CODEWITHSUNDEEP
c++visual-studio-2010dynamiccudaparallel-processing
Hadi
Hadi

Reputation: 307

Dynamic parallelism programming supported by GTX 550 Ti graphic card?

I am trying to implement simple dynamic parallelism programming example like..

__global__ ChildKernel(void* data){
    //Operate on data
}
__global__ ParentKernel(void *data){
    ChildKernel<<<16, 1>>>(data);
}
// In Host Code
ParentKernel<<<256, 64>>(data);

They said it is supported in CUDA 5.0 and above. I have CUDA 5.0 installed and working on my system but is the device that is GTX 550 Ti support dynamic parallelism ? Thank you.

Upvotes: 0

Views: 155

Answers (1)

Robert Crovella
Robert Crovella

Reputation: 151799

No, it's not supported.

Dynamic Parallelism requires compute capability 3.5

Your device has compute capability 2.1

You'll find this requirement listed at the bottom of page 1 of the user guide I linked to in your last question on this subject.

Upvotes: 2

Related Questions