Databases
StackOverflow2013
Postsdbo
POcopy data which is allocated in device from device to host
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POcopy data which is allocated in device from device to host
    primarykey
    data
    text
    <p>I have a pointer which is dynamically allocated in device,then how can I copy it from device to host. </p> <pre><code>#include &lt;stdio.h&gt; #define cudaSafeCall(call){ \ cudaError err = call; \ if(cudaSuccess != err){ \ fprintf(stderr, "%s(%i) : %s.\n", __FILE__, __LINE__, cudaGetErrorString(err)); \ exit(EXIT_FAILURE); \ }} #define cudaCheckErr(errorMessage) { \ cudaError_t err = cudaGetLastError(); \ if(cudaSuccess != err){ \ fprintf(stderr, "%s(%i) : %s : (code %d) %s.\n", __FILE__, __LINE__, errorMessage, err, cudaGetErrorString(err)); \ exit(EXIT_FAILURE); \`` }} struct num{ int *a; int b; }; __device__ struct num *gun; int main() { int i; char c[100]; struct num *dun,*cun; cudaSafeCall(cudaSetDevice(1)); cun=(struct num*)malloc(10*sizeof(struct num)); cudaSafeCall(cudaMalloc(&amp;dun,10*sizeof(struct num))); cudaSafeCall(cudaMemcpyToSymbol(gun,&amp;dun,sizeof(struct num*))); __global__ void kernel(); kernel&lt;&lt;&lt;1,10&gt;&gt;&gt;(); cudaSafeCall(cudaDeviceSynchronize()); cudaCheckErr(c); cudaSafeCall(cudaMemcpyFromSymbol(&amp;dun,gun,sizeof(struct num*))); cudaSafeCall(cudaMemcpy(cun,dun,10*sizeof(struct num),cudaMemcpyDeviceToHost)); for(i=0;i&lt;10;i++) cudaSafeCall(cudaMalloc(&amp;csu[i].a,10*sizeof(int))); cudaSafeCall(cudaGetSymbolAddress((void**)csu[0].a,(void**)gun[0].a)); for(i=0;i&lt;10;i++) cun[i].a=(int*)malloc(10*sizeof(int)); for(i=0;i&lt;10;i++) cudaSafeCall(cudaMemcpy(cun[i].a,dun[i].a,10*sizeof(int),cudaMemcpyDeviceToHost)); printf("%d ",cun[8].b); printf("%d ",cun[8].a[8]); cudaSafeCall(cudaFree(dun)); free(cun); } __global__ void kernel() { int i; int tid=threadIdx.x; gun[tid].b=tid; gun[tid].a=(int*)malloc(10*sizeof(int));/*this is dynamically allocated in device.*/ for(i=0;i&lt;10;i++) gun[tid].a[i]=tid+i; } </code></pre> <p>In this program, it always comes to a "segmentation fault" in </p> <pre><code>cudaSafeCall(cudaMemcpy(cun[i].a,dun[i].a,10*sizeof(int),cudaMemcpyDeviceToHost)) </code></pre> <p>Why? And what can I do to copy this data from device to host? </p>
    singulars
    1. PO
      singulars
      1. PTAnswer
    1. This table or related slice is empty.
    1. PTQuestion
    1. UStalonmies
    1. UShelena
    plurals
    1. This table or related slice is empty.
    1. PL
      singulars
      1. LTLinked
    1. This table or related slice is empty.
    1. PO
      singulars
      1. PTAnswer
    1. VO
      singulars
      1. POcopy data which is allocated in device from device to host
      1. This table or related slice is empty.
      1. VTUpMod
    1. COThe dynamic memory allocation is supported only for latest cuda devices (since 2.0 version). If you want to compile for this architecture, add this nvcc parameter `-arch=sm_20` or `-arch=sm_21`.
      singulars
      1. POcopy data which is allocated in device from device to host
      1. USYappie
    2. COYes,I have used this parameter,and my cun[8].b is correct.Just don't know how to pass element a in the struct.
      singulars
      1. POcopy data which is allocated in device from device to host
      1. UShelena
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload