Sign in to follow this  
madgallagher

OpenGL Strange Video Memory Increase

Recommended Posts

Hi Guys

 

I have a weird issue on opengl. I am rendering a lot of tris (millions) using vertex arrays.

Here is the issue.

 

If I load up the tris and display as shaded with a constant color, the video memory is showing around 600MB.

 

If I then change the color array values to be varied at different vertex, the memory usage increases to around 1GB.

If I then spin the model around, the memory usage slowly decreases to around 600MB (after a couple of minutes).

 

But, if I load up the model straight away with varied shading, the memory is only 600MB.

 

This is on linux, so is it a driver issue ? Or is there something more obvious I am neglecting.

 

I am using nvidia-smi to check the graphics memory usage.

 

Cheers in advance for any help !

Share this post


Link to post
Share on other sites

My first question is "Are you using the Linux drivers from your graphics card manufacturer or Nouveau?" If you are Nouveau I recommend you get the official Linux drivers for you video card before moving on.

 

Once thats all set and done and it still doesn't work, they my next question is "Are you using any type of culling?" If you are, its possible that when switching from the constant color to the varied shading the culling gets reset and isn't re-called until you move the camera (spin the model) or load the model right away with the varied shading. You'll need to find way to make sure that the culler is "always active". If you are attempting to use deferred shading rendering, there could also be some issues there, but I wouldn't be able to help you much as I'm just starting to learn about using defShading with OpenGL.

 

Also, have you tried testing it with a lower-poly model (maybe in the thousands of tris)?

 

Lastly, what version of OpenGL are you using and what is your graphics card?

Share this post


Link to post
Share on other sites

Hi

 

I am using the latest linux driver from Nvidia. The card is pretty old. Its an FX3800, but we see the same issue on an FX4000.

The same routines are used for both types of shading. I'm simply changing the values in the color array. Nothing else.

 

We get the same "doubling" of memory usage on smaller models too. Its like when the color array is updated, the memory

is not re-allocated on the card properly.

Share this post


Link to post
Share on other sites

I'm simply changing the values in the color array.

How are you doing this? Is it a VBO or a client-side vertex array? What hints was it created with?

 

the memory usage slowly decreases to around 600MB (after a couple of minutes).

Is this actually a problem? If the driver doesn't need that memory for other tasks right now, they it's ok for it to delay releasing it.

Edited by Hodgman

Share this post


Link to post
Share on other sites

OK, the render code is basically as follows:

 

glShadeModel(GL_SMOOTH);

glEnable(GL_COLOR_MATERIAL)

glColorMaterial(GL_FRONT_AND_BACK,GL_AMBIENT);

glColorMaterial(GL_FRONT_AND_BACK,GL_SPECULAR);

glColorMaterial(GL_FRONT_AND_BACK,GL_DIFFUSE);

 

glEnableClientState(GL_VERTEX_ARRAY);

glVertexPointer(....);

glEnableClientState(GL_NORMAL_ARRAY);

glNormalPointer(...);

glEnableClientState(GL_COLOR_ARRAY);

glColorPointer(....);

glDrawElements(........);

 

So, I am just using arrays, not VBOs. All I do is update the values in the color array that glColorPointer points to before I come into this routine.

 

The problem is, on a 1GB graphics card, if I am using 600MB and then change the color, it swamps the card and

the system starts badly lagging because it is trying to double the memory usage.

Share this post


Link to post
Share on other sites

By you use of glColorPointer() and the FX3000/4000, I'm guessing your using OpenGL 2.0. I personally don't have too much experience with anything earlier than OpenGL 3.1, so please take my help as a grain of salt.

 

Do you disable client state after you are finished writing to it or at program close? If you use the later, try disabling it after you are finished drawing everything. I'm not sure weather or not it will help, but its worth a shot in my opinion.

 

Is it possible for you to use VBOs rather than arrays? My understanding is that even back in OpenGL 2, it was more efficient to use buffers rather than arrays.

Share this post


Link to post
Share on other sites

Yes, sorry I forgot to add in the glDisableClientState commands. These are all deleted after the tris are drawn each time.

 

I added in VBOs and got the same effect even when deleting the VBO. Even simply using glBegin/End for the rendering (shock horror !!) causes the same effect.

 

It is kind of driving me crazy..................

Share this post


Link to post
Share on other sites

It's driving me crazy that I don't know how else I can help you...blink.png

 

Just one quick question... Are you using C++ or C or another language? Also, if you are using C++ or C, are you using any libraries like GLEW or SDL?

Share this post


Link to post
Share on other sites
I suspect that this may be normal behaviour rather than a driver issue. The fact that you're seeing this kind of video RAM usage at all with client arrays indicates that the geometry is being cached in video RAM. So you specify the arrays and draw, all good. Then you change the color array and now you've got two copies cached - the original (which the driver is presumably keeping around in case you need to go back to it) and the new. After a while, if you haven't gone back to the original, the driver decides to throw out it's old copy.

You may be able to control some of this behaviour with some use of GL_EXT_compiled_vertex_array, or you may be better off not using a vertex array for your constant color case (just glColor4f it instead).

Share this post


Link to post
Share on other sites

Thanks, I'll check out the GL_EXT_compiled _vertex_array. I get the same problem when I have varied colours

at the vertex and just change the values to say RGB ->RBG. Basically, any change to the colour array causes

the problem.

 

I appreciate all your comments gentlemen.

Share this post


Link to post
Share on other sites

mhagain, on 28 Mar 2013 - 12:50, said:
I suspect that this may be normal behaviour rather than a driver issue. The fact that you're seeing this kind of video RAM usage at all with client arrays indicates that the geometry is being cached in video RAM. So you specify the arrays and draw, all good. Then you change the color array and now you've got two copies cached - the original (which the driver is presumably keeping around in case you need to go back to it) and the new. After a while, if you haven't gone back to the original, the driver decides to throw out it's old copy.

You may be able to control some of this behaviour with some use of GL_EXT_compiled_vertex_array, or you may be better off not using a vertex array for your constant color case (just glColor4f it instead).

Pretty much this. For the most part, the driver knows best. If your changing values in your submitted data, it's highly likely the driver is simply caching the data, without releasing the old buffer immediately.

This is not uncommon to see in windows either. In short, i'd recommend not worrying about what the driver is doing, unless it's actually causing a problem.

Share this post


Link to post
Share on other sites

Well I think it is a problem for him as he said that when he changes the value, his graphics card's memory usage shoots up from 600M to 1G and then it starts to lag. If the driver is caching the data, then he would need to find a way to remove the values he's editing from the cache.

Share this post


Link to post
Share on other sites
The problem is in the "idea" to store 600MB in a single vertex array and transfer it in each draw call. That is not the way to go. I guess the performace is terrible if the data is not cached. The driver is probably trying to save performance by caching data and that's the problem. It could be solved by making smaller buffers.

Share this post


Link to post
Share on other sites
The only real way to handle this that I can think of is to use two different fragment shaders, one with the constant colour as a uniform. From the looks of the OP's sample code he's not using shaders at all, but running on prehistoric hardware shouldnt be a problem here; the kind of hardware that can handle this volume of data will always have shaders available anyway.

Share this post


Link to post
Share on other sites

Well I think it is a problem for him as he said that when he changes the value, his graphics card's memory usage shoots up from 600M to 1G and then it starts to lag. If the driver is caching the data, then he would need to find a way to remove the values he's editing from the cache.

ah, i missed where he said it was lagging, my bad.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this  

  • Forum Statistics

    • Total Topics
      627744
    • Total Posts
      2978895
  • Similar Content

    • By DelicateTreeFrog
      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!
    • By JJCDeveloper
      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
    • By AyeRonTarpas
      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?  
    • By ferreiradaselva
      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. 
  • Popular Now