKeOpenGLFence** fence )
{
#if GL_APPLE_fence
GLenum error = glGetError();

/* Generate a new fence */
glGenFencesAPPLE( 1, &(*fence)->fence );
OGL_DISPDBG_RB( KE_ERROR, "Error creating new fence!" );

return true;
#else
DISPDBG_RB( KE_ERROR, "GL_APPLE_fence not supported!" );
#endif
}

bool KeOpenGLInsertFenceAPPLE( IKeOpenGLFence** fence )
{
#if GL_APPLE_fence
GLenum error = glGetError();

/* Set the fence */
glSetFenceAPPLE( (*fence)->fence );
OGL_DISPDBG_RB( KE_ERROR, "Error setting new fence!" );

return true;
#else
DISPDBG_RB( KE_ERROR, "GL_APPLE_fence not supported!" );
#endif
}

bool KeOpenGLTestFenceAPPLE( IKeOpenGLFence* fence )
{
#if GL_APPLE_fence
if( glTestFenceAPPLE( fence->fence ) )
return true;

return false;
#else
DISPDBG_RB( KE_ERROR, "GL_APPLE_fence not supported!" );
#endif

}

void KeOpenGLBlockOnFenceAPPLE( IKeOpenGLFence* fence )
{
#if GL_APPLE_fence
glFinishFenceAPPLE( fence->fence );
#else
DISPDBG( KE_ERROR, "GL_APPLE_fence not supported!" );
#endif

}

void KeOpenGLDeleteFenceAPPLE( IKeOpenGLFence* fence )
{
#if GL_APPLE_fence
glDeleteFencesAPPLE( 1, &fence->fence );
#else
DISPDBG( KE_ERROR, "GL_APPLE_fence not supported!" );
#endif

}

bool KeOpenGLIsFenceAPPLE( IKeOpenGLFence* fence )
{
#if GL_APPLE_fence
if( glTestFenceAPPLE( fence->fence ) )
return true;

return false;
#else
DISPDBG_RB( KE_ERROR, "GL_APPLE_fence not supported!" );
#endif

}

/*
* ARB synchronization functions (GL_ARB_sync)
*/
bool KeOpenGLCreateFenceARB( IKeOpenGLFence** fence )
{
#if GL_ARB_sync
/* ARB_sync creates and inserts the fence with the same API call */
return true;
#else
DISPDBG_RB( KE_ERROR, "GL_ARB_sync not supported!" );
#endif
}

bool KeOpenGLInsertFenceARB( IKeOpenGLFence** fence )
{
#if GL_ARB_sync
GLenum error = glGetError();

/* Create sync object.  It will automatically be set in the unsignaled state
if successful. */
(*fence)->sync = glFenceSync( GL_SYNC_GPU_COMMANDS_COMPLETE, 0 );
OGL_DISPDBG_RB( KE_ERROR, "Error creating and setting new sync object!" );

return true;
#else
DISPDBG_RB( KE_ERROR, "GL_ARB_sync not supported!" );
#endif
}

bool KeOpenGLTestFenceARB( IKeOpenGLFence* fence )
{
#if GL_ARB_sync
int signaled = GL_UNSIGNALED;

/* Test this sync object for it's status and return the result */
glGetSynciv( fence->sync, GL_SYNC_STATUS, sizeof( int ), NULL, &signaled );

return signaled ? true : false;
#else
DISPDBG_RB( KE_ERROR, "GL_ARB_sync not supported!" );
#endif
}

void KeOpenGLBlockOnFenceARB( IKeOpenGLFence* fence )
{
#if GL_ARB_sync
/* Stall the current thread until this sync object is signaled */
//glWaitSync( fence->sync, 0, GL_TIMEOUT_IGNORED );
GLenum ret = glClientWaitSync( fence->sync, GL_SYNC_FLUSH_COMMANDS_BIT, GL_TIMEOUT_IGNORED );
#else
DISPDBG( KE_ERROR, "GL_ARB_sync not supported!" );
#endif

}

void KeOpenGLDeleteFenceARB( IKeOpenGLFence* fence )
{
#if GL_ARB_sync
glDeleteSync( fence->sync );
#else
DISPDBG( KE_ERROR, "GL_ARB_sync not supported!" );
#endif
}

bool KeOpenGLIsFenceARB( IKeOpenGLFence* fence )
{
#if GL_ARB_sync
return glIsSync( fence->sync ) ? true : false;
#else
DISPDBG_RB( KE_ERROR, "GL_ARB_sync not supported!" );
#endif
}

/*
* APPLE synchronization functions (GL_APPLE_sync)
*/
bool KeOpenGLCreateFenceAPPLEiOS( IKeOpenGLFence** fence )
{
#if GL_APPLE_sync
/* ARB_sync creates and inserts the fence with the same API call */
return true;
#else
DISPDBG_RB( KE_ERROR, "GL_APPLE_sync not supported!" );
#endif
}

bool KeOpenGLInsertFenceAPPLEiOS( IKeOpenGLFence** fence )
{
#if GL_APPLE_sync
GLenum error = glGetError();

/* Create sync object.  It will automatically be set in the unsignaled state
if successful. */
(*fence)->sync = glFenceSyncAPPLE( GL_SYNC_GPU_COMMANDS_COMPLETE_APPLE, 0 );
OGL_DISPDBG_RB( KE_ERROR, "Error creating and setting new sync object!" );

return true;
#else
DISPDBG_RB( KE_ERROR, "GL_APPLE_sync not supported!" );
#endif
}

bool KeOpenGLTestFenceAPPLEiOS( IKeOpenGLFence* fence )
{
#if GL_APPLE_sync
int signaled = GL_UNSIGNALED;

/* Test this sync object for it's status and return the result */
glGetSyncivAPPLE( fence->sync, GL_SYNC_STATUS_APPLE, sizeof( int ), NULL, &signaled );

return signaled ? true : false;
#else
DISPDBG_RB( KE_ERROR, "GL_APPLE_sync not supported!" );
#endif
}

void KeOpenGLBlockOnFenceAPPLEiOS( IKeOpenGLFence* fence )
{
#if GL_APPLE_sync
/* Stall the current thread until this sync object is signaled */
//glWaitSync( fence->sync, 0, GL_TIMEOUT_IGNORED );
GLenum ret = glClientWaitSyncAPPLE( fence->sync, GL_SYNC_FLUSH_COMMANDS_BIT_APPLE, GL_TIMEOUT_IGNORED );
#else
DISPDBG( KE_ERROR, "GL_APPLE_sync not supported!" );
#endif

}

void KeOpenGLDeleteFenceAPPLEiOS( IKeOpenGLFence* fence )
{
#if GL_APPLE_sync
glDeleteSyncAPPLE( fence->sync );
#else
DISPDBG( KE_ERROR, "GL_APPLE_sync not supported!" );
#endif
}

bool KeOpenGLIsFenceAPPLEiOS( IKeOpenGLFence* fence )
{
#if GL_APPLE_sync
return glIsSyncAPPLE( fence->sync ) ? true : false;
#else
DISPDBG_RB( KE_ERROR, "GL_APPLE_sync not supported!" );
#endif
}

bool ( *KeOpenGLCreateFence[4] )( IKeOpenGLFence** ) = { KeOpenGLCreateFenceARB, KeOpenGLCreateFenceNV, KeOpenGLCreateFenceAPPLE, KeOpenGLCreateFenceAPPLEiOS };
bool ( *KeOpenGLInsertFence[4] )( IKeOpenGLFence** ) = { KeOpenGLInsertFenceARB, KeOpenGLInsertFenceNV, KeOpenGLInsertFenceAPPLE, KeOpenGLInsertFenceAPPLEiOS };
bool ( *KeOpenGLTestFence[4] )( IKeOpenGLFence* ) = { KeOpenGLTestFenceARB, KeOpenGLTestFenceNV, KeOpenGLTestFenceAPPLE, KeOpenGLTestFenceAPPLEiOS };
void ( *KeOpenGLBlockOnFence[4] )( IKeOpenGLFence* ) = { KeOpenGLBlockOnFenceARB, KeOpenGLBlockOnFenceNV, KeOpenGLBlockOnFenceAPPLE, KeOpenGLBlockOnFenceAPPLEiOS };
void ( *KeOpenGLDeleteFence[4] )( IKeOpenGLFence* ) = { KeOpenGLDeleteFenceARB, KeOpenGLDeleteFenceNV, KeOpenGLDeleteFenceAPPLE, KeOpenGLDeleteFenceAPPLEiOS };
bool ( *KeOpenGLIsFence[4] )( IKeOpenGLFence* ) = { KeOpenGLIsFenceARB, KeOpenGLIsFenceNV, KeOpenGLIsFenceAPPLE, KeOpenGLIsFenceAPPLEiOS };

/*
* Name: IKeOpenGLFence::Destroy
* Desc: Handles destruction of this interface instance.
*/
void IKeOpenGLFence::Destroy()
{
/* Destroy the fence */
KeOpenGLDeleteFence[vendor]( this );

/* Delete this instance */
delete this;
}

/*
* Name: IKeOpenGLFence::Insert
* Desc: Inserts GPU fence object into the execution pipeline.
*/
bool IKeOpenGLFence::Insert()
{
/* Sanity check */
if( !fence && !sync )
return false;

IKeOpenGLFence* f = this;

return KeOpenGLInsertFence[vendor]( (IKeOpenGLFence**) &f );
}

/*
* Name: IKeOpenGLFence::Test
* Desc: Returns true if this all GPU commands have been completed since
*		 this fence was set.  If there are still GPU commands pending,
*		 returns false.
*/
bool IKeOpenGLFence::Test()
{
if( !fence && !sync )
return false;

return KeOpenGLTestFence[vendor]( this );
}

/*
* Name: IKeOpenGLFence::Block
* Desc: Stalls the current thread until the fence has been crossed.
*/
void IKeOpenGLFence::Block()
{
KeOpenGLBlockOnFence[vendor]( this );
}

/*
* Name: IKeOpenGLFence::Valid
* Desc: Tests this fence object for a valid fence.
*/
bool IKeOpenGLFence::Valid()
{
if( !fence && !sync )
return false;

return KeOpenGLIsFence[vendor]( this );
}


Now, i know this can be better written, I just never got around to doing it yet.  My über is here, better get outside.  Be right back... again!

Shogun.

##### Share on other sites

OpenGL has core and extensions.

Extensions start by being proposed by a vendor. GL_NV_sync is such example (it was proposed and implemented by NVIDIA, although other vendors can also implement it if they want).

When an extension becomes really useful/widespread but needs some tweaking (i.e. a different behavior in edge cases, a different interface to accomodate for certain hardware) it may be promoted to ARB (IIRC ARB stands for Architecture Review Board, but don't quote me on that). Which is what happened when GL_NV_sync became GL_ARB_sync.

In some cases, an extension is vendor agnostic but it's not popular/stable enough to be ARB, so its name will say EXT.

Once it becomes really useful it can get into core. GL_ARB_sync became into core in OpenGL 3.2; which means it's guaranteed/mandatory to be present starting from OpenGL 3.2

GL_APPLE_sync (GL ES) & GL_APPLE_fence (GL) are Apple's way of doing it. However since Apple already supports GL 3.2; that means you can use GL_ARB_sync.

Long story short you only need to aim for three:

1. GL_ARB_sync (Windows, Linux, OSX, all vendors; unless you don't target GL core >= 3.2)
2. GL_APPLE_sync (iOS)
3. EGL_ANDROID_native_fence_sync (Android)

The rest are just historic relics that vendors must support so old games can still run.

Cheers

Good response, that's a general rule of thumb to remember.  However, I'm still going to support GL_NV_fence not for Windows, but for Android since the EGL extension requires you to have the EGL context/display setup by your own code in advance.  I've always initialized OpenGL ES using Java in the past, and I plan to use SDL2 to initialize OpenGL ES for Android, so I believe it's safe to assume I won't have easy access to the EGLDisplay pointer.  Because of this, I will likely use GL_NV_fence since it is supported on Android.  At least, it is supported on every Android devices I own with 5.0+ installed.  But I will cross that bridge when I come to it.

Shogun

##### Share on other sites

Because of this, I will likely use GL_NV_fence since it is supported on Android. At least, it is supported on every Android devices I own with 5.0+ installed. But I will cross that bridge when I come to it.

If I have ever learned any wrt to OpenGL (ES) is never assume anything, unless its core functionality. The feature maybe on every device you own, but are you developing your application for personal/widespread usage? If the latter, then you will get bit sooner than later. If you notice all the different sync function entry points are somewhat similar in flavor. So to make you life easier ( ...and I know others will object to this ), but why don't you just write a simple wrapper that encapsulate all the different sync function currently in the wild. Extension checking is straightforward and simple. In fact, I did the same thing for sync objects in my render and I don't have to worry about which sync type is being used since I use a higher level object. Work so far on Android, Linux and Windows. I'll cross the Mac bridge whenever I decide to support it, but even then all I have to do is add the support to the wrapper object.

## Create an account

Register a new account

• ### Forum Statistics

• Total Topics
627757
• Total Posts
2978950
• ### Similar Content

• Hello! As an exercise for delving into modern OpenGL, I'm creating a simple .obj renderer. I want to support things like varying degrees of specularity, geometry opacity, things like that, on a per-material basis. Different materials can also have different textures. Basic .obj necessities. I've done this in old school OpenGL, but modern OpenGL has its own thing going on, and I'd like to conform as closely to the standards as possible so as to keep the program running correctly, and I'm hoping to avoid picking up bad habits this early on.
Reading around on the OpenGL Wiki, one tip in particular really stands out to me on this page:
For something like a renderer for .obj files, this sort of thing seems almost ideal, but according to the wiki, it's a bad idea. Interesting to note!
So, here's what the plan is so far as far as loading goes:
Set up a type for materials so that materials can be created and destroyed. They will contain things like diffuse color, diffuse texture, geometry opacity, and so on, for each material in the .mtl file. Since .obj files are conveniently split up by material, I can load different groups of vertices/normals/UVs and triangles into different blocks of data for different models. When it comes to the rendering, I get a bit lost. I can either:
Between drawing triangle groups, call glUseProgram to use a different shader for that particular geometry (so a unique shader just for the material that is shared by this triangle group). or
Between drawing triangle groups, call glUniform a few times to adjust different parameters within the "master shader", such as specularity, diffuse color, and geometry opacity. In both cases, I still have to call glBindTexture between drawing triangle groups in order to bind the diffuse texture used by the material, so there doesn't seem to be a way around having the CPU do *something* during the rendering process instead of letting the GPU do everything all at once.
The second option here seems less cluttered, however. There are less shaders to keep up with while one "master shader" handles it all. I don't have to duplicate any code or compile multiple shaders. Arguably, I could always have the shader program for each material be embedded in the material itself, and be auto-generated upon loading the material from the .mtl file. But this still leads to constantly calling glUseProgram, much more than is probably necessary in order to properly render the .obj. There seem to be a number of differing opinions on if it's okay to use hundreds of shaders or if it's best to just use tens of shaders.
So, ultimately, what is the "right" way to do this? Does using a "master shader" (or a few variants of one) bog down the system compared to using hundreds of shader programs each dedicated to their own corresponding materials? Keeping in mind that the "master shaders" would have to track these additional uniforms and potentially have numerous branches of ifs, it may be possible that the ifs will lead to additional and unnecessary processing. But would that more expensive than constantly calling glUseProgram to switch shaders, or storing the shaders to begin with?
With all these angles to consider, it's difficult to come to a conclusion. Both possible methods work, and both seem rather convenient for their own reasons, but which is the most performant? Please help this beginner/dummy understand. Thank you!

• I want to make professional java 3d game with server program and database,packet handling for multiplayer and client-server communicating,maps rendering,models,and stuffs Which aspect of java can I learn and where can I learn java Lwjgl OpenGL rendering Like minecraft and world of tanks

• A friend of mine and I are making a 2D game engine as a learning experience and to hopefully build upon the experience in the long run.

-What I'm using:
C++;. Since im learning this language while in college and its one of the popular language to make games with why not.     Visual Studios; Im using a windows so yea.     SDL or GLFW; was thinking about SDL since i do some research on it where it is catching my interest but i hear SDL is a huge package compared to GLFW, so i may do GLFW to start with as learning since i may get overwhelmed with SDL.
-Questions
Knowing what we want in the engine what should our main focus be in terms of learning. File managements, with headers, functions ect. How can i properly manage files with out confusing myself and my friend when sharing code. Alternative to Visual studios: My friend has a mac and cant properly use Vis studios, is there another alternative to it?

• Both functions are available since 3.0, and I'm currently using glMapBuffer(), which works fine.
But, I was wondering if anyone has experienced advantage in using glMapBufferRange(), which allows to specify the range of the mapped buffer. Could this be only a safety measure or does it improve performance?
Note: I'm not asking about glBufferSubData()/glBufferData. Those two are irrelevant in this case.
• By xhcao
Before using void glBindImageTexture(    GLuint unit, GLuint texture, GLint level, GLboolean layered, GLint layer, GLenum access, GLenum format), does need to make sure that texture is completeness.

• 11
• 10
• 10
• 23
• 14