New parallelization (#108)

* initial work * new parallelization preliminary done * small fix in tests * commented unnecessary code * updated construction sites * hopefully fixed parallelization * small test fix * initial work * new parallelization preliminary done * small fix in tests * commented unnecessary code * updated construction sites * hopefully fixed parallelization * small test fix * only print when verbosity > 0, otherwise nothing is printed (bad) * pre-commit fixes * added ncores to config + implemented setting ncores for external programs * fixed tests * added check for number of cores available vs needed * pre-commit * test fix * fixed tm implementation * updated main.py * tqdm progress bar * updated dependencies * mypy types import * moved warnings in correct bracket * updated default config toml * added final output of molecules and timing * fixed time * better time info * print formatting * shift block setup to parallel.py Signed-off-by: Marcel Müller <[email protected]> * avoid UnboundLocalError Signed-off-by: Marcel Müller <[email protected]> * some code formatting and printout adjustments Signed-off-by: Marcel Müller <[email protected]> * shifted CHANGELOG entry to correct position Signed-off-by: Marcel Müller <[email protected]> * update CODEOWNERS file Signed-off-by: Marcel Müller <[email protected]> --------- Signed-off-by: Marcel Müller <[email protected]> Co-authored-by: Marcel Müller <[email protected]>
grimme-lab · Jan 20, 2025 · 7364159 · 7364159
1 parent 89a3d58
commit 7364159
Show file tree

Hide file tree

Showing 18 changed files with 426 additions and 129 deletions.
diff --git a/.github/CODEOWNERS b/.github/CODEOWNERS
@@ -1,12 +1,12 @@
 # These owners will be the default owners for everything in
 # the repo. Unless a later match takes precedence,
-# @marcelmbn and @thfroitzheim will be requested for
+# the names mentioned after the '*' will be requested for
 # review when someone opens a pull request.
 *  @marcelmbn @jonathan-schoeps
 
 # These parts are specifically owned by some people
 /src/mindlessgen/cli @marcelmbn
-/src/mindlessgen/generator @marcelmbn
+/src/mindlessgen/generator @marcelmbn @lmseidler
 /src/mindlessgen/molecules @marcelmbn @jonathan-schoeps
-/src/mindlessgen/prog @marcelmbn @jonathan-schoeps
-/src/mindlessgen/qm @marcelmbn
+/src/mindlessgen/prog @marcelmbn @jonathan-schoeps @lmseidler
+/src/mindlessgen/qm @marcelmbn @jonathan-schoeps
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -9,9 +9,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - to set the elemental composition it is now possible to use dicts with not only int but also the element symbols (str)
 - dict keys for elemental compositions will now always be checked for validity
 - Renamed GP3-xTB to g-xTB
+- Nothing will be printed while multiple molecules are generated in parallel, tqdm-based progress bar instead
+- Some debugging statements from generate had to be removed (esp. w.r.t. early stopping)
 
 ### Added
 - `GXTBConfig` class for the g-xTB method, supporting SCF cycles check
+- support for TURBOMOLE as QM engine.
+- updated the parallelization to work over the number of molecules
 
 ### Fixed
 - version string is now correctly formatted and printed
@@ -45,7 +49,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - support for `python-3.13`
 - option to set a fixed molecular charge, while ensuring `uhf = 0`
 - `element_composition` and `forbidden_elements` can now be directly set to a `dict` or `list`, respectively, via API access
-- support for TURBOMOLE as QM engine.
 
 ### Breaking Changes
 - Removal of the `dist_threshold` flag and in the `-toml` file.

diff --git a/environment.yml b/environment.yml
@@ -9,5 +9,6 @@ dependencies:
   - pre-commit
   - pytest
   - tox
+  - tqdm
   - pip:
       - covdefaults
diff --git a/mindlessgen.toml b/mindlessgen.toml
@@ -58,6 +58,8 @@ hlgap = 0.5
 # > Debug this step. Leads to more verbose output as soon as the refinement part is reached. Options: <bool>
 # > If `debug` is true, the process is terminated after the first (successful or not) refinement step.
 debug = false
+# > Number of cores to be used for geometry optimizations. Single-points will continue to use one core each.
+ncores = 4
 
 [postprocess]
 # > Engine for the post-processing part. Options: 'xtb', 'orca', 'turbomole'
@@ -70,6 +72,8 @@ opt_cycles = 5
 # > If `debug` is true, the process is terminated after the first (successful or not) post-processing step.
 # > Note: This option is only relevant if the 'postprocess' option in the 'general' section is set to 'true'.
 debug = false
+# > Number of cores to be used for both single-point calculations and geometry optimizations.
+ncores = 4
 
 [xtb]
 # > Path to the xtb executable. The names `xtb` and `xtb_dev` are automatically searched for. Options: <str | Path>

diff --git a/pyproject.toml b/pyproject.toml
@@ -22,7 +22,7 @@ classifiers = [
     "Topic :: Scientific/Engineering",
     "Typing :: Typed",
 ]
-dependencies = ["numpy", "networkx", "toml"]
+dependencies = ["numpy", "networkx", "toml", "tqdm"]
 dynamic = ["version"]
 
 [project.urls]
@@ -41,6 +41,7 @@ dev = [
     "tox",
     "setuptools_scm>=8",
     "types-toml",
+    "types-tqdm",
 ]
 
 [project.scripts]
-Original file line number
+Diff line change
@@ Expand Up / @@ -9,5 +9,6 @@ dependencies: @@
       - pre-commit
       - pytest
       - tox
+      - tqdm
       - pip:
           - covdefaults