creation of main shader to ShaderComposer and
collection of ShaderComponent to osg::State.
Also added very basic shader set up in osgshadecomposition example.
In State::initializeExtensionProcs() the _glMaxTextureUnits is calculated based on osg::getGLVersionNumber().
At least for ES 2.0 this function will return 0.f since the version string will look like "OpenGL ES 2.0 ...".
My proposal doesn't touch getGLVersionNumber(), since desktop OpenGL 2.0 isn't OpenGL ES 2.0.
So i changed the conditions in State::initializeExtensionProcs() for getting the number via glGetIntegerv()."
osg::Camera* c = createCamera();
c->attach( osg::Camera::COLOR_BUFFER0, texture3d, 0,
osg::Camera::FACE_CONTROLLED_BY_GEOMETRY_SHADER );
it works also for cubemap textures and 2d texture arrays
"
bool observer_ptr<>::lock(ref_ptr<T>&) to avoid the temporary ref_ptr<>'s
being created and destroyed on the stack along with the associated ref/unref() operations
Changes:
- new mouse wheel zoom/movement/center functionality
- ability to fix vertical axis (important for CAD)
- possibility to specify values as absolute values or relative to model size
- kind of backward compatibility by flags passed to constructor
- and much more
- restructuring classes to use kind of hierarchy and standard way of event processing (handle methods). This way, there is much more code reusability and it is more easy to develop new kinds of manipulators.
Briefly, the new architecture keeps MatrixManipulator as base abstract class. StandardManipulator is the feature-rich standard manipulator with two main descendant classes: OrbitManipulator and FirstPersonManipulator. OrbitManipulator is base class for all trackball style manipulators, based on center, rotation and distance from center. FirstPersonManipulator is base for walk or fly style manipulators, using position and rotation for camera manipulation.
"
Changes by Robert: Replaced osg::Vec3 by osg::Vec3d, introduced DEFAULT_SETTINGS enum and usage. Added frame time member variables in prep for improving throw animation when vysync is off.
osg::TransferFunction1D* tf(new osg::TransferFunction1D());
tf->allocate(18);
tf->setColor(-10000,osg::Vec4( 1.0, 1.0, 1.0,1.0));
tf->setColor(-1e-6,osg::Vec4( 1.0, 1.0, 1.0,1.0));
tf->setColor( 0,osg::Vec4( 0.0, 0.0, 1.0,1.0));
Remark: The value -1e-6 may be added (falsely) because of a rounding error.
The attached fix prevents assigning NaN values to the image.
"
Texture2DMultismaple as name suggests provides means to directly access subsamples of rendered FBO target. (GLSL 1.5 texelFetch call).
Recently I was working on deferred renderer with OSG, during that I noticed there is no support for multisampled textures (GL_ARB_texture_multisample extension). After consultations with Paul Martz and Wojtek Lewandowski I added Texture2DMultisample class and made few necessary changes around osg::FrameBufferObject, osg::Texture and osgUtil::RenderStage classes."
and from follow email:
"Fixed. According to ARB_texture_multisample extension specification multisample textures don't need TexParameters since they can only be fetched with texelFetch."
- compile errors on windows when compiled with UNICODE flag
- warnings for duplicate WIN32_LEAN_AND_MEAN. I think this should
better fixed by adding WIN32_LEAN_AND_MEAN to vcproj preprocessor
list."
Many IDEs will capture the output of processes and highlight stderr messages in red. After upgrading from OSG 1.2, I was noticing that none of the error messages were being highlighted, which I find quite useful."
* Change OcclusionQueryNode::getPassed to take a NodeVisitor rather than the distance from BS center to the eye point. Change where CullVisitor calls this method to use the new parameters.
* getPassed now exits early and returns true to avoid blinking / blink-in of geometry for the first frame or for out-of-range LOD children coming back into view.
* getPassed now considers the distance from the near plane to the bounding sphere (rather than eye point to bounding sphere) when determining if the viewer is "inside" the bounding sphere or not."
The fix is basically as follows:
if( GL3 )
Query indexed extension strings.
else
Query the old way.
The "else" branch is re-indented but otherwise shouldn't contain any changes."
From Robert Osfield, added #if !defined(OSG_GLES1_AVAILABLE) && !defined(OSG_GLES2_AVAILABLE) to new block to prevent it
being compiled under GLES where no glGetStringi exists.
The fix is basically as follows:
if( GL3 )
Query indexed extension strings.
else
Query the old way.
The "else" branch is re-indented but otherwise shouldn't contain any changes."
I create a compositeviewer with two views that share a context. One view is deleted. Texture::TextureObjectSet::discardAllTextureObjects is called, but this does not reset _tail. Now the texture object is again created and addToBack is called from Texture::TextureObjectSet::takeOrGenerate. In addToBack (line 612) _tail is now not 0 (although the list should be empty) and a loop is created from the texture object to itself. Then when the second view is deleted, Texture::TextureObjectSet::deleteAllTextureObjects loops forever. Setting _tail to 0 fixes it for me (see attached)."
which utilises a global recursive mutex that is dedicated to manage Observer and ObserverSet.
The new global mutex for observers avoids problems with deadlocks that were occurring previously when
an osg::Refenced object was being deleted at the same time as on osg::ObserverNodePath.
the extra capability of making it possible for Observers to assume ownership of a object that would otherwsie be deleted.
Added a thread safe ref_ptr<T> observer_ptr<T>::lock() method for robust access to an observed object. This
makes observer_ptr<> more equivilant to boosts weak_ptr.
Introduced the OSG_NOTIFY_DISABLE Cmake variable + include/osg/Config #define to control whether the OpenSceneGraph build
should disable the notification system completely. By setting OSG_NOTIFY_DISABLE to ON in CMake and then rebuilding the
the OSG you can get a slightly smaller (~1%) and more slightly efficient library which can be good for shipping applications,
but with downside of reduced ability to detect runtime problems and their causes.
almost all Linux distributions. Although it is 100% compatible with ld,
by default it gives an error if a library has unresolved symbols at link
time, that is, it has set -Wl,--no-undefined by default. Debian folks
have found that libosg.so and libosgDB.so use some functions belonging
to libdl.so {dlsym,dlopen,dlclose,dlerror} without linking to it.
My changes link those two libraries to libdl.so explicitly in the same
way it is already done for libm.so and librt.so."
From Robert Osfield, I didn't merge the change of parameter type of IntersectKdTree::intersect() as the internal maths is all done in Vec3s. Keeping Vec3 here hasn't effected the test results.
1) Add getShadowComparison() accessor function to osg::Texture class
2) Modify ReaderWriterTiff::writeTifStream() and _readColor() (in Image.cpp) to handle pixelFormat==GL_DEPTH_COMPONENT as if it were GL_LUMINANCE
3) Modify the Texture classes of the ive and osg plug-ins so that they save/load the following Texture members: _use_shadow_comparison, _shadow_compare_func and _shadow_texture_mode
"
glActiveTexture on every State::apply after more than one texunits have been
used.
This is against 2.9.6 (I think SVN head is the same)
Quick Synopsis:
New functions:
State::applyModeOnTexUnit
State::applyAttributeOnTexUnit
State::applyModeMapOnTexUnit
State::applyAttributeMapOnTexUnit
State::applyModeListOnTexUnit
State::applyAttributeListOnTexUnit
All copies of the normal versions, but they also set the active tex unit if
absolutely necessary (i.e. only if they call something OpenGL).
State::apply (*2)
State::applyTextureAttribute
Changed to call the above functions and no longer call setActiveTextureUnit
themselves.
State::setActiveTextureUnit
Made inline, so the benefit of having applyModeOnTexUnit (etc) inline
is retained.
"
To enable the automatic attachment of the required update callback to call osg::Image::update(..) subclasses from osg::Image will
need to implement the osg::Image::requestUpdateCall() and return true, and implement the osg::Image::update(NodeVisitor*) method to recieve the update call during the update traversal.
There is a small bug with your change and SubloadCallback, the texture object is destroy at each call of Texture2D::apply because the modified count is never updated when using SubloadCallback.
I have made a small fix to avoid that, see attachement."
From Robert Osfield, added an if (_image.valid()) before the getModified() that Fabien added to avoid problems when no _image is assigned but an subload callback is.
As Robert pointed out, i opted for modifying the resize code where the break is to minimize code changes, avoiding the duplicate resize of the viewport with the use of a vector and a search for duplicates. Not very elegant (avoiding an effect of a cause), another approach could be ripping out the method osg::Camera::setViewport(osg::Viewport*) which is more inline with Roberts rationale behind not to share viewports between cameras and left only its overloaded method setViewport(x,y,width,height). But this approach need some refactoring due to the intense use of the method. Notice also that the resize works well without this change if no sharing occurs, and the user of the method can opt for always call setViewport with a new instance.
"
Note from Robert Osfield, changed this submission to use an std::set<Viewport*> rather than an std::vector<> as it keeps the code a bit cleaner and more compact.
was able to follow the problem until following addition to Texture.cpp:
// GLES doesn't cope with internal formats of 1,2,3 and 4 so map them to
the appropriate equivilants.
if (_internalFormat==1) _internalFormat = GL_ALPHA;
if (_internalFormat==2) _internalFormat = GL_LUMINANCE_ALPHA;
if (_internalFormat==3) _internalFormat = GL_RGB;
if (_internalFormat==4) _internalFormat = GL_RGBA;
The problem is that internal format "1" corresponds to GL_LUMINANCE, not
GL_ALPHA. I double checked this from the Red Book. Fixed version is
attached to the email."
DisplaySettings now define COLOR and DEPTH as defaults for implicit buffers. Consequently Camera by default uses the same defaults through USE_DISPLAY_SETTINGS_MASK. However, particular Camera mask can be easily overriden through Camera::setImplicitBufferAttachmentMask method. I hope, that in this way we can have global control over implicit buffer defaults, and we can still retain fine grained control at Camera level.
I have also replaced original unsigned ints used to store masks to signed ints because complier resolves enums as signed integer (I got a number of warnings with unsigned int)."
settimeofday), OSG animations will freeze your application because
osg::Timer uses gettimeofday internally on non-Win32 platforms. This
is wrong and should be replace with times(2) or clock_gettime(2).
The attached patch fixes the issue in a binary-compatible way by using
clock_gettime when it's available, and falling back to gettimeofday
when it's not."
Fixed to osg::Texture for GLES support.
Added automatic GLenum mode mappings in osg::PrimitiveSet to provide a fallback for non support glDrawArray/glDrawElement modes.
Added finer gained error checking during StateSet::compile().
Moved the handling of DisplaySettings into Traits constructor.
Added support for s/getGLContextVersion(), s/getGLContextFlags() and s/getGLContextProfileMask() to osg::DisplaySettings.
Added command line and env var support for setting the GLContextVersion, GLContextFlags and GLContextProfileMask to osg::DisplaySettings.
Removed EXT postfix of FrameBufferObject functions, and added support for checking non EXT versions frame buffer object GL functions.
Introduced usage of OSG_GL*_FEATURES to avoid some #if #else #endif code blocks.
Using a submissions from Paul Martz as a guide added perliminary GL3 support to a range of OSG classes
Added support for automatic aliasing of vertex, normal, color etc. arrays to Vertex Attribute equivelants.
Added new osg::GLBeginEndAdapter class for runtime conversion from glBegin/glEnd codes to vertex arrray equivelants.
Added automatic shader source conversion from gl_ to osg_ builtins.
checking _gl_get_query_objectuiv_arb which is a different function
pointer. Changed to check the function being called.
This is against the subversion trunk."
The Texture Pool can be enabled by setting the env var OSG_TEXTURE_POOL_SIZE=size_in_bytes.
Note, setting a size of 1 will result in the TexturePool allocating the minimum number of
textures it can without having to reuse TextureObjects from within the same frame.
Note from Robert Osfield, I've temporarily re-enabled the old focing of of color and depth attachment to avoid regressions on some OpenGL driver. We'll revist this once
we have a mechanism for controlling this override at runtime.
#define FORCE_COLOR_ATTACHMENT 1
#define FORCE_DEPTH_ATTACHMENT 1
Therefore I have changed all the occurances of atof by asciiToFloat or
asciiToDouble.
I believe that it is safe to do so at least for all the plugins.
Included here are also asciiToFloat conversion of environment variables. One
might argue that these should be locale dependent. But IMO these should be
set and interpreted by osg independent of the current locale.
"
- osg::Texture sets GL_MAX_TEXTURE_LEVEL if image uses fewer mipmaps than
number from computeNumberOfMipmaps (and it works!)
- DDS fix to read only available mipmaps
- DDS fixes to read / save 3D textures with mipmaps ( packing == 1 is
required)
- Few cosmetic DDS modifications and comments to make code cleaner (I hope)
Added _isTextureMaxLevelSupported variable to texture extensions. It
could be removed if OSG requires OpenGL version 1.2 by default.
Added simple ComputeImageSizeInBytes function in DDSReaderWrites. In
my opinion it would be better if similar static method was defined for
Image. Then it could be used not only in DDS but other modules as well (I
noticed that Texture/Texture2D do similar computations).
Also attached is an example test.osg model with DDS without last mipmaps to
demonstrate the problem. When loaded into Viewer with current code and moved
far away, so that cube occupies 4 pixels, cube becomes red due to the issue
I described in earlier post. When you patch DDS reader writer with attched
code but no osg::Texture yet, cube becomes blank (at least on my
Windows/NVidia) When you also merge osg::Texture patch cube will look right
and mipmaps will be correct."
database options like recently added to PagedLOD.
Also there is a change to the traverse method:
The previous ProxyNode checks the VisitorType to be a CULL_VISITOR and the
presence of a request handler to submit a database request.
In contrast to that PagedLOD uses the request handler if it is there - even if
the visitor type is not a cull visitor.
The change removes the cull visitor test from the ProxyNode so that it behaves
like the PagedLOD.
I believe that the presence of a request handler in a visitor might be
sufficient to trigger the requests as this is done in the PagedLOD anyway.
Based on rev 10332."
GL_EXT_framebuffer_blit extension is available. This is due to the early
return from the constructor if this is missing.
As far as I read the standard extension documents, this blit call is not
required to have packed depth stencil fbos.
The change fixes this and allows packed stencil attachments on machines
without the multisample blit command."
constructor issues:
Within the NullStream & NotifyStream classes the base class objects
(ostream) were being initialised before the class member _buffer objects
which they referenced, causing a crash.
I had to move the file scope g_NullStream & g_NotifyStream initialisation
into a function to get them to work.
Also there was a missing osg:: qualifier on getNotifyHandler."
Two implementations of NotifyHandler are currently available:
- StandardNotifyHandler, calls fputs(message, stderr) for severity <= WARN and fputs(message, stdout) for severity > WARN
- WinDebugNotifyHandler, windows users can redirect notifications to windows debug output, notifications can be viewed in output window of the debugger i.e. MSVC or DebugView (http://technet.microsoft.com/en-us/sysinternals/bb896647.aspx) (see screenshot).
I have seen on osg-users that some people do std::cerr.rdbuf(otherStream.rdbuf()) to redirect notifications. This trick will no longer work since osg::notify() returns internal osg::NotifyStream not std::cout or std::cerr. You can use osg::notify().rdbuf(otherStream.rdbuf()) to do this instead.
Additionally I've made some minor fixes:
- Minor imrovements to osg::notify documentation
- NullStream could crash by deleting stream buffer other than default NullStreamBuffer in the destructor i.e. after osg::notify(osg::DEBUG_FP).rdbuf(otherStream.rdbuf())"
Introduced a new callback osgDB::FindFileCallback that overrides the default behavior of findDataFile/findLibraryFile.
Introduced support for assigning ReaderWriter::Options directory to PagedLOD.
Introduced new osgDB::FileLocationCallback for assistancing the DatabasePager to know when a file is hosted on a local or remote file system.
Introduced a new FindFileCallback to Registry to compliement the existing ReadFileCallback and WriteFileCallback.
Added support for assign Find, Read and WriteFileCallbacks to osdDB::Options to enable plugins/applications to override the callbacks just for that
read/write call and any nested file operations
I have attached a fixed file where the traits are checked in the PrimitiveShapeFunctor where appropriate. They are checked for Box, Cone Capsule and Cylinder. These just mirror the checks that were already done in the DrawShapeVisitor.
(another instance where if the ShapeDrawable had just been osg::Geometry, there wouldn't have been a problem... :-) )
I also fixed a small typo in the file in two places ("implementated" --> "implemented")."
-Added copying of shaders and attributes in osg::Program copy constructor.
-Changed StateSet::compare function to compare Uniforms and their
override values. Previously it compared a RefUniformPair."
Also, there was also a small bug in osgDB's CMakeLists.txt that was causing an error when I tested with CMake 2.4.4.
IF(${OSG_DEFAULT_IMAGE_PLUGIN_FOR_OSX} STREQUAL "quicktime")
was changed to
IF(OSG_DEFAULT_IMAGE_PLUGIN_FOR_OSX STREQUAL "quicktime")
"
Node::Node(Node &node, copyop) :
_stateSet(copyop(node.getStateSet()),
It doesn't call the setStateSet method of osg::Node (or osg::Drawable). So the parent
list of the state set is not updated with the new node (drawable)."
On destruction of some static variables, the global referenced mutex is used
to lock access to the parent lists of state attributes, nodes and so on.
This even happens past the mutex is already destroyed.
This change to Referenced.cpp revision 9851 uses the same technique like the
DeleteHandlerPointer already in Referenced.cpp to return an zero pointer for
the global referenced lock if it is already destroyed."
DisplaySettings.cpp: OSG_COMPIlE_CONTEXTS -> OSG_COMPILE_CONTEXTS
AnimtkViewer.cpp: is a 3d poker game client -> is an example for viewing
osgAnimation animations"
It\u2019s a one line change against OSG 2.8.0 (see line 196). I\u2019ve already tested the change, and confirmed it\u2019s fixing the crashes described above."
used but never restored to the decimal notation. That made OSG print messages
like the following after some notifications:
Warning: detected OpenGL error 'invalid value' after RenderBin::draw(,)
RenderStage::drawInner(,) FBO status= 0x8cd5
[...]
Scaling image 'brick_side.JPG' from (1b4,24f) to (200,200) <--- Values in hex
because of previous error.
[...]"
circumstances under which this bug occur are rather specific, but the
basic problem occurs when one translation unit other than libosg.so
constructs an object that is a subclass of osg::Shape and another
translation unit other than libosg.so tries to perform a dynamic_cast or
other RTTI-based operation on that object. Under these circumstances,
the RTTI operation will fail. In my case, the translation units involved
were an application and osgdb_ive.so. The application constructed a
scene graph that included instantiations of subclasses of osg::Shape.
Depending on how the user ran the application, it would write the scene
graph to an IVE file using osgDB::writeNodeFile(). The dynamic_cast
operations in DataOutputStream::writeShape() would fail on the first
subclass of osg::Shape that was encountered. This is because there were
two different RTTI data objects for all osg::Shape subclasses being
compared: one in the application and one in osgdb_ive.so.
The fix for this is simple. We must ensure that at least one member
function of each of the subclasses of the polymorphic type osg::Shape is
compiled into libosg.so so that there is exactly one RTTI object for
that type in libosg.so. Then, all code linking against libosg.so will
use that single RTTI object. The following message from a list archive
sort of explains the issue and the solution:
http://aspn.activestate.com/ASPN/Mail/Message/1688156
While the posting has to do with Boost.Python, the problem applies to
C++ libraries in general."
The fix was to convert the osg::State to use C pointers for the set of applied PerContexProgram objects, and use the osg::Oberver mechanism to avoid dangling pointers for being maintained in osg::State.
methods
getProjectionMatrixAsOrtho()
getProjectionMatrixAsFrustum()
getProjectionMatrixAsPerspective()
getViewMatrixAsLookAt() (2x)
are now const, as they only call const methods of osg::Matrixf/d.
"
returns true if (the extension string is supported or GL version is greater than or equal to a specified version) and
non extension disable is used. This makes it possible to disable extensions that are now
available as parts of the core OpenGL spec.
Updated Texture.cpp is use this method.
VertexBufferObject::compileBuffer().
The offsets of newly added Arrays were not properly
calculated. This submission tries to find a
matching empty slot when the total size of
the VBO has not changed (e.g. when an array
is replaced by another array of the same size).
This fixes the overwriting issue that I showed in my posting
"Bug in VertexBufferObject::compileBuffer" on OSG-Users.
"
unsigned int Image::computeNumComponents(GLenum pixelFormat)
so I added these types to the switch statement:
case(GL_RED_INTEGER_EXT): return 1;
case(GL_GREEN_INTEGER_EXT): return 1;
case(GL_BLUE_INTEGER_EXT): return 1;
case(GL_ALPHA_INTEGER_EXT): return 1;
case(GL_RGB_INTEGER_EXT): return 3;
case(GL_RGBA_INTEGER_EXT): return 4;
case(GL_BGR_INTEGER_EXT): return 3;
case(GL_BGRA_INTEGER_EXT): return 4;
case(GL_LUMINANCE_INTEGER_EXT): return 1;
case(GL_LUMINANCE_ALPHA_INTEGER_EXT): return 2;
That's all... now it computes the number of components and, thus, the image size
correctly."
--------------------------------------------
Descripton:
The patch does provide a new class PixelDataBufferObject which is capable of allocating memory on the GPU side (PBO memory) of arbitrary size. The memory can then further be used to be enabled into read mode (GL_PIXEL_UNPACK_BUFFER_ARB) or in write mode (GL_PIXEL_PACK_BUFFER_ARB). Enabling the buffer into write mode will force the driver to write data from bounded textures into that buffer (i.e. glGetTexImage). Using buffer in read mode give you the possibility to read data from the buffer into a texture with e.g. glTexSubImage or other instuctions. Hence no data is copied over the CPU (host memory), all the operations are done in the GPU memory.
--------------------------------------------
Compatibility:
The new class require the unbindBuffer method from the base class BufferObject to be virtual, which shouldn't break any functionality of already existing classes. Except of this the new class is fully orthogonal to existing one, hence can be safely added into already existing osg system.
--------------------------------------------
Testing:
The new class was tested in the current svn version of osgPPU. I am using the new class to copy data from textures into the PBO and hence provide them to CUDA kernels. Also reading the results back from CUDA is implemented using the provided patch. The given patch gives a possibility of easy interoperability between CUDA and osg (osgPPU ;) )
--------------------------------------------
I think in general it is a better way to derive the PixelBufferObject class from PixelDataBufferObject, since the second one is a generalization of the first one. However this could break the current functionality, hence I haven't implemented it in such a way. However I would push that on a stack of wished osg 3.x features, since this will reflect the OpenGL PBO functionality through the classes better.
"
"To reproduce the bug:
1. Create a template osg::Sequence node (and underlying geometry) but do not attach the node to the current active scenegraph.
2. At some point during the rendering loop (perhaps on a keystroke) clone the sequence node (I use the call:
dynamic_cast<osg::Node*>(templateNode -> clone( osg::CopyOp( (osg::CopyOp::CopyFlags)osg::CopyOp::DEEP_COPY_NODES ) ) )
3. Set the cloned sequence node duration to a value that makes the animation run slower (i.e. 2.0).
4. Start the cloned sequence (using setMode()).
5. Repeat steps 2 \u2013 4 and observe that the cloned sequences do not run slow but run as fast, appearing to ignore the duration that has been set on them.
Looking at the \u2018good documentation\u2019 (2.4 source code), I see that _start is being set to _now (osg::Sequence::setMode(), line 192). Should this not _start not be set to -1.0?"
* When used PDS RenderStage::runCameraSetUp sets flag that FBO has already stencil,depth buffer attached. Prevents adding next depth buffer.
* Sets correct traits for p-buffer if used PDS and something goes wrong with FBO setup or p-buffer is used directly.
* Adds warning to camera if user add depth/stencil already attached through PDS.
* Sets blitMask when use blit to resolve buffer.
There is also new example with using multisampled FBO."
I have not reverted added Compiler options. I assume that one may want to have warnings enabled for the application but may not want to see them while OSG libraries and examples compile.
Modified files:
osg/Export - now explicitly includes osg/Config to make sure OSG_DISABLE_MSVC_WARNINGS is read
osg/Config.in - declares OSG_DISABLE_MSVC_WARNINGS flag to be added to autogenerated osg/Config
CMakeLists.txt - declares OSG_DISABLE_MSVC_WARNINGS as option with default ON setting
"
The fix was to check whether glGetString( GL_VERSION ) returned a null pointer (Ref. svn diff below). The altered src/osg/GLExtensions.cpp is zipped and attached to this email."
set.
The optimization is based on the observation that matrix matrix multiplication
with a dense matrix 4x4 is 4^3 Operations whereas multiplication with a
transform, or scale matrix is only 4^2 operations. Which is a gain of a
*FACTOR*4* for these special cases.
The change implements these special cases, provides a unit test for these
implementation and converts uses of the expensiver dense matrix matrix
routine with the specialized versions.
Depending on the transform nodes in the scenegraph this change gives a
noticable improovement.
For example the osgforest code using the MatrixTransform is about 20% slower
than the same codepath using the PositionAttitudeTransform instead of the
MatrixTransform with this patch applied.
If I remember right, the sse type optimizations did *not* provide a factor 4
improovement. Also these changes are totally independent of any cpu or
instruction set architecture. So I would prefer to have this current kind of
change instead of some hand coded and cpu dependent assembly stuff. If we
need that hand tuned stuff, these can go on top of this changes which must
provide than hand optimized additional variants for the specialized versions
to give a even better result in the end.
An other change included here is a change to rotation matrix from quaterion
code. There is a sqrt call which couold be optimized away. Since we divide in
effect by sqrt(length)*sqrt(length) which is just length ...
"
they are written to disk, either inline or as an external file. Added support for
this in the .ive plugin. Default of WriteHint is NO_PREFERNCE, in which case it's
up to the reader/writer to decide.
occlusion query result is not ready for retrieval when the app tries to
retrieve it. This fix adds an application-level wait loop to ensure the
result is ready for retrieval. This code is not compiled by default; add "-D
FORCE_QUERY_RESULT_AVAILABLE_BEFORE_RETRIEVAL" to get this code.
Full, gory details, to the best of my recollection:
The conditions under which we encountered this issue are as follows: 64-bit
processor, Mac/Linux OS, multiple NVIDIA GPUs, multiple concurrent draw
threads, VRJuggler/SceneView-based viewer, and a scene graph containing
OcclusionQueryNodes. Todd wrote a small test program that produces an almost
instant crash in this environment. We verified the crash does not occur in a
similar environment with a 32-bit processor, but we have not yet tested on
Windows and have not yet tested with osgViewer.
The OpenGL spec states clearly that, if an occlusion query result is not yet
ready, an app can go ahead and attempt to retrieve it, and OpenGL will
simply block until the result is ready. Indeed, this is how
OcclusionQueryNode is written, and this has worked fine on several platforms
and configurations until Todd's test program.
By trial and error and dumb luck, we were able to workaround the crash by
inserting a wait loop that forces the app to only retrieve the query after
OpenGL says it is available. As this should not be required (OpenGL should
do this implicitly, and more efficiently), the wait loop code is not
compiled by default. Developers requiring this work around must explicitly
add "-D FORCE_QUERY_RESULT_AVAILABLE_BEFORE_RETRIEVAL" to the compile
options to include the wait loop."
New attribute DatabasePager::_expiryFrames sets number of frames a PagedLOD child is kept in memory. The attribute is set with DatabasePager::setExpiryFrames method or OSG_EXPIRY_FRAMES environmental variable.
New attribute PagedLOD::PerRangeData::_
frameNumber contains frame number of last cull traversal.
Children of PagedLOD are expired when time _AND_ number of frames since last cull traversal exceed OSG_EXPIRY_DELAY _AND_ OSG_EXPIRY_FRAMES respectively. By default OSG_EXPIRY_FRAMES = 1 which means that nodes from last cull/rendering
traversal will not be expired even if last cull time exceeds OSG_EXPIRY_DELAY. Setting OSG_EXPIRY_FRAMES = 0 revokes previous behaviour of PagedLOD.
Setting OSG_EXPIRY_FRAMES > 0 fixes problems of children reloading in lazy rendering applications. Required behaviour is achieved by manipulating OSG_EXPIRY_DELAY and OSG_EXPIRY_FRAMES together.
Two interface changes are made:
DatabasePager::updateSceneGraph(double currentFrameTime) is replaced by DatabasePager::updateSceneGraph(const osg::FrameStamp &frameStamp). The previous method is in #if 0 clause in the header file. Robert, decide if You want to include it.
PagedLOD::removeExpiredChildren(double expiryTime, NodeList &removedChildren) is deprecated (warning is printed), when subclassing use PagedLOD::removeExpiredChildren(double expiryTime, int expiryFrame, NodeList &removedChildren) instead. "
/users/mvalle/OSG/OpenSceneGraph/src/osg/BufferObject.cpp: In member function `virtual void osg::ElementBufferObject::compileBuffer(osg::State&) const':
/users/mvalle/OSG/OpenSceneGraph/src/osg/BufferObject.cpp:600: warning: cast to pointer from integer of different size"
MemoryBarrier() is used in the implementation, so it should be checked.
This in effect disables the faster atomic ops on msvc 7.1 and older, even if
only the MemoryBarrier() call is missing. But it ensures for the fist cut
that it will build everywhere. If somebody cares for msvc 7.1 enough and has
one for testing installed, he might provide the apropriate defines to guard
that MemoryBarrier() call.
I tested that msvc8 32/64bit still passes the configure tests and compiles.
"
implementation of the atomic increment and decrement into a implementation
file.
This way inlining and compiler optimization can no longer happen for these
implementations, but it fixes compilation on win32 msvc targets. I expect
that this is still faster than with with mutexes.
Also the i386 gcc target gets atomic operations with this patch. By using an
implementation file we can guarantee that we have the right compiler flags
available."
via StateSet::setNestedRenderBin(bool) whether the new RenderBin should be nested
with the existing RenderBin, or be nested with the enclosing RenderStage.
I added a call to allocateDataArray() if rhs has (at least) one valid array, which should allocate the right array according to the type. Since the type was copied from rhs, it should create the same array as rhs has, so then it should copy the data in the following lines.
"
is based on suggested fix from Marco for fixing a crash due to lack of
thread safety in std::ofstream("/dev/null"); The fix is to use a custom stream
buffer that just discards all data. The implementation is also twice as fast
as the old /dev/null based approach.
nvidia 6 months ago and the issue is still "in progress". I've given up
waiting for them!
Platform - various Intel Windows XP SP2 PCs with various nvidia cards
including GeForce 8800 GTS and Quadro FX 4500, and various driver
versions including the latest WHQL 175.16.
I investigated your concerns about glGenerateMipmapEXT being slower than
GL_GENERATE_MIPMAP_SGIS, and for power-of-two textures, to my surprise
it is. For a 512*512 texture, glGenerateMipmapEXT takes on average 10ms,
while GL_GENERATE_MIPMAP_SGIS takes on average 6ms. Therefore I have
modified the code to only use glGenerateMipmapEXT if the texture has a
non-power-of-two width or height. I am resubmitting all the files
previously submitted (only "Texture.cpp" has significant changes since
my previous submission, I've also replaced tabs with spaces in
"Texture").
"
GL_GENERATE_MIPMAP_SGIS is very slow (over half a second for a 720*576
texture). However, glGenerateMipmapEXT() performs well (16ms for the
same texture), so I have modified the attached files to use
Texture::generateMipmap() if glGenerateMipmapEXT is supported, instead
of enabling & disabling GL_GENERATE_MIPMAP_SGIS."
Notes, from Robert Osfield, I've tested the out of the previous path using
GL_GENERATE_MIPMAP_SGIS and non power of two textures on NVidia 7800GT and
Nvidia linux drivers with the image size 720x576 and only get compile times
of 56ms, so the above half second speed looks to be a driver bug. With
Muchael's changes the cost goes done to less than 5ms, so it's certainly
an effective change, even given that Michael's poor expereiences with
GL_GENERATE_MIP_SGIS do look to be a driver bug.
