PyManga's C++ kernel #388

Lgz-tud · 2025-09-16T12:20:50Z

No description provided.

mcwimm

Great! It's much clearer than before. I haven't tried it yet.

In general, I'm still not 100% convinced by the structure if we want to include it for more modules. Perhaps we could make both versions (original and CPP) available without having multiple modules (as before). I think you're almost there.

This would also enable us to create benchmarks for comparing the implementations individually.

mcwimm · 2025-09-30T07:09:12Z

Benchmarks/ModuleBenchmarks/Resource/AboveGround/AsymmetricZOI/AsymmetricZOI.xml

I'm don't get why this benchmark file is located in this location and what the difference to the previous one is.

Yes, they are identical. We can directly reference the previous XML file without placing it in a new location.

mcwimm · 2025-09-30T07:11:01Z

CMakeLists.txt

@@ -0,0 +1,251 @@
+# ============================================================
+# CmakeLists overview (pymanga_cpp) — What this CMake does


If this is something a user needs to do before using cpp modules, we need to provide this information elsewhere, e.g., in the source code and input documentation?

CMakeLists.txt serves as our build script, instructing CMake on how to compile and link our source code. The introductory remarks could be included in the readme, though such complexity is unnecessary. I have written this in such detail solely for your review. Typically, when publishing a model, such exhaustive explanations are unnecessary. Often, it suffices to state in the README that we use CMake, much like iland's Qt implementation: https://github.com/edfm-tum/iland-model

mcwimm · 2025-09-30T07:14:37Z

ResourceLib/AboveGround/AsymmetricZOI/AsymmetricZOI.cpp

@@ -0,0 +1,154 @@
+// ResourceLib/AboveGround/AsymmetricZOI/AsymmetricZOI.cpp


Could you please check whether we can include these files in our source code and input documentation? If so, could you also check what the comment format is to do this automatically?

yes, I will check it.

mcwimm · 2025-09-30T07:16:03Z

ResourceLib/AboveGround/AsymmetricZOI/AsymmetricZOI.md

+The C++ core parallelizes the grid evaluation using OpenMP, allowing efficient simulation of large stands with many individuals and fine spatial resolution.
+
+
+# Compilation


I guess this will be the same for all modules. So maybe we should find another position in the documentation where we can describe this once.

Yes, I agree with your point.

mcwimm · 2025-09-30T07:17:07Z

ResourceLib/AboveGround/AsymmetricZOI/AsymmetricZOI.md

- Define the grid on which to calculate above-ground resource availability.
- The total size is 20x20 m, and a cell is 0.25 x 0.25 m (20 m / 80 cells = 0.25 m/cell).
+- Define the large grid on which to calculate above-ground resource availability.
+- The total size is 1000x1000 m, and a cell is 0.25 x 0.25 m (20 m / 80 cells = 0.25 m/cell).


I would keep the examples here small as they are supposed to run very quickly for all users.

No problem. This can be adjusted as required.

mcwimm · 2025-09-30T07:22:41Z

ResourceLib/AboveGround/AsymmetricZOI/AsymmetricZOI.py

+#!/usr/bin/env python3  # Shebang for Unix-like systems to run the script with Python 3
+# -*- coding: utf-8 -*-  # File encoding declaration

+import numpy as np  


Is it possible to implement a switch, e.g., at project level, so that we can keep the plain python implementation. Depending on the settings in the project file, the python or cpp module is used.

Not every user will be able to change the c-code or wants to compile C++.

Yes, need to add a new conditional function. Then expose a parameter interface, allowing users to freely utilise the module by controlling whether to enable C++.

mcwimm · 2025-09-30T07:24:23Z

ResourceLib/AboveGround/AsymmetricZOI/AsymmetricZOI.py

+    The class computes per-plant aboveground resource factors via asymzoi.compute_aboveground_resources.  
    """
-    def __init__(self, args):
-        """


Docstrings (""" THIS IS A COMMENT""") will be automatically included in our source code documentation. Could you please leave them where they are?

mcwimm · 2025-09-30T07:28:03Z

ResourceLib/AboveGround/AsymmetricZOI/AsymmetricZOI.py

+        gx, gy = self._require_grid()  # Retrieve validated grid arrays (2D and same shape)
+
+        # Convert inputs to C-contiguous float32 arrays to match the C++ interface and save memory
+        xe = np.ascontiguousarray(np.asarray(self.xe, dtype=np.float32))  # X positions of plants


Try to avoid making comments that say the same thing as the code (Check out the first rule here: https://stackoverflow.blog/2021/12/23/best-practices-for-writing-code-comments/).

Lgz-tud · 2025-09-30T07:51:27Z

I shall be taking a week's leave from the afternoon of today until the 8th of October. I cannot confirm whether the new submission will be completed today. Should it not be finalised today, the new pull request will be processed after the 8th.

…ames.

[RL] Using C++ to Refactor AsymmetricZOI

0c72072

Lgz-tud changed the title ~~[RL] Using C++ to Refactor AsymmetricZOI~~ PyManga's C++ kernel Sep 16, 2025

[RL] fixed broken link to Visual Studio download page

3b94cae

mcwimm reviewed Sep 30, 2025

View reviewed changes

Lgz-tud added 2 commits November 3, 2025 12:00

[RL] Add optional cpp or python backends

c9c7292

[RL] Update the test files for the BenchmarkList.xml and input path n…

f1f5b72

…ames.

		@@ -0,0 +1,251 @@
		# ============================================================
		# CmakeLists overview (pymanga_cpp) — What this CMake does

		@@ -0,0 +1,154 @@
		// ResourceLib/AboveGround/AsymmetricZOI/AsymmetricZOI.cpp

		The C++ core parallelizes the grid evaluation using OpenMP, allowing efficient simulation of large stands with many individuals and fine spatial resolution.


		# Compilation

PyManga's C++ kernel #388

Are you sure you want to change the base?

PyManga's C++ kernel #388

Uh oh!

Conversation

Lgz-tud commented Sep 16, 2025

Uh oh!

mcwimm left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Lgz-tud commented Sep 30, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants