0

I am new to intel-oneapi, I followed intel's instructions to install oneApi on Ubuntu 20.04.

I successfully compiled vector-add example. Unfortunately, during runtime I get the following exception:

pi_die: CUDA support was not enabled at compilation time
terminate called without an active exception
Aborted (core dumped)

Do I need special flags to run it on nvidia GPU? when I switch sycl::default_selector to sycl::cpu_selector then code is executed without exception.

Piotr Barejko
  • 608
  • 4
  • 15

1 Answers1

1

If you are targetting Nvidia hardware using DPC++ you currently need to re-compile DPC++ with this support. The instructions for doing this are in the repository here.

Once you have done that you can use the appropriate device selector to choose the Nvidia GPU for executing.

Rod Burns
  • 2,104
  • 13
  • 24