Conversation
…obal iteration space to stencils
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
…Modeling/gt4py into edges_and_corners_proposal
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
Co-Authored-By: Jeremy McGibbon <jeremym@vulcan.com>
|
I will amend this to reflect the current regional computation concept. |
egparedes
left a comment
egparedes
left a comment
There was a problem hiding this comment.
I left some comments about extending a bit more the examples and the explanations and a suggestion to improve the syntax.
egparedes
left a comment
egparedes
left a comment
There was a problem hiding this comment.
I have some minor comments to clarify some ambiguous points, but overall I think now the definition of the feature is clear. I would also like that someone else with experience in the GTScript parallel model (or the GridTools parallel model) reviewed the proposal to look for possible ambiguities in a C++ implementation. @havogt ?
docs/GDPs/gdp-0002-regions.rst
Outdated
| Axis Offsets therefore interally hold another offset, which can be set by adding or subtracting from the subscript. | ||
| For example ``I[0] - 2`` is itself an Axis Offsets that refers to 2 points before the start of the compute domain. | ||
|
|
||
| Axis Offsets may be assigned to variables in Python and/or used as externals in GT4Py in order to define ``region``. |
There was a problem hiding this comment.
Do you mean that an AxisOffset object can be created and manipulated in Python or that AxisOffset indices can be Python variables? It's not 100% clear to me.
There was a problem hiding this comment.
Both. They are a proper class with __add__ __sub__, and the r-variants.
Co-authored-by: Enrique G. Paredes <18477+egparedes@users.noreply.github.com>
|
@egparedes @mbianco Is this ready to be accepted then? |
egparedes
left a comment
egparedes
left a comment
There was a problem hiding this comment.
Last two minor changes to clarify and this is finally ready to be merged for me ASAP.
|
I am OK with merging this! |
5 participants
Discussion about GDP-2. Credit goes to @rheacangeo who pushed this forward.