Merge branch 'develop' into jmm/make-get-parent-pointer-consistent

.github/PULL_REQUEST_TEMPLATE.md

@@ -18,6 +18,10 @@ detail.  Why is this change required?  What problem does it solve?-->
 - [ ] Adds a test for any bugs fixed. Adds tests for new features.
 - [ ] Code is formatted
 - [ ] Changes are summarized in CHANGELOG.md
+- [ ] Change is breaking (API, behavior, ...)
+  - [ ] Change is *additionally* added to CHANGELOG.md in the breaking section
+  - [ ] PR is marked as breaking
+  - [ ] Short summary API changes at the top of the PR (plus optionally with an automated update/fix script)
 - [ ] CI has been triggered on [Darwin](https://re-git.lanl.gov/eap-oss/parthenon/-/pipelines) for performance regression tests.
 - [ ] Docs build
 - [ ] (@lanl.gov employees) Update copyright on changed files

CHANGELOG.md

@@ -2,6 +2,23 @@
 
 ## Current develop
 
+### Added (new features/APIs/variables/...)
+
+### Changed (changing behavior/API/variables/...)
+
+### Fixed (not changing behavior/API/variables/...)
+
+### Infrastructure (changes irrelevant to downstream codes)
+
+### Removed (removing behavior/API/varaibles/...)
+
+### Incompatibilities (i.e. breaking changes)
+- [[PR 974]](https://github.com/parthenon-hpc-lab/parthenon/pull/974) Change GetParentPointer to always return T*
+
+
+## Release 23.11
+Date: 2023-11-16
+
 ### Added (new features/APIs/variables/...)
 - [[PR 962]](https://github.com/parthenon-hpc-lab/parthenon/pull/962) Add support for in-situ histograms/profiles
 - [[PR 911]](https://github.com/parthenon-hpc-lab/parthenon/pull/911) Add infrastructure for geometric multi-grid
@@ -21,8 +38,9 @@
 - [[PR 868]](https://github.com/parthenon-hpc-lab/parthenon/pull/868) Add block-local face, edge, and nodal fields and allow for packing
 
 ### Changed (changing behavior/API/variables/...)
-- [[PR 974]](https://github.com/parthenon-hpc-lab/parthenon/pull/974) Change GetParentPointer to always return T*
-- [[PR 976]](https://github.com/parthenon-hpc-lab/parthenon/pull/976) Move UserWorkBeforeLoop to be after first o
+- [[PR 977]](https://github.com/parthenon-hpc-lab/parthenon/pull/977) Change to CalVer from SemVer
+- [[PR 975]](https://github.com/parthenon-hpc-lab/parthenon/pull/975) Construct staged integrators via arbitrary name
+- [[PR 976]](https://github.com/parthenon-hpc-lab/parthenon/pull/976) Move UserWorkBeforeLoop to be after first output
 - [[PR 965]](https://github.com/parthenon-hpc-lab/parthenon/pull/965) Allow leading whitespace in input parameters
 - [[PR 926]](https://github.com/parthenon-hpc-lab/parthenon/pull/926) Internal refinement op registration
 - [[PR 897]](https://github.com/parthenon-hpc-lab/parthenon/pull/897) Deflate compression filter is not called any more if compression is soft disabled

CMakeLists.txt

@@ -20,7 +20,7 @@ cmake_minimum_required(VERSION 3.16)
 # Imports machine-specific configuration
 include(cmake/MachineCfg.cmake)
 
-project(parthenon VERSION 0.8.0 LANGUAGES C CXX)
+project(parthenon VERSION 23.11 LANGUAGES C CXX)
 
 if (${CMAKE_VERSION} VERSION_GREATER_EQUAL 3.19.0)
   cmake_policy(SET CMP0110 NEW)

CONTRIBUTING.md

@@ -39,26 +39,20 @@ Use GitHub labels as appropriate and feel free to directly add/tag people to the
 
 ### Summary of Branching Model and Versioning
 
-Two main branches exist:
-- `stable` contains the latest "stable" release
-- `develop` contains all approved changes since the previous release
+Only a single main branch called `develop` exists and all PRs should be merged
+into that branch.
+Individual versions/releases are tracked by tags.
 
-We aim at creating a new release everyone 4 to 6 weeks.
+We aim at creating a new release every 6 months.
 The decision on creating a new release is made during the bi-weekly calls.
-A release consists of of merging `develop` into `stable` and create a new tag for that version
-using a modified [semantic versioning](https://semver.org/) scheme.
-Releases will be tagged `v0.MAJOR.MINOR` given the current rapid development.
-
-- MAJOR is incremented for API incompatible changes
-- MINOR is incremented for backwards compatible changes and bug fixes
-
-This scheme will be reevaluated once a future version is considered to be the first official stable release.
-
-The main idea behind separating `stable` from `develop` is to allow for more in-depth nightly testing
-on the latter.
-This specifically applies to downstream codes so that incompatibilities (e.g., due to to an
-updated API) are discovered early.
-
+Following steps need to be done for a new release:
+
+- Create a new tag for that version using a modified [calender versioning](https://calver.org/) scheme.
+Releases will be tagged `vYY.0M` i.e., short years and zero-padded months.
+- Update the version in the main `CMakeLists.txt`.
+- Update the `CHANGELOG.md` (i.e., add a release header and create new empty
+categories for the "Current develop" section.
+- Sent a mail to the mailing list announcing the new release.
 
 ### Contributing Code
 

src/time_integration/butcher_integrator.cpp

@@ -36,8 +36,12 @@ namespace parthenon {
  * alpha_k = c_k
  * c_k = b_k
  */
-ButcherIntegrator::ButcherIntegrator(ParameterInput *pin)
-    : StagedIntegrator(pin->GetOrAddString("parthenon/time", "integrator", "rk2")) {
+
+//----------------------------------------------------------------------------------------
+//! \class ButcherIntegrator::ButcherIntegrator(const std::string &name)
+//! \brief Constructs a ButcherIntegrator instance given a string (e.g., rk2, rk3..)
+
+ButcherIntegrator::ButcherIntegrator(const std::string &name) : StagedIntegrator(name) {
   if (name_ == "rk1") {
     nstages = nbuffers = 1;
     Resize_(nstages);
@@ -219,6 +223,17 @@ ButcherIntegrator::ButcherIntegrator(ParameterInput *pin)
   }
 }
 
+//----------------------------------------------------------------------------------------
+//! \class ButcherIntegrator::ButcherIntegrator(ParameterInput *pin)
+//! \brief Constructs a ButcherIntegrator instance given ParameterInput *pin
+
+ButcherIntegrator::ButcherIntegrator(ParameterInput *pin)
+    : ButcherIntegrator(pin->GetOrAddString("parthenon/time", "integrator", "rk2")) {}
+
+//----------------------------------------------------------------------------------------
+//! \fn  void ButcherIntegrator::Resize_(int nstages)
+//! \brief Resizes ButcherIntegrator registers given a supplied integer nstages
+
 void ButcherIntegrator::Resize_(int nstages) {
   a.resize(nstages);
   for (int i = 0; i < a.size(); ++i) {

src/time_integration/low_storage_integrator.cpp

@@ -43,8 +43,13 @@ namespace parthenon {
  * Stone et al., ApJS (2020) 249:4
  * See equations 11 through 15.
  */
-LowStorageIntegrator::LowStorageIntegrator(ParameterInput *pin)
-    : StagedIntegrator(pin->GetOrAddString("parthenon/time", "integrator", "rk2")) {
+
+//----------------------------------------------------------------------------------------
+//! \class LowStorageIntegrator::LowStorageIntegrator(const std::string &name)
+//! \brief Constructs a LowStorageIntegrator instance given a string (e.g., rk2, rk3..)
+
+LowStorageIntegrator::LowStorageIntegrator(const std::string &name)
+    : StagedIntegrator(name) {
   if (name_ == "rk1") {
     nstages = 1;
     nbuffers = 1;
@@ -155,4 +160,11 @@ LowStorageIntegrator::LowStorageIntegrator(ParameterInput *pin)
   MakePeriodicNames_(stage_name, nstages);
 }
 
+//----------------------------------------------------------------------------------------
+//! \class LowStorageIntegrator::LowStorageIntegrator(ParameterInput *pin)
+//! \brief Constructs a LowStorageIntegrator instance given ParameterInput *pin
+
+LowStorageIntegrator::LowStorageIntegrator(ParameterInput *pin)
+    : LowStorageIntegrator(pin->GetOrAddString("parthenon/time", "integrator", "rk2")) {}
+
 } // namespace parthenon

src/time_integration/staged_integrator.hpp

@@ -49,6 +49,7 @@ class StagedIntegrator {
 class LowStorageIntegrator : public StagedIntegrator {
  public:
   LowStorageIntegrator() = default;
+  explicit LowStorageIntegrator(const std::string &name);
   explicit LowStorageIntegrator(ParameterInput *pin);
   std::vector<Real> delta;
   std::vector<Real> beta;
@@ -60,6 +61,7 @@ class LowStorageIntegrator : public StagedIntegrator {
 class ButcherIntegrator : public StagedIntegrator {
  public:
   ButcherIntegrator() = default;
+  explicit ButcherIntegrator(const std::string &name);
   explicit ButcherIntegrator(ParameterInput *pin);
   // TODO(JMM): Should I do a flat array with indexing instead?
   std::vector<std::vector<Real>> a;