• Announcements

    • khawk

      Download the Game Design and Indie Game Marketing Freebook   07/19/17

      GameDev.net and CRC Press have teamed up to bring a free ebook of content curated from top titles published by CRC Press. The freebook, Practices of Game Design & Indie Game Marketing, includes chapters from The Art of Game Design: A Book of Lenses, A Practical Guide to Indie Game Marketing, and An Architectural Approach to Level Design. The GameDev.net FreeBook is relevant to game designers, developers, and those interested in learning more about the challenges in game development. We know game development can be a tough discipline and business, so we picked several chapters from CRC Press titles that we thought would be of interest to you, the GameDev.net audience, in your journey to design, develop, and market your next game. The free ebook is available through CRC Press by clicking here. The Curated Books The Art of Game Design: A Book of Lenses, Second Edition, by Jesse Schell Presents 100+ sets of questions, or different lenses, for viewing a game’s design, encompassing diverse fields such as psychology, architecture, music, film, software engineering, theme park design, mathematics, anthropology, and more. Written by one of the world's top game designers, this book describes the deepest and most fundamental principles of game design, demonstrating how tactics used in board, card, and athletic games also work in video games. It provides practical instruction on creating world-class games that will be played again and again. View it here. A Practical Guide to Indie Game Marketing, by Joel Dreskin Marketing is an essential but too frequently overlooked or minimized component of the release plan for indie games. A Practical Guide to Indie Game Marketing provides you with the tools needed to build visibility and sell your indie games. With special focus on those developers with small budgets and limited staff and resources, this book is packed with tangible recommendations and techniques that you can put to use immediately. As a seasoned professional of the indie game arena, author Joel Dreskin gives you insight into practical, real-world experiences of marketing numerous successful games and also provides stories of the failures. View it here. An Architectural Approach to Level Design This is one of the first books to integrate architectural and spatial design theory with the field of level design. The book presents architectural techniques and theories for level designers to use in their own work. It connects architecture and level design in different ways that address the practical elements of how designers construct space and the experiential elements of how and why humans interact with this space. Throughout the text, readers learn skills for spatial layout, evoking emotion through gamespaces, and creating better levels through architectural theory. View it here. Learn more and download the ebook by clicking here. Did you know? GameDev.net and CRC Press also recently teamed up to bring GDNet+ Members up to a 20% discount on all CRC Press books. Learn more about this and other benefits here.
Sign in to follow this  
Followers 0

glTexSubImage2D() performance

2 posts in this topic

I'm trying to implement a 2D GUI system for my OpenGL test program. I have the following definitions:

#define E_RMASK 0x00ff0000
#define E_RSHIFT 16
#define E_GMASK 0x0000ff00
#define E_GSHIFT 8
#define E_BMASK 0x000000ff
#define E_BSHIFT 0
#define E_AMASK 0xff000000
#define E_ASHIFT 24
#define E_CFMT GL_BGRA

typedef struct {
int32_t x;
int32_t y;
int32_t w;
int32_t h;
} Erect_t;

typedef struct {
int32_t w;
int32_t h;
uint8_t bpp;
uint32_t fmt;
void *d;

/* private members */
bool dirty;
Erect_t clip;
} Eimage_t;

typedef struct {
Eimage_t *img;
GLuint id;
} Etexture_t;

static inline uint32_t pixel_blend( uint32_t dst, uint32_t src )
uint32_t as = ( src & E_AMASK ) >> E_ASHIFT;
uint32_t ad = ( dst & E_AMASK ) >> E_ASHIFT;
uint32_t a = min( as + ad, 0x000000ff );

if ( !as ) return dst;

#define RGB_MASK ( ~E_AMASK )
#define RB_MASK ( E_RMASK | E_BMASK )
#define AG_MASK ( E_AMASK | E_GMASK )

uint32_t rb = ( ( ( src & RB_MASK ) * as ) + ( ( dst & RB_MASK ) * ( 0xff - as ) ) ) & AG_MASK;
uint32_t g = ( ( ( src & E_GMASK ) * as ) + ( ( dst & E_GMASK ) * ( 0xff - as ) ) ) & E_RMASK;

#undef RB_MASK
#undef AG_MASK

return ( ( a << E_ASHIFT ) & E_AMASK ) | ( ( ( rb | g ) >> 8 ) & RGB_MASK );

uint32_t engineMapColor( uint8_t r, uint8_t g, uint8_t b, uint8_t a )
return ( ( ( uint32_t )r & 0x000000ff ) << E_RSHIFT ) +
( ( ( uint32_t )g & 0x000000ff ) << E_GSHIFT ) +
( ( ( uint32_t )b & 0x000000ff ) << E_BSHIFT ) +
( ( ( uint32_t )a & 0x000000ff ) << E_ASHIFT );

Eimage_t *engineImageCreate( uint32_t w, uint32_t h )
Eimage_t *ret = ( Eimage_t * )calloc( 1, sizeof( Eimage_t ) );

ret->w = w;
ret->h = h;
ret->bpp = 4;
ret->fmt = engineState()->Cfmt;
ret->d = calloc( 4, w * h );
img->clip.x = 0;
img->clip.y = 0;
img->clip.w = img->w;
img->clip.h = img->h;

return ret;
void engineImageFillRect( Eimage_t *img, const Erect_t *r, uint32_t color )
int32_t start_x = min( ( int32_t )img->w, ( r ? max( r->x, img->clip.x ) : 0 ) );
int32_t start_y = min( ( int32_t )img->h, ( r ? max( r->y, img->clip.y ) : 0 ) );
int32_t end_x = ( r ? min( ( int32_t )img->clip.w, r->x + r->w ) : ( int32_t )img->clip.w );
int32_t end_y = ( r ? min( ( int32_t )img->clip.h, r->y + r->h ) : ( int32_t )img->clip.h );
uint32_t *data = ( uint32_t * )img->d + start_y * img->w;

for ( int32_t j = start_y; j < end_y; j++ ) {
int32_t i = end_x - start_x;
int32_t c = ( i + 7 ) / 8;
switch ( i % 8 ) {
case 0: do { { data[start_x + i] = pixel_blend( data[start_x + i], color ); i -= 1; }
case 7: { data[start_x + i] = pixel_blend( data[start_x + i], color ); i -= 1; }
case 6: { data[start_x + i] = pixel_blend( data[start_x + i], color ); i -= 1; }
case 5: { data[start_x + i] = pixel_blend( data[start_x + i], color ); i -= 1; }
case 4: { data[start_x + i] = pixel_blend( data[start_x + i], color ); i -= 1; }
case 3: { data[start_x + i] = pixel_blend( data[start_x + i], color ); i -= 1; }
case 2: { data[start_x + i] = pixel_blend( data[start_x + i], color ); i -= 1; }
case 1: { data[start_x + i] = pixel_blend( data[start_x + i], color ); i -= 1; } } while ( --c > 0 );
data += img->w;

img->dirty = true;

Etexture_t *engineTextureCreate( uint32_t w, uint32_t h )
Etexture_t *ret = ( Etexture_t * )calloc( 1, sizeof( Etexture_t ) );
ret->img = engineImageCreate( w, h );

glGenTextures( 1, &ret->id );
glBindTexture( GL_TEXTURE_2D, ret->id );
glTexImage2D( GL_TEXTURE_2D, 0, ret->img->bpp, ret->img->w, ret->img->h, 0, ret->img->fmt, GL_UNSIGNED_BYTE, ret->img->d );

return ret;

void engineTextureBind( Etexture_t *tex )
if ( tex ) {
glBindTexture( GL_TEXTURE_2D, tex->id );
if ( tex->img->dirty ) {
glTexSubImage2D( GL_TEXTURE_2D, 0, 0, 0, tex->img->w, tex->img->h, tex->img->fmt, GL_UNSIGNED_BYTE, tex->img->d );
tex->img->dirty = false;

Then I have (Estate_t is a structure holding various flags, the width/height of the window, etc. I'm not pasting it here since it's pretty big):
static void begin_2d_overlay( Estate_t *st )
glDisable( GL_DEPTH_TEST );
glDisable( GL_LIGHTING );
glEnable( GL_BLEND );
if ( st->wireframe )
glPolygonMode( GL_FRONT_AND_BACK, GL_FILL );

glMatrixMode( GL_PROJECTION );
glOrtho( 0.0, st->wm_width, st->wm_height, 0.0, -1, 1 );

glMatrixMode( GL_MODELVIEW );

glColor3f( 1.0, 1.0, 1.0 );

static void end_2d_overlay( Estate_t *st )
glEnable( GL_DEPTH_TEST );
glEnable( GL_LIGHTING );
glDisable( GL_BLEND );

glMatrixMode( GL_PROJECTION );

glMatrixMode( GL_MODELVIEW );

if ( st->wireframe )
glPolygonMode( GL_FRONT_AND_BACK, GL_LINE );

and my rendering function:

Etexture_t *overlay = engineTextureCreate( st->wm_width, st->wm_height );

static void render_function( Estate_t *st )
/* draw the 3D scene here */

begin_2d_overlay( st );

Erect_t r;
r.x = rand() % st->wm_width;
r.y = rand() % st->wm_height;
r.w = rand() % 300 + 1;
r.h = rand() % 300 + 1;
engineImageFillRect( overlay->img, &r, engineMapColor( rand() % 255, rand() % 255, rand() % 255, 128 ) );

overlay->img->dirty = false; /* this shouldn't be here, see below code block for explanation */

engineTextureBind( overlay );
glBegin( GL_QUADS );
glTexCoord2f( 0.0, 1.0 ); glVertex2f( 0.0, st->wm_height );
glTexCoord2f( 1.0, 1.0 ); glVertex2f( st->wm_width, st->wm_height );
glTexCoord2f( 1.0, 0.0 ); glVertex2f( st->wm_width, 0.0 );
glTexCoord2f( 0.0, 0.0 ); glVertex2f( 0.0, 0.0 );

end_2d_overlay( st );

/* more stuff like swapping OpenGL buffers, etc. */

Notice the "overlay->img->dirty = false;" line. When uncommented it causes engineTextureBind() to not call glTexSubImage2D() which in turn causes my application to jump from rendering ~360 frames per second to rendering ~3600 frames per second (yes, 10 times more). I only added that line to see if the slowdown is caused by my rectangle drawing code or something else. After multiple tries I managed to isolate glTexSubImage() as the cause of the slow-down.

My questions are:
(1) is this anywhere near the proper way to do a 2D GUI with OpenGL ?
(2) am I doing something wrong that causes glTexSubImage() to run that slow ?
(3) should I be using something else than glTexSubImage() to update the GUI ?

[Edited by - ileonte on November 21, 2010 4:29:57 AM]

Share this post

Link to post
Share on other sites
The difference between 360fps and 3600fps is only 2ms render time. That is not much. Don't use fps to measure performance, use seconds per frame instead.
1) No, you should do all the rendering in OpenGL. Currently you are doing a lot of computation on the CPU and then sending huge amount of pixel data to the GPU. I don't know at what resolution you are working but the situation will become worth on higher resolutions. At some point even the CPU might become the bottleneck.
2) I have not checked the code in details but using the format the GPU uses internally can improve performance. PBOs are also a good choice.
Fast Texture Transfers
3) Yes, render everything using OpenGL or use a library that does this, like Qt.

Share this post

Link to post
Share on other sites
The optimal glTex(Sub)Image2D parameters are:

internalformal: GL_RGBA
format: GL_BGRA
type: GL_UNSIGNED_INT_8_8_8_8_REV

Use these (and set up your source data to be happy with them) and you'll get good performance on all hardware with no stalls, no copying back to system memory to do the update, and likely no need to use a PBO (which may give you a lot of work for zero performance gain otherwise).

The reason why is that OpenGL stores your texture using 4 components irrespective of what you specify in your glTex(Sub)Image2D call - check the spec if you doubt me. Using any combination of parameters that doesn't match the real internal storage will mean that your driver will need to copy, expand, swizzle, fill in unspecified components and all the rest itself, and - in especially evil cases (Intel) - copy the texture data back to system memory.

People often choose suboptimal formats and internalformats such as GL_RGB (or GL_BGR) because they think it "saves memory" - it doesn't. OpenGL will always expand textures to 4 components, so not only does it not "save memory" but it also slows down your program.

A format of GL_BGRA (instead of GL_RGBA) can give up to a 6x increase in transfer speed on even recent NVIDIA hardware.

GL_UNSIGNED_INT_8_8_8_8_REV instead of GL_UNSIGNED_BYTE is optional but I recommend it as it can give a hint to some drivers that you're transferring 4-component texels and therefore put them into a more optimal special case path (Intel again).

Mandatory reading:
http://www.opengl.org/wiki/Common_Mistakes#Texture_upload_and_pixel_reads (all of it)
http://www.opengl.org/wiki/Common_Mistakes#Paletted_textures (the last sentence)

The best code structure for efficient use of glTexSubImage2D looks like this:
Identify what needs to be updated.
Update textures.
Draw the scene.
The worst looks like this:
Update subrect.
Draw stuff.
Update another subrect.
Draw stuff.
Update another subrect.
Draw stuff.
The latter will result in multiple pipeline stalls per frame as OpenGL will need to wait until you're finished drawing with the texture before it can update each next subrect.

Using these techniques I've had reports of an over 90x overall performance improvement from one user - not to be sniffed at.

[Edited by - mhagain on November 21, 2010 9:07:54 AM]

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  
Followers 0

  • Similar Content

    • By Solid_Spy
      Hello, I have been working on SH Irradiance map rendering, and I have been using a GLSL pixel shader to render SH irradiance to 2D irradiance maps for my static objects. I already have it working with 9 3D textures so far for the first 9 SH functions.
      In my GLSL shader, I have to send in 9 SH Coefficient 3D Texures that use RGBA8 as a pixel format. RGB being used for the coefficients for red, green, and blue, and the A for checking if the voxel is in use (for the 3D texture solidification shader to prevent bleeding).
      My problem is, I want to knock this number of textures down to something like 4 or 5. Getting even lower would be a godsend. This is because I eventually plan on adding more SH Coefficient 3D Textures for other parts of the game map (such as inside rooms, as opposed to the outside), to circumvent irradiance probe bleeding between rooms separated by walls. I don't want to reach the 32 texture limit too soon. Also, I figure that it would be a LOT faster.
      Is there a way I could, say, store 2 sets of SH Coefficients for 2 SH functions inside a texture with RGBA16 pixels? If so, how would I extract them from inside GLSL? Let me know if you have any suggestions ^^.
    • By KarimIO
      EDIT: I thought this was restricted to Attribute-Created GL contexts, but it isn't, so I rewrote the post.
      Hey guys, whenever I call SwapBuffers(hDC), I get a crash, and I get a "Too many posts were made to a semaphore." from Windows as I call SwapBuffers. What could be the cause of this?
      Update: No crash occurs if I don't draw, just clear and swap.
      static PIXELFORMATDESCRIPTOR pfd = // pfd Tells Windows How We Want Things To Be { sizeof(PIXELFORMATDESCRIPTOR), // Size Of This Pixel Format Descriptor 1, // Version Number PFD_DRAW_TO_WINDOW | // Format Must Support Window PFD_SUPPORT_OPENGL | // Format Must Support OpenGL PFD_DOUBLEBUFFER, // Must Support Double Buffering PFD_TYPE_RGBA, // Request An RGBA Format 32, // Select Our Color Depth 0, 0, 0, 0, 0, 0, // Color Bits Ignored 0, // No Alpha Buffer 0, // Shift Bit Ignored 0, // No Accumulation Buffer 0, 0, 0, 0, // Accumulation Bits Ignored 24, // 24Bit Z-Buffer (Depth Buffer) 0, // No Stencil Buffer 0, // No Auxiliary Buffer PFD_MAIN_PLANE, // Main Drawing Layer 0, // Reserved 0, 0, 0 // Layer Masks Ignored }; if (!(hDC = GetDC(windowHandle))) return false; unsigned int PixelFormat; if (!(PixelFormat = ChoosePixelFormat(hDC, &pfd))) return false; if (!SetPixelFormat(hDC, PixelFormat, &pfd)) return false; hRC = wglCreateContext(hDC); if (!hRC) { std::cout << "wglCreateContext Failed!\n"; return false; } if (wglMakeCurrent(hDC, hRC) == NULL) { std::cout << "Make Context Current Second Failed!\n"; return false; } ... // OGL Buffer Initialization glClear(GL_DEPTH_BUFFER_BIT | GL_COLOR_BUFFER_BIT); glBindVertexArray(vao); glUseProgram(myprogram); glDrawElements(GL_TRIANGLES, indexCount, GL_UNSIGNED_SHORT, (void *)indexStart); SwapBuffers(GetDC(window_handle));  
    • By Tchom
      Hey devs!
      I've been working on a OpenGL ES 2.0 android engine and I have begun implementing some simple (point) lighting. I had something fairly simple working, so I tried to get fancy and added color-tinting light. And it works great... with only one or two lights. Any more than that, the application drops about 15 frames per light added (my ideal is at least 4 or 5). I know implementing lighting is expensive, I just didn't think it was that expensive. I'm fairly new to the world of OpenGL and GLSL, so there is a good chance I've written some crappy shader code. If anyone had any feedback or tips on how I can optimize this code, please let me know.
      Vertex Shader
      uniform mat4 u_MVPMatrix; uniform mat4 u_MVMatrix; attribute vec4 a_Position; attribute vec3 a_Normal; attribute vec2 a_TexCoordinate; varying vec3 v_Position; varying vec3 v_Normal; varying vec2 v_TexCoordinate; void main() { v_Position = vec3(u_MVMatrix * a_Position); v_TexCoordinate = a_TexCoordinate; v_Normal = vec3(u_MVMatrix * vec4(a_Normal, 0.0)); gl_Position = u_MVPMatrix * a_Position; } Fragment Shader
      precision mediump float; uniform vec4 u_LightPos["+numLights+"]; uniform vec4 u_LightColours["+numLights+"]; uniform float u_LightPower["+numLights+"]; uniform sampler2D u_Texture; varying vec3 v_Position; varying vec3 v_Normal; varying vec2 v_TexCoordinate; void main() { gl_FragColor = (texture2D(u_Texture, v_TexCoordinate)); float diffuse = 0.0; vec4 colourSum = vec4(1.0); for (int i = 0; i < "+numLights+"; i++) { vec3 toPointLight = vec3(u_LightPos[i]); float distance = length(toPointLight - v_Position); vec3 lightVector = normalize(toPointLight - v_Position); float diffuseDiff = 0.0; // The diffuse difference contributed from current light diffuseDiff = max(dot(v_Normal, lightVector), 0.0); diffuseDiff = diffuseDiff * (1.0 / (1.0 + ((1.0-u_LightPower[i])* distance * distance))); //Determine attenuatio diffuse += diffuseDiff; gl_FragColor.rgb *= vec3(1.0) / ((vec3(1.0) + ((vec3(1.0) - vec3(u_LightColours[i]))*diffuseDiff))); //The expensive part } diffuse += 0.1; //Add ambient light gl_FragColor.rgb *= diffuse; } Am I making any rookie mistakes? Or am I just being unrealistic about what I can do? Thanks in advance
    • By yahiko00
      Not sure to post at the right place, if not, please forgive me...
      For a game project I am working on, I would like to implement a 2D starfield as a background.
      I do not want to deal with static tiles, since I plan to slowly animate the starfield. So, I am trying to figure out how to generate a random starfield for the entire map.
      I feel that using a uniform distribution for the stars will not do the trick. Instead I would like something similar to the screenshot below, taken from the game Star Wars: Empire At War (all credits to Lucasfilm, Disney, and so on...).

      Is there someone who could have an idea of a distribution which could result in such a starfield?
      Any insight would be appreciated
    • By afraidofdark
      I have just noticed that, in quake 3 and half - life, dynamic models are effected from light map. For example in dark areas, gun that player holds seems darker. How did they achieve this effect ? I can use image based lighting techniques however (Like placing an environment probe and using it for reflections and ambient lighting), this tech wasn't used in games back then, so there must be a simpler method to do this.
      Here is a link that shows how modern engines does it. Indirect Lighting Cache It would be nice if you know a paper that explains this technique. Can I apply this to quake 3' s light map generator and bsp format ?
  • Popular Now