Hi Filipe,
it worked, but your code had some small errors, I suppose typos.
The right code is:
// define a structure in the header
typedef struct {
GLfloat coords[3];
} MYCoords;
// allocate a MYCoords pointer and *not* a GLfloat pointer
MYCoords *pVertex = malloc(mTotVertex * sizeof(MYCoords));
// then assign the values...
pVertex[p].coords[0] = pointC[0];
pVertex[p].coords[1] = pointC[1];
pVertex[p].coords[2] = pointC[2];
// Then it works well with the VBO
glBufferData(GL_ARRAY_BUFFER, (sizeof(GLfloat) * mTotVertex * 3),
pVertex, GL_STATIC_DRAW);
// Then release the pointer
free(pVertex);
pVertex = nil;
Thanks for the "pointer"!!!
Best Regards
--
Lorenzo
email: email@hidden
> From: Filipe Varela <email@hidden>
> Date: Fri, 25 May 2007 12:55:18 +0100
> To: Lorenzo <email@hidden>
> Cc: email@hidden
> Subject: Re: Too many vertexes
>
> That problem isn't related to available memory or the gpu memory. The
> whole thing takes
>
> sizeof(GLfloat) * 3 * 480000 bytes which is way more than what you're
> allowed to statically allocate.
> use a GLfloat pointer like this
>
> typedef struct {
> GLfloat coords[3];
> } vertex_t;
>
> GLfloat *dataPtr = malloc(mTotVertexes * vertex_t);
>
> What you're doiing is actually allocating the thing twice. Once
> statically with the declaration
> GLfloat vertexes[mTotVertexex][3]; //evaluates to static allocation
> of vertexes[480000][3]
>
> this memory (if you could allocate it statically) would not be used
> by the rest of your program since you are reallocating it dynamically.
>
> Question: why the double *?
>
> Rewrite of your code... Should work
>
>> int mTotVertexes = 480000;
>> GLfloat *vertexes;
>> glBufferData(GL_ARRAY_BUFFER, (sizeof(GLfloat) * mTotVertexes *
>> 3),
>> vertexes, GL_STATIC_DRAW);
>>
>> So I have tried to calloc the array this way.
>>
>> vertexes = (GLfloat *)malloc(mTotVertexes * 3 * sizeof(GLfloat));
>
> Remember: you can use a simple vector (1d) to store objects
> (vertexes) with more than one coordinate.
> To iterate through the vector and set/get axis coordinates
>
> for(i=0;i<mTotVertexes;i+=3) {
> x = vertexes[+0];
> y = vertexes[i+1];
> z = vertexes[i+2];
> }
>
> And finally (THIS IS REALLY IMPORTANT): Do not ever do this
> vertexes[i] = (GLfloat*)calloc(3, sizeof(GLfloat));
> don't ever iterate through an array and allocate memory for each
> item. If you want to do this, you might as well use a linked list.
> The whole purpose of using arrays/vbos is to get a continuous sector
> in memory which can be easily copied back and forth with a single
> memcpy. By dynamically allocating EACH vertex's coordinates, you're
> incresing the probabilities of NOT having your data in a continuous
> block. The end result is a delay between reads/writes of each vertex
> while the gpu uses dma to search the address of the vertex before
> copy ( i think).
>
> Cheers,
> Filipe
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Mac-opengl mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/mac-opengl/email@hidden
This email sent to email@hidden
