• Advertisement
Sign in to follow this  

[Solved] DirectX Load Time Question - 256 x 256 Map - no textures

This topic is 1967 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Edit: RulerOfNothing was correct, I was using too many calls to cout to check all my vertice/indice values.

Hey, I've been programming a while, but pretty new to directx/win32 programming and i've finally run into something I can't really find on google, so i've decided to create an account and ask you guys! (Yay! or groan maybe? haha)

Anyway, my question is basically this; I have a basic engine going using directx/c++ that'll basically let you input a width and it'll draw a square map with that width/height. At small values it takes no time at all to load, but if I start increasing it to 50+ it starts to take a while to load, at 100x100 it takes ~30 seconds, and at 256 x 256 it takes atleast a minute (maybe 2-3 - feels like forever!) to load.

This seems pretty excessive considering I have little else, just a sky box and low poly model with a basic texture on the screen.

I assume this isn't normal based off current games loading times, I don't know much about culling & I also have some basic lighting going on?

Do I just have a poor algorithm or is there some culling option or something else I can do to get the time down?

Here's the code that draws the vertices/indicies;

This basically calculates everything & puts it in arrays;


void loadMap(int width){

SecureZeroMemory(arrayx, sizeof(arrayx));
SecureZeroMemory(arrayy, sizeof(arrayy));
SecureZeroMemory(arrayz, sizeof(arrayz));
SecureZeroMemory(indx, sizeof(indx));
SecureZeroMemory(indy, sizeof(indy));
SecureZeroMemory(indz, sizeof(indz));

int widthx = width;
int counter = 0;
int xtrack = 0;
int zvalue = 0;
int limit = (widthx+1) * (widthx+1);
while(counter < limit){

arrayx[counter] = xtrack;
arrayy[counter] = 0;
arrayz[counter] = zvalue;

if( xtrack > 0 && xtrack % widthx == 0)
{
xtrack = 0;
++zvalue;
} else { ++xtrack;}
++counter;
}
counter = 0;
cout << "\n x: ";
while(counter < limit){
cout << arrayx[counter];
++counter;
}
counter = 0;
cout << "\n y: ";
while(counter < limit){
cout << arrayy[counter];
++counter;
}

counter = 0;
cout << "\n z: ";
while(counter < limit){
cout << arrayz[counter];
++counter;
}

/********* indicies loop **********/
int totaltriangles = (widthx * widthx) * 2;
counter = 0;
int yoffset = widthx+1;
int zoffset = 1;
int zup = width + 2;
int xoffset = 0;
int eor = 1; //end of row
int eorMax = width * 2; //value eor resets at
while ( counter < totaltriangles){
indx[counter] = xoffset;
indy[counter] = yoffset;

if(counter % 2 == 0 || eor == eorMax ){
++xoffset;}
if(counter % 2 == 1){
indz[counter] = zup; ++zup;
++yoffset;
++zoffset;
} else {
indz[counter] = zoffset;

}

if(eor < eorMax){++eor;}
else{eor = 1; ++yoffset; ++zoffset; ++zup;}
++counter;
}
counter = 0;
cout << "\n\n Indicy Arrays: \n x:";
while(counter < totaltriangles){
cout << indx[counter];
++counter;
}
counter = 0;
cout << "\n y: ";
while(counter < totaltriangles){
cout << indy[counter];
++counter;
}
counter = 0;
cout << "\n z: ";
while(counter < totaltriangles){
cout << indz[counter];
++counter;
}
cout << "\n";
}


And this sends it all to the vertex/index buffers;


void init_graphics(void)
{
loadMesh();
buildSkyBox();
loadMap(MAP_WIDTH); //Load map data into arrays based on map size
int counter = 0; //Generic Loop Counter
limit = (MAP_WIDTH+1) * (MAP_WIDTH+1); //Calculate number of vertices
int sizearray = sizeof(CUSTOMVERTEX) * limit; //Calculate size of array to store vertices

CUSTOMVERTEX *vertices = new CUSTOMVERTEX[limit]; //Allocate memory on the heap
while(counter < limit){ //fill vertices array with our map data

vertices[counter].x = arrayx[counter];
vertices[counter].y = arrayy[counter];
vertices[counter].z = arrayz[counter];
vertices[counter].NORMAL.x = 0.0f;
vertices[counter].NORMAL.y = 0.1f;
vertices[counter].NORMAL.z = 0.0f;

++counter;
}

// create the vertex and store the pointer into v_buffer, which is created globally
d3ddev->CreateVertexBuffer(limit*sizeof(CUSTOMVERTEX),
0,
CUSTOMFVF,
D3DPOOL_MANAGED,
&v_buffer,
NULL);
VOID* pVoid; // the void pointer
v_buffer->Lock(0, 0, (void**)&pVoid, 0); // lock the vertex buffer
memcpy(pVoid, vertices, sizearray); // copy the vertices to the locked buffer
v_buffer->Unlock(); // unlock the vertex buffer

triangleCount = (MAP_WIDTH*MAP_WIDTH)*2; //total number of triangles we'll be using
int sizeofarrayind = triangleCount * 3; //work out how big of an array we need to store all the data
short *indices = new short[sizeofarrayind]; //allocate memory on the heap for our array
counter = 0; //Generic counter
int track = 0; //Some sort of tracker
int sizeofindices = sizeof(short) * sizeofarrayind; //calculate the memory size of our array
while(counter < triangleCount){ //fill the area with our map data
indices[track] = indx[counter];
indices[track+1] = indy[counter];
indices[track+2] = indz[counter];

track += 3;
++counter;
}
// create an index buffer interface called i_buffer
d3ddev->CreateIndexBuffer(sizeofarrayind*sizeof(short),
0,
D3DFMT_INDEX16,
D3DPOOL_MANAGED,
&i_buffer,
NULL);
// lock i_buffer and load the indices into it
i_buffer->Lock(0, 0, (void**)&pVoid, 0);
memcpy(pVoid, indices, sizeofindices);
i_buffer->Unlock();
delete[] vertices;
delete[] indices;
}


Any help would be appreciated! Thanks!

Nb: I know how rough some of my code is, it'll be more rigorous later I swear wink.png

Zul

Share this post


Link to post
Share on other sites
Advertisement
Hey thanks for the response, you were absolutely right! Now it's taking around 3 seconds for a 256 map happy.png Of course it was something that simple! ha!

Much appreciated,
Zul

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement