[idefixs]

Thanks you for your advice. All CPU intensive operations are multi-threaded (i.e. 4 threads on my 4-core CPU). I will be looking into OpenGL optimisations.
Here's the code that matters:

Code:

    glEnable (GL_BLEND);
    glEnable (GL_TEXTURE_2D);
    glDisable (GL_DEPTH_TEST);

    glMatrixMode (GL_PROJECTION);
    glPushMatrix();
    glLoadIdentity();
    glOrtho (-1, 1, -1, 1, -1, 1);

    glMatrixMode (GL_MODELVIEW);
    glPushMatrix();
    glLoadIdentity();
    
    glPolygonMode (GL_FRONT, GL_FILL);

    glBindTexture (GL_TEXTURE_2D, fishbmc_texture);

    // get a lock on the texture pixels
    while (!screenbuffer->lock())
        usleep (1);
    
    // update texture
    glTexSubImage2D (GL_TEXTURE_2D, 0, 0, 0, configuration->width(), configuration->height(), GL_RGBA, GL_UNSIGNED_BYTE, pixels);
    
    screenbuffer->unlock();

    glBegin (GL_QUADS);
    glTexCoord2f (fishbmc_left, 0.0f);     glVertex3f (-1.0f, -1.0f, 1.0f);
    glTexCoord2f (fishbmc_right, 0.0f);    glVertex3f (1.0f, -1.0f, 1.0f);
    glTexCoord2f (fishbmc_right, 1.0f);    glVertex3f (1.0f, 1.0f, 1.0f);
    glTexCoord2f (fishbmc_left, 1.0f);    glVertex3f (-1.0f, 1.0f, 1.0f);
    glEnd();

    glPopMatrix();
    glMatrixMode (GL_PROJECTION);
    glPopMatrix();