2

I have a geforce gpu i.e GTX 750Ti . It has compute capability 3.5. In the documentation of dynamic parallelism it is mentioned that requirement for dynamic parallilism requires 3.5 but it is not mentioned about Geforce or Tesla. sample program has mentioned minspec tesla3.5

and Simple program in geforce is not running , Is dynamic parallilism not supported in geforce GTX 750 Ti?

Malacu
  • 191
  • 2
  • 10

1 Answers1

4

GTX 750 Ti is compute capability 5.0.

Dynamic Parallelism is supported on any CUDA GPU that is compute capability 3.5 or higher.

So yes, your GTX 750Ti supports dynamic parallelism.

Robert Crovella
  • 143,785
  • 11
  • 213
  • 257
  • whenever I use cudadevrt.lib and changed Generate Relocatable code device to yes , even the simple program of squaring doesnot execute the kernal. – Malacu Sep 01 '14 at 04:31
  • 2
    When you asked about this [already](http://stackoverflow.com/questions/25564798/dynamic-parallelism-in-cuda-giving-wrong-output) you were advised to add [proper cuda error checking](http://stackoverflow.com/questions/14038589/what-is-the-canonical-way-to-check-for-errors-using-the-cuda-runtime-api) to your program. Have you done that? If so, what was the result? Also try `cuda-memcheck`. Finally, there are a number of CUDA [CDP sample codes](http://docs.nvidia.com/cuda/cuda-samples/index.html#simple-print--cuda-dynamic-parallelism-), you should try building and running one of those. – Robert Crovella Sep 01 '14 at 04:44
  • running CDP radix sort gives Invalid Item[1]: 107 greater than 13 – Malacu Sep 01 '14 at 05:50
  • What about 750 non-ti? wikpedia is blocked in my country. – huseyin tugrul buyukisik Jun 19 '17 at 20:39