Databases
StackOverflow2013
Postsdbo
POCopying a dynamically allocated 2D array from host to device in CUDA
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POCopying a dynamically allocated 2D array from host to device in CUDA
    primarykey
    data
    text
    <p>I want to copy a dynamically allocated 2D array from host to device to get its Discrete Fourier Transform. </p> <p>I'm using below code to copy the array to the device </p> <pre><code>cudaMalloc((void**)&amp;array_d, sizeof(cufftComplex)*NX*(NY/2+1)); cudaMemcpy(array_d, array_h, sizeof(float)*NX*NY, cudaMemcpyHostToDevice); </code></pre> <p>This works fine with static arrays, i get the intended output from my fft. But it doesn't work with dynamic arrays. After little bit searching I learnt I can not copy dynamic arrays like this from host to device. So I found this solution.</p> <pre><code>cudaMalloc((void**)&amp;array_d, sizeof(cufftComplex)*NX*(NY/2+1)); for(int i=0; i&lt;NX; ++i){ cudaMemcpy(array_d+ i*NY, array_h[i], sizeof(float)*NY, cudaMemcpyHostToDevice); } </code></pre> <p>But it's also not doing the task properly since I get wrong values from my fft.</p> <p>Given below is my fft code. </p> <pre><code>cufftPlanMany(&amp;plan, NRANK, n,NULL, 1, 0,NULL, 1, 0,CUFFT_R2C,BATCH); cufftSetCompatibilityMode(plan, CUFFT_COMPATIBILITY_NATIVE); cufftExecR2C(plan, (cufftReal*)data, data); cudaThreadSynchronize(); cudaMemcpy(c, data, sizeof(float)*NX*NY, cudaMemcpyDeviceToHost); </code></pre> <p>How can I overcome this problem ?</p> <p>EDIT</p> <p>given below is the code</p> <pre><code>#define NX 4 #define NY 5 #define NRANK 2 #define BATCH 10 #include "cuda_runtime.h" #include "device_launch_parameters.h" #include &lt;cufft.h&gt; #include &lt;stdio.h&gt; #include &lt;iostream&gt; int check(); int main() { // static array float b[NX][NY] ={ {0.7943 , 0.6020 , 0.7482 , 0.9133 , 0.9961}, {0.3112 , 0.2630 , 0.4505 , 0.1524 , 0.0782}, {0.5285 , 0.6541 , 0.0838 , 0.8258 , 0.4427}, {0.1656 , 0.6892 , 0.2290 , 0.5383 , 0.1067} }; // dynamic array float **a = new float*[NX]; for (int r = 0; r &lt; NX; ++r) { a[r] = new float[NY]; for (int c = 0; c &lt; NY; ++c) { a[r][c] = b[r][c]; } } // arrray to store the results - host side float c[NX][NY] = { 0 }; cufftHandle plan; cufftComplex *data; int n[NRANK] = {NX, NY}; cudaMalloc((void**)&amp;data, sizeof(cufftComplex)*NX*(NY/2+1)); cudaMemcpy(data, b, sizeof(float)*NX*NY, cudaMemcpyHostToDevice); /* Create a 2D FFT plan. */ cufftPlanMany(&amp;plan, NRANK, n,NULL, 1, 0,NULL, 1, 0,CUFFT_R2C,BATCH); cufftSetCompatibilityMode(plan, CUFFT_COMPATIBILITY_NATIVE); cufftExecR2C(plan, (cufftReal*)data, data); cudaThreadSynchronize(); cudaMemcpy(c, data, sizeof(float)*NX*NY, cudaMemcpyDeviceToHost); cufftDestroy(plan); cudaFree(data); return 0; } </code></pre>
    singulars
    1. PO
      singulars
      1. PTAnswer
    1. This table or related slice is empty.
    1. PTQuestion
    1. USOptimus
    1. USOptimus
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PO
      singulars
      1. PTAnswer
    1. VO
      singulars
      1. POCopying a dynamically allocated 2D array from host to device in CUDA
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. POCopying a dynamically allocated 2D array from host to device in CUDA
      1. This table or related slice is empty.
      1. VTUpMod
    3. VO
      singulars
      1. POCopying a dynamically allocated 2D array from host to device in CUDA
      1. USDabo
      1. VTFavorite
    1. COWe will need to see a short, complete, compilable example, I think. What is `c` in the FFT code? How is it allocated? is it like `array_h` in the first code? what is `data` in the FFT code? is it the same as `array_d` in the first code? I could go on.
      singulars
      1. POCopying a dynamically allocated 2D array from host to device in CUDA
      1. UStalonmies
    2. COMost importantly, how is `array_h` defined and initialised?
      singulars
      1. POCopying a dynamically allocated 2D array from host to device in CUDA
      1. USAngew
    3. COsorry. i added the code ....
      singulars
      1. POCopying a dynamically allocated 2D array from host to device in CUDA
      1. USOptimus
 

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