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:

Quote

There is a temptation to just have 1 shader that does all : 100 lights and textures and bump mapping that can be turned on and off at the flip of a boolean uniform. Boolean uniforms and for loops all over the place. Consider writing separate shaders for different cases. Consider writing a tool that auto-generates some of your shaders.

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!

Setting uniform variables is expensive, but I don't know how it compares to calling glUseProgram(). If you set 10 or so boolean uniform variables, then it would probably be slower than just calling glUseProgram() and not setting as many. Let's say you set 2 uniforms and then 10 boolean uniforms for a "master shader". If you broke them down into different shaders, then each would have to set the first 2 uniforms, and let's say they each set 1 more. Then you could run 3-4 different shaders in approximately the same amount of time (assuming run time for master shader is equal to the sum of the multiple smaller shaders, and I said 3-4 since you'd have to call glUseProgram more often.)

So, personally I'd break them apart, because there's no point in setting multiple uniform booleans to check for things that the shader isn't going to use. It would be like checking if you need to run, and maybe even setting up, a particle, light, and fur shader on something like a user interface. It just needs to know how to texture it and position it.

Storing the shaders in a file then having to load it (either during loading a model or rendering), then link it, then run it would be slower than loading them all at the start, linking them, and then just running them as needed. It might not actually be slower, but you're in the middle of loading a file, or rendering when you start reading the shader, and linking it, so it would increase the time it took to load the file or render. So its best to do it once at the start of the game, and if you need to reload it during runtime, add the ability to do so when a key is pressed.

50 minutes ago, Promit said:

Changing uniform parameters is cheaper than changing program, and this holds true for a very large number of uniforms. Once you've changed any uniform, changing more is all blocked into one operation and the incremental cost is very small. Uniform branching in a single shader is a perfectly reasonable way to develop, though I've had performance trouble with it on certain mobile platforms.

Now whether uniform branching in a single shader is a good design for readability and maintainability is up to you. The ideal case is probably to move to uniform buffer objects and swap the entire block of settings in one go. But it's certainly not inherently advantageous to have separate shaders.

Do you mean the initial setting of the data is cheaper than switching shaders, or just changing them after they've been set once? If you only meant that changing the uniform after its been set once, then I wasn't referring to that, just the initial setting of each uniform. It is better to not set variables that won't be used. Even if it doesn't cost much, you can still increase performance slightly by not setting them, this is the main problem I have with using one "master shader".

Thank you for your replies, guys! The idea of using one shader with uniforms is that the individually generated shaders would all just be copies of the same source code, but with the diffuse color, specularity, etc tweaked during load time based on the contents of the corresponding .obj materials. In this situation, "master shader" may be a bad name for it, since it really only contains the necessities that would apply to all objects (instead of being overloaded with features which aren't often used), with the objects tweaking the shader's uniforms for specularity (a value in the range of 0.0 to 1.0), diffuse color, etc based on the particular .obj material in use at the time, versus switching shaders altogether. Would the most benefits come out of doing it this way given these particular circumstances?