Verify global_barrier, indexing, private array inside device_func

[diptorupd]

• Have you provided a meaningful PR description?

The existing func decorator has several design limitation that does not allow usage of features such as barrier, dpex.local.array and dpex.private.array inside a func decorated function. These limitations reduced the utility of the feature. See: https://github.com/soda-inria/sklearn-numba-dpex/blob/747914481371ee7fe6527240a7bb57104d207115/sklearn_numba_dpex/kmeans/kernels/lloyd_single_step.py#L142

All these limitations are addressed by the new experimental.device_func decorator. The PR adds unit tests to verify that kernel_api.global_barrier and kernel_api.PrivateArray can be used in a device_func decorated function.

• Have you added a test, reproducer or referred to an issue with a reproducer?
• Have you tested your changes locally for CPU and GPU devices?
• Have you made sure that new changes do not introduce compiler warnings?
• If this PR is a work in progress, are you filing the PR as a draft?

[coveralls]

coverage: 82.781%. remained the same
when pulling d89104d on device_func_unit_tests
into 97f9069 on main.

[ZzEeKkAa]

LGTM! Thank you