Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>Start by doing proper <a href="https://stackoverflow.com/questions/14038589/what-is-the-canonical-way-to-check-for-errors-using-the-cuda-runtime-api">cuda error checking</a> on all your cuda API calls (e.g. cudaMemcpy, etc.) and kernel calls.</p> <p>When you do that you'll discover that your kernels are not running successfully. These types of things won't work:</p> <pre><code>uchar4 *devPtr; // you've just created an unallocated NULL host pointer size_t img1_size = IMAGESIZE_MAX; kernel&lt;&lt;&lt;grids,threads&gt;&gt;&gt;(devPtr); // this kernel will fail uchar4 *devPtr2; // you've just created an unallocated NULL host pointer size_t img2_size = IMAGESIZE_MAX; kernel2&lt;&lt;&lt;grids,threads&gt;&gt;&gt;(devPtr2); // this kernel will fail </code></pre> <p><code>devPtr</code> and <code>devPtr2</code> in the above code are NULL pointers. You haven't allocated any storage associated with them. Furthermore, since you are passing them to device kernels, they need to be allocated with <code>cudaMalloc</code> or similar API function, in order for the pointers to be usable in device code.</p> <p>Since they are not allocated with <code>cudaMalloc</code>, as soon as you try to dereference those pointers in device code, you'll create a kernel fault. This will be evident if you do error checking, as you will have "unspecified launch failure" or similar report from those kernels.</p> <p>I think there are probably a number of other problems in your code, but first you should do proper cuda error checking and at least get your code to the point where everything you've written is, in fact, running.</p> <p>And the code you've posted doesn't actually compile.</p> <p>After fixing the compile errors I also discovered that you have another infinite loop:</p> <pre><code>cudaMalloc ( (uchar4 **)&amp;pBufferCurrent, sizeTotal + sizeof(size) + size); cudaMalloc ( (uchar4 **)&amp;pBuffer, sizeTotal + sizeof(size) + size); do { if (!pBufferCurrent) { break; } pBuffer = pBufferCurrent; pBufferCurrent += sizeTotal; imageget ( pBufferCurrent + sizeof(size), size, devPtr); sizeTotal += (sizeof(size) + size); } while (a==1); </code></pre> <p>Since <code>a</code> is initialized to 1 in your loop, and nothing in the loop modifies <code>a</code>, the loop will never exit based on the <code>while</code> condition. Since pBufferCurrent is also never zero if it's been properly set up by <code>cudaMalloc</code>, the <code>break</code> will never be taken.</p> <p>If you <code>malloc</code> or <code>cudaMalloc</code> a pointer called <code>pBufferCurrent</code>, it's hard for me to imagine under what circumstances this would ever make sense:</p> <pre><code>pBufferCurrent += sizeTotal; </code></pre> <p>and although this is legal, I don't see how this makes sense:</p> <pre><code>pBuffer = pBufferCurrent; </code></pre> <p>You just created an allocation for <code>pBuffer</code> using <code>cudaMalloc</code>, but the first thing you do is throw it away?</p>
    singulars
    1. This table or related slice is empty.
    1. POC/CUDA - Modifying CUDA/GL interop example to store image in a memory buffer
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USCommunity
    1. USRobert Crovella
    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. This table or related slice is empty.
    1. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    1. COI allocated memory for the two pointers devptr and devptr2 using cudaMalloc and I used the "HANDLE_ERROR" CUDA error checking function that came along with the example code. The code actually compiles for me but I am not sure why the code is not compiling for you?
      singulars
      1. PO
      1. USEagle
    2. COI don't see any statements in the code you've posted that are using `cudaMalloc` to allocate the pointers `devptr` and `devptr2`. Are you looking at the code you've posted in this question? Regarding compilation, the reason is because the code you're compiling is not what you've posted in this question. Look at the code *in the question*, specifically at the last call to `cudaGraphicsResourceGetMappedPointer`. Does that line contain a semicolon terminator? It does not. It will not compile correctly.
      singulars
      1. PO
      1. USRobert Crovella
    3. COI'm sorry, I had forgotten to add an a++ statement in the do-while loop. I am editing the code that I pasted now. I am trying my best to write a program for a memory buffer that can store images coming from an OCT system. I realize that I am a beginner in C/CUDA programming and I am trying to learn as fast as I can. It would be great if you could help me write a memory buffer program that is correct. I am stumped here. Every program I write either has an infinite loop to it or some other problem
      singulars
      1. PO
      1. USEagle
 

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