Add a target inside the experimental module#1213
Conversation
| from numba_dpex.core.pipelines import kernel_compiler | ||
| from numba_dpex.core.types import DpnpNdArray | ||
|
|
||
| from .target import dpex_exp_kernel_target |
There was a problem hiding this comment.
do we want to avoid relative imports?
There was a problem hiding this comment.
We can, but in the past I have found relative imports to be the only way out of avoiding circular imports.
There was a problem hiding this comment.
It is okay for now then. We just need to rethink design to avoid circular imports
| """ | ||
|
|
||
| targetdescr = dpex_kernel_target | ||
| targetdescr = dpex_exp_kernel_target |
There was a problem hiding this comment.
I can't see usage of targetdescr
There was a problem hiding this comment.
It is needed by the superclass: https://github.com/numba/numba/blob/a4664180ddc91e4c8a37dd06324f4e229a76df91/numba/core/dispatcher.py#L796
The variable is used in the super().__init__: https://github.com/numba/numba/blob/a4664180ddc91e4c8a37dd06324f4e229a76df91/numba/core/dispatcher.py#L823
There was a problem hiding this comment.
Sorry, my fault. It was confused, because it is defined on class level. I was expecting to see it in init with self.*.
Add a target inside the experimental module 6f670d6
2 participants
Have you provided a meaningful PR description?
Changes introduced in the experimental module should not have side effects on the
numba_dpex.corecomponents. For this reason, an experimental target is being introduced intonumba_dpex.experimental. Any future typing or lowering changes for any experimental feature should be introduced here.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?