the only way to free some cpu time is the use of the shader program ,
shaders are running calcs on the rendering hardwer ( the graphic card ) and frees cpu time for some other operation , I played with glsl ( opengl shader api ) and it is powerfull thing , language is like C and you can quickly catch syntax ...

there is a lot of tutorials for glsl , just search for modern opengl glsl ,

http://ogldev.atspace.co.uk/index.html

http://www.arcsynthesis.org/gltut/index.html

http://openglbook.com/

http://www.opengl-tutorial.org/

this is editor , you can see your shader in real time ,really nice and powerfull :
http://www.geeks3d.com/glslhacker/

gl&hf

ps
this 2 library helped me a lot in fast app development , but in C , you'll have to port headers to fasm style ofc ...
http://www.glfw.org/
http://glew.sourceforge.net/

pps
this is all for opengl 3.* and higher ,
there isn't glvertex3* functions here , no need really , just do some tutorials and you will found why ...
http://www.opengl.org/sdk/docs/man3/

can you use the real keyboard scancodes to tell the keys to the program? then, i will have the qsdz control, and any keaybord mapping will have the same physical correspondance.

the game is interresting, i think there are too much explosions by default, we don't see the enemies.

tip: if you sit in a corner, it is easy to kill enemies


enemies can be cool in real 3D, and not only projected to a surface

Can you give a quick example of what this means? Are you talking about scanning the keyboard state each frame instead of reading them in the message loop?

Also, the graphics are continually being updated. You can see a screenshot of the most recent version on my first post. The particle counts for each enemy are given by WINDMILL_PAR_COUNT, DIAMOND_PAR_COUNT, ect in the source

i mean, the hardware keyboard layout is the same for everybody (qwerty, azerty, bepo, etc...)
but the soft layout changes (qwerty, azerty, bepo) and makes it difficult to use the alpha keys reliably for the moves.

in my case, the keys wasd are not disposed as a directional cross, the hardware scancodes are "geographic" and i think they are better suitable for the game dev than the character keys.

oh, i get it. What api functions would I use for that? I'm assume GetKeyboardState will not work because it 'copies the status of the 256 virtual keys to the specified buffer' and virtual keys are not what you are talking about...

http://www.youtube.com/watch?v=e1TCHfDE7jI

there i can see a very similar graphics thing on some things.. its yours?

too similiar... except the top bar got score, time, and highest score.

i wonder wich one is earlier. because some stuff can be disassemblered, but whatever. If it was stolen, then should report this. if the person made this and put it there then there is nothing else to do, if this is sthe stolen copy, then perhaps should reconsider allowing the person to handle moneys with common projects.

a dot will always be a dot, and the drawer of the dot will not change anything to that, a dot is a dot.

then, a geometry war 2 is a geometry war 2 and cannot be anything else.

maybe you can make a map with inverted colors... white instead of black, and so on.

about the keyboard, sorry, i don't have any solution, i am still a pure noob in win api. but i am sure there is a way to interpret keys by scancode. even in windows.

but really, your dot looks very similar to their dot... maybe they use your dot in fact sue them until you or they don't have more money, the world will be a little more balanced after that

but before (or after like me) saying, they copied your idea etc... can you juste read the title bar of this program? what do you see?

i see:

aGW2 clone.

the word clone here means, it is a copy of the original.

then, close this nonsense, we don't care who had the idea first, otherwise, stop breath now cause jesus and ramses 2 had the idea before you.

There was a person who posted this video (well i dunno) so this "person" have posted a video about another game wich seemed like many things were stolen from gta san andreas or such and some stuff only modified. plus some stuff added. the game is called saints row 4.
so perhaps stolen in some way? but u never know.. in reality making 3d stuff in high quality is rather old stuff and already used to fake some persons on videos, so perhaps such is the case recently. this is btw a very good reason never to pay for webcam strip porn things.

there are already websits exploiting such a thing. for example Chaturbate or similar seems very likely to be such.. or contain many such individduals at least. but well.

this actually confuse me a lot, as i just read it today, maybe i overlook this sentence from the beginning,


http://en.wikipedia.org/wiki/Geometry_Wars:_Retro_Evolved_2 (2008)

oh. i dont read lengthy topics, just look at pictures and some questions. fine then.

tthsqe wrote to me we can make some game so now i have a 2 person team. tthsqe have not yet added me on skype.

I almost have the graphics updated and all of the enemy types included.
What this game needs is some sound and some level types or AI.

Sound: I have no idea about this and have never created any sound files before.

AI: this is actually fairly simple. The game logic will have the current game state and simply needs to determine where and when and what types of enemies to spawn.

Currently the game just spawn enemies randomly, which is quite boring.

i'd like to join. i need to practice a little more win32+openGL with fasm, i have not a lot of time to give for this project, but why not just try to add some features like sound, IA, skins, menus.... do you manage the sources with a versionning repository?

i dont think an open forum post is the best place to discuss this game, because it is obviously intended to be open. Also would there be some money out of it or something?

[ Post removed by author. ]

well
i only recently started learning opengl so dont hold me on the terminology
(and some design details probably)

modern opengl does not use glBegin() and end
it's become a state machine and far more complex to set up

speaking for opengl 3.3
data from buffers gets processed in vertex shaders then fragment shaders
how you process data is up to you ofc
shaders are written in GLSL, that is basically math wrapped in C

first you need to make an opengl context, GLFW does that for you
GLFW also gives you a window and input and all you need for graphics
API is nice and can most all be done in callbacks, what is great

then you set up opengl capabilities like DEPTH_TEST

now comes the fun part
most all these things need to be generated (created in the state machine)
then bound (made current, that will be used in subsequent calls)
then filled with some data

it is actually a VERY flexible system, but this is how i did a simple picture viewer
after creating a window, setting up callbacks and stuff

generated and bound a vertex array that will hold buffers
glGenVertexArrays(1, &vao);
glBindVertexArray(vao);

made buffers for vertices (all 4 of them) and UV coordinates, then filled them
(this is just for vertices, its same for UV's and anything you want rly)
glGenBuffers(1, vbo);
glBindBuffer(GL_ARRAY_BUFFER, vbo[0]);
glBufferData(GL_ARRAY_BUFFER, 4 * sizeof(struct vec3), flat, GL_STATIC_DRAW);
glVertexAttribPointer((GLuint) 0, 3, GL_DOUBLE, GL_FALSE, sizeof(struct vec3), 0);
glEnableVertexAttribArray(0);

where "flat" is a pointer to my vertices
last 2 calls put the buffer object in the first place of a vertex array
afaik a buffer can be used directly easily, but like this i have 1 opengl call in rendering loop (shouldn't matter rly)

then you have to load shader sources to opengl, create shaders in the state engine, compile them, link them into a program and make them current to be used in rendering
also a program has to be crated too

vertexsource = filetobuf("vertexshader.vert");
fragmentsource = filetobuf("fragmentshader.frag");
vertexshader = glCreateShader(GL_VERTEX_SHADER);
fragmentshader = glCreateShader(GL_FRAGMENT_SHADER);
glShaderSource(vertexshader, 1, (const GLchar **)&vertexsource, 0);
glShaderSource(fragmentshader, 1, (const GLchar **)&fragmentsource, 0);
glCompileShader(vertexshader);
glCompileShader(fragmentshader);
shaderprogram = glCreateProgram();
glAttachShader(shaderprogram, vertexshader);
glAttachShader(shaderprogram, fragmentshader);
glLinkProgram(shaderprogram);
glUseProgram(shaderprogram);

fun yet ?

now also a sampler is needed to feed the pixel shader with.. well pixels
it takes a couple parameters too

glGenSamplers(1, &sampler);
glSamplerParameteri(sampler, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
glSamplerParameteri(sampler, GL_TEXTURE_MIN_FILTER, GL_LINEAR_MIPMAP_LINEAR);
glSamplerParameterf(sampler, GL_TEXTURE_MAX_ANISOTROPY_EXT, 16.0f);

this is not much use in your game, but you can give it a texture instead of just a color

glGenTextures(1, &texture);
glBindTexture(GL_TEXTURE_2D, texture);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, X, Y, 0, GL_RGB, GL_UNSIGNED_BYTE, picture);

last call tells the shader to use the RGB texture pointed to by "picture" (actually it uploads the "picture" to "texture" and it is used in the frag shader)

drawing itself is called simply with

glDrawArrays(GL_QUADS, 0, 4);

in my case its a QUAD made of 4 vertices
this "draws" whatever is stored in buffers associated with the current (bound) array

and ofc
glfwSwapBuffers(window);
to flip the buffer to screen

shaders are more or less unrelated programs that process the data sent to them
vertex shader sets the vertices in stone
my vertex shader takes the vertices and UV's, makes vertices... vertices using gl_Position
and passes UV's to the fragment shader directly
my vertices are 3 floats, but opengl wants 4 so it gets a "1.0" as fourth
also i dont apply no transformation matrix, since i do "2D" (there is no 2D in opengl afaik)

#version 330

in vec3 inPosition;
in vec2 vertexUV;

out vec2 UV;

void main()
{
gl_Position = vec4(inPosition, 1.0);
UV = vertexUV;
}

note the IN to OUT design

fragment shaders do more work giving pixes on screen their colors
one pixel at a time

#version 330

in vec2 UV;
out vec3 outputColor;

uniform sampler2D sampler;

void main()
{
outputColor = texture( sampler, UV ).rgb ;
}

it uses the sampler to color a pixel based on the position of a triangle and its UV

so ye its a lot more work to set up but it is powerful for complex scenes

if you want il post my whole code
just need to translate the comments to english

funny, this makes me think there is a simpler way to just put a picture on the screen

PS i assumed you know 3D math better then me so i left it out
i could probably learn about it from you
also a 2D picture viewer does not need it
but there is great things about matrices and such in online tutorials
i just didnt find any opengl 3.3 in C so i went on my own

also great game
inspires me to get back to opengl

I did use a bunch of immediate mode drawing. The problem that see with opengl is that I do not know exactly what the gpu does, nor do I know what the driver dies, so I am always reduced to a "top-down" approach of modifying others' code. Can we have a fragment shader sample from more than one texture? How would you set this up? Also, can we sample from a texture of floats (not bytes)?

edit: The shaders are making a bit more sense now.
I think I found a list of attributes at http://www.opengl.org/sdk/docs/tutorials/ClockworkCoders/attributes.php
so we just need to bind the texture to "channels" 0 through 7 and then we can access them with gl_MultiTexCoord0 through 7.