You pass the address of a stack-allocated structure to pthread_create, this will be destroyed as soon as the main thread leaves init(), so probably before the loader thread has a chance to read the...

Nvidia called this "pseudo instancing", and showed some significant performance gains on old GPUs by using vertex attributes. Performance on DX10+ GPUs may be different.

AMD does it that way on the southern island chips according to their open ISA documents:



The GPU has some dedicated interpolation instructions for the fragment shader.

This will make all varyings behave as if they were declared "noperspective".

Also note that after viewport transformation the vertex coordinates are usually converted to fixed-point numbers with...

The internalFormat parameter (the third one) for glTexImage2D controls the format in which OpenGL stores the texture internally. Passing GL_R32F means GL should store the data a single-channel float...

opengl32.dll exports only the OpenGL 1.1 functions directly, all other functions have to be accessed via function pointers. The pointers can be retrieved by wglGetProcAddress.

This is obviously...

Most OpenGL versions are upward compatible, the only GL version which removed some APIs was the Version 3.1. However in Version 3.2 the ARB defined the concept of GL "profiles" and defined the "core"...

You only need to create a 1.1 context to get the max. supported compatibility version and a 3.1 context to get the max. supported core version via the GL_VERSION string. As you need the 1.1 context...

Another question: Has setting max_vertices = 4 (and still emitting only 3 vertices as before) an impact on the performance ?

The render-to-texture semantics of the R and RG formats are simple, but those of the luminance and intensity formats are not. And in a mostly programmable pipeline you can easily replace them with RG...

Maybe it's not a bad name per-se, but I think it is confusing when you read a math book on homogeneous coordinates and projective geometry to understand the math behind the OpenGL transforms.

I...

Well, to be more precise: gl_position contains the homogeneous coordinates of the already projected vertex, there's no further perspective transformation in the GL pipeline. The "perspective divide"...

Alfonse is of course right, I meant enabling depth clamp.

Actually OpenGL can deal with a literally infinite distant far plane already, it's just not possible to set up this projection via GLU (but then GLU isn't part of OpenGL), it doesn't even have a big...

The answer actually depends on the GL profile you use. In a core profile the glDrawArrays without bound vertex array should work, but in a compatibility profile the glDrawArrays call *requires* a...

You copy only half of your result matrix "res" back to parameter "resMat" (assuming sizeof(double) = 8, sizeof(float) = 4).

You should take a look at EGL_KHR_create_context, I think it covers everything glx can do.

Actually the fragment shader will be run also for some pixels outside the rendered triangle, because at least 2x2 pixel blocks are needed to compute derivates.

There was also a talk about this on the XDC 2012, you can find it on YouTube.

I see two problems:

In the call to glDrawElements you pass the offset of the end of your indexes, that should be the offset of the start of the indexes: (GLvoid *)0.

In the glBufferSubData...

The MVP transformation is not a rotation, so you can't represent it with a quaternion.

I don't know how specific cards implement scissor, but I think most implement the scissor test directly in the rasterizer. See...

The OpenGL Specs say that the scissor test comes after rasterization, but the OpenGL Specs only define behaviour, not implementation. That means a driver may implement the scissor test in any way...

While it can't interpret glIntercept logs, http://apitrace.github.com/ can intercept and retrace OpenGL and DirectX calls.

Do you use a GL_MIN_FILTER that requires mipmaps and you don't have the correct mipmaps for that texture ?