Get Segmentation fault in CUDA online course: Exercise: Array Manipulation on both the Host and Device

I did all excersizes on the CUDA online course up to the exercise

Exercise: Array Manipulation on both the Host and Device

I get a segmentation fault (from the host) when trying to initialize the array. It seems that the

cudaMallocManaged(&a, size)

is problematic.

i don’t seem to attach a file here ( I would attach the snapshot of the Debugging session.)

So here is a link to the snapshot

https://drive.google.com/open?id=1RXOovhUdj60YKgjuC5zxDTmi2axjPhGG

Any advise?

Could it be that Tesla C2075 GPU does not support CUDA managed memory?

If so, how should I replace

cudaMallocMAnaged()

using

cudaMemcpy

?

In general, how can I check if a GPU support this managed memory model?