Merge branch 'MFlowCode:master' into master

Author: danieljvickers

.claude/rules/common-pitfalls.md

@@ -59,7 +59,7 @@ Before submitting a PR:
 - [ ] `./mfc.sh precheck -j 8` (5 CI lint checks)
 - [ ] `./mfc.sh build -j 8` (compiles)
 - [ ] `./mfc.sh test --only <relevant> -j 8` (tests pass)
-- [ ] If adding parameters: all 4 locations updated
+- [ ] If adding parameters: definitions.py (_r + _nv) updated; cmake reconfigured; case_validator.py if constraints
 - [ ] If modifying `src/common/`: all three targets tested
 - [ ] If changing output: golden files regenerated for affected tests
 - [ ] One logical change per commit

.claude/rules/parameter-system.md

@@ -23,20 +23,22 @@ MFC has ~3,400 simulation parameters defined in Python and read by Fortran via n
    - Reads `&user_inputs` namelist
    - Each parameter must be declared in the namelist statement
 
-## Adding a New Parameter (4-location checklist)
+## Adding a New Parameter (2-location checklist)
 
-YOU MUST update the first 3 locations. Missing any causes silent failures or compile errors.
-Location 4 is required only if the parameter has physics constraints.
+Fortran declarations and namelist bindings are now auto-generated from definitions.py
+at CMake configure time — no manual Fortran edits needed for simple scalar parameters.
 
-1. **`toolchain/mfc/params/definitions.py`**: Add parameter with type, default, constraints
-2. **`src/*/m_global_parameters.fpp`**: Declare the Fortran variable in the relevant
-   target(s). If the param is used by simulation only, add it there. If shared, add to
-   all three targets' m_global_parameters.fpp.
-3. **`src/*/m_start_up.fpp`**: Add to the Fortran `namelist` declaration in the relevant
-   target(s).
-4. **`toolchain/mfc/case_validator.py`**: Add validation rules if the parameter has
+1. **`toolchain/mfc/params/definitions.py`**: Add parameter with `_r()` (type, default,
+   constraints) AND add it to `NAMELIST_VARS` via `_nv()` for the relevant target(s).
+   After editing, re-run cmake (or `./mfc.sh build`) to regenerate the Fortran includes.
+2. **`toolchain/mfc/case_validator.py`**: Add validation rules if the parameter has
    physics constraints. Include `PHYSICS_DOCS` entry with title, category, explanation.
 
+**Exceptions — still require manual Fortran edits:**
+- Array variables (e.g. `logical, dimension(num_fluids_max)`) → declare in `src/*/m_global_parameters.fpp`
+- Derived-type members (`fluid_pp%attr`, `patch_icpp(i)%attr`) → declare in the relevant derived type
+- Case-optimization parameters → add to `CASE_OPT_PARAMS` and the `#:else` block in `src/simulation/m_global_parameters.fpp`
+
 ## Case Files
 - Case files are Python scripts (`.py`) that define a dict of parameters
 - Validated with `./mfc.sh validate case.py`

.coderabbit.yaml

@@ -1,15 +1,3 @@
 # Spencer H. Bryngelson (sbryngelson) is the
 # default owner of all files in the repository.
 *                  @sbryngelson
-
-# @core-devs
-src/               @MFlowCode/core-devs
-docs/              @MFlowCode/core-devs
-toolchain/         @MFlowCode/core-devs
-tests/             @MFlowCode/core-devs
-benchmarks/        @MFlowCode/core-devs
-mfc.sh             @MFlowCode/core-devs
-CMakeLists.txt     @MFlowCode/core-devs
-
-# @physics-devs
-src/               @MFlowCode/physics-devs

.github/CODEOWNERS

@@ -8,17 +8,20 @@ ARG FC_COMPILER
 ARG COMPILER_PATH
 ARG COMPILER_LD_LIBRARY_PATH
 
+ENV DEBIAN_FRONTEND=noninteractive
+ENV TZ=UTC
+
 RUN apt-get update -y && \
-    apt-get install -y software-properties-common ca-certificates gnupg && \
+    apt-get install -y software-properties-common ca-certificates gnupg wget && \
     add-apt-repository ppa:deadsnakes/ppa && \
     apt-get update -y && \
-    if [ "$TARGET" != "gpu" ]; then \
+    if [ "$TARGET" = "cpu" ]; then \
         apt-get install -y \
             build-essential git make cmake gcc g++ gfortran bc \
             python3.12 python3.12-venv python3-pip \
             openmpi-bin libopenmpi-dev libfftw3-dev \
             mpich libmpich-dev; \
-    else \
+    elif [ "$TARGET" = "gpu" ]; then \
         apt-get install -y \
             build-essential git make cmake bc \
             python3.12 python3.12-venv python3-pip \
@@ -30,15 +33,16 @@ RUN apt-get update -y && \
 
 ENV OMPI_ALLOW_RUN_AS_ROOT=1
 ENV OMPI_ALLOW_RUN_AS_ROOT_CONFIRM=1
+ENV HYDRA_LAUNCHER=fork
 ENV PATH="/opt/MFC:$PATH"
 
 COPY ../ /opt/MFC
 
 ENV CC=${CC_COMPILER}
 ENV CXX=${CXX_COMPILER}
 ENV FC=${FC_COMPILER}
-ENV PATH="${COMPILER_PATH}:$PATH"
-ENV LD_LIBRARY_PATH="${COMPILER_LD_LIBRARY_PATH}:${LD_LIBRARY_PATH:-}"
+ENV PATH="${COMPILER_PATH}:/opt/mpich/bin:$PATH"
+ENV LD_LIBRARY_PATH="${COMPILER_LD_LIBRARY_PATH}:/opt/mpich/lib:${LD_LIBRARY_PATH:-}"
 
 # Pre-install numpy into the venv before mfc.sh runs, as it's required at
 # build time by several dependencies (pandas, cantera, matplotlib, etc.) that
@@ -50,14 +54,14 @@ RUN python3.12 -m venv /opt/MFC/build/venv && \
 RUN echo "TARGET=$TARGET CC=$CC_COMPILER FC=$FC_COMPILER" && \
     cd /opt/MFC && \
     if [ "$TARGET" = "gpu" ]; then \
-      ./mfc.sh build --gpu -j $(nproc); \
+      ./mfc.sh build --gpu acc -j $(nproc); \
     else \
       ./mfc.sh build -j $(nproc); \
     fi
 
 RUN cd /opt/MFC && \
     if [ "$TARGET" = "gpu" ]; then \
-      ./mfc.sh test -a --dry-run --gpu -j $(nproc); \
+      ./mfc.sh test -a --dry-run --gpu acc -j $(nproc); \
     else \
       ./mfc.sh test -a --dry-run -j $(nproc); \
     fi

.github/Dockerfile

@@ -30,6 +30,8 @@ yml: &yml
   - '.github/workflows/bench.yml'
   - '.github/workflows/test.yml'
   - '.github/workflows/formatting.yml'
+  - '.github/workflows/fp-stability.yml'
+  - '.github/workflows/convergence.yml'
 
 checkall: &checkall
   - *fortran_src

.github/file-filter.yml

@@ -1,10 +1,10 @@
 #!/bin/bash
 
-# Pre-builds all benchmark cases with --case-optimization.
-# No GPU hardware needed — compilation only.
+# Pre-builds all benchmark cases with --case-optimization using --dry-run so
+# binaries are cached before the GPU run job. No simulation is executed.
 # Can run in two modes:
 #   1. Direct (Frontier login nodes): pass cluster/device/interface as args
-#   2. Inside SLURM (Phoenix): uses $job_device/$job_interface from submit-slurm-job.sh
+#   2. Inside SLURM (Phoenix/frontier_amd): uses $job_device/$job_interface
 # Usage: bash prebuild-case-optimization.sh [<cluster> <device> <interface>]
 
 set -e
@@ -22,14 +22,18 @@ case "$cluster" in
     *) echo "ERROR: Unknown cluster '$cluster'"; exit 1 ;;
 esac
 
-source .github/scripts/clean-build.sh
-clean_build
+# Phoenix starts fresh (no prior dep build); other clusters pre-build deps via
+# build.sh first, so we must preserve them and only clean MFC target staging.
+if [ "$cluster" = "phoenix" ]; then
+    source .github/scripts/clean-build.sh
+    clean_build
+else
+    find build/staging -maxdepth 1 -regex '.*/[0-9a-f]+' -type d -exec rm -rf {} + 2>/dev/null || true
+    find build/install -maxdepth 1 -regex '.*/[0-9a-f]+' -type d -exec rm -rf {} + 2>/dev/null || true
+fi
 
 . ./mfc.sh load -c "$flag" -m g
 
-# Set GPU build flags from interface — this is always a GPU build.
-# Don't use gpu-opts.sh since $job_device may be "cpu" when submitted
-# to a CPU SLURM partition (no GPU hardware needed for compilation).
 case "$job_interface" in
     acc) gpu_opts="--gpu acc" ;;
     omp) gpu_opts="--gpu mp" ;;
@@ -38,5 +42,5 @@ esac
 
 for case in benchmarks/*/case.py; do
     echo "=== Pre-building: $case ==="
-    ./mfc.sh build -i "$case" --case-optimization $gpu_opts -j 8
+    ./mfc.sh run "$case" --case-optimization $gpu_opts -j 8 --dry-run
 done

.github/scripts/prebuild-case-optimization.sh

@@ -23,13 +23,14 @@ benchmarks=(
 
 # For Frontier/Frontier AMD: deps were fetched on the login node via --deps-only;
 # build case-optimized binaries here on the compute node before running.
-# For Phoenix: prebuild-case-optimization.sh already built everything in a prior SLURM job.
+# For Phoenix and frontier_amd: prebuild-case-optimization.sh already built
+# everything in a prior SLURM job (via --dry-run), so skip the build here.
 #
 # Clean stale MFC target staging before building. On self-hosted CI runners,
 # corrupted intermediate files from a prior failed build (e.g. CCE optcg crash)
 # can persist and poison subsequent builds. Each case-opt config gets its own
 # hash-named staging dir, but install dirs and other artifacts may be stale.
-if [ "$job_cluster" != "phoenix" ]; then
+if [ "$job_cluster" != "phoenix" ] && [ "$job_cluster" != "frontier_amd" ]; then
     # Clean stale MFC target dirs (hash-named) from prior builds, but
     # preserve dependency dirs (hipfort, fftw, etc.) since the compute
     # node has no internet to re-fetch them.

.github/scripts/run_case_optimization.sh

@@ -51,7 +51,7 @@ case "$cluster" in
         compiler_flag="f"
         account="CFD154"
         job_prefix="MFC"
-        qos="develop"
+        qos="hackathon"
         extra_sbatch=""
         test_time="01:59:00"
         bench_time="01:59:00"
@@ -61,7 +61,7 @@ case "$cluster" in
         compiler_flag="famd"
         account="CFD154"
         job_prefix="MFC"
-        qos="develop"
+        qos="hackathon"
         extra_sbatch=""
         test_time="01:59:00"
         bench_time="01:59:00"
@@ -92,7 +92,7 @@ if [ "$device" = "cpu" ]; then
         frontier|frontier_amd)
             sbatch_device_opts="\
 #SBATCH -n 32
-#SBATCH -p service"
+#SBATCH -p batch"
             ;;
     esac
 elif [ "$device" = "gpu" ]; then
@@ -120,7 +120,7 @@ elif [ "$device" = "gpu" ]; then
         frontier|frontier_amd)
             sbatch_device_opts="\
 #SBATCH -n 8
-#SBATCH -p service"
+#SBATCH -p batch"
             ;;
     esac
 else

.github/scripts/submit-slurm-job.sh

@@ -0,0 +1,57 @@
+name: Convergence
+
+on:
+  push:
+    branches: [master]
+  pull_request:
+    types: [opened, synchronize, reopened, ready_for_review]
+  workflow_dispatch:
+
+env:
+  OMPI_MCA_rmaps_base_oversubscribe: 1
+
+jobs:
+  file-changes:
+    name: Detect File Changes
+    runs-on: ubuntu-latest
+    outputs:
+      checkall: ${{ steps.changes.outputs.checkall }}
+    steps:
+      - name: Clone
+        uses: actions/checkout@v4
+
+      - name: Detect Changes
+        uses: dorny/paths-filter@v3
+        id: changes
+        with:
+          filters: ".github/file-filter.yml"
+
+  convergence:
+    name: "Convergence"
+    runs-on: ubuntu-latest
+    needs: [file-changes]
+    if: >-
+      !cancelled() &&
+      needs.file-changes.result == 'success' &&
+      (needs.file-changes.outputs.checkall == 'true' || github.event_name == 'workflow_dispatch')
+    timeout-minutes: 240
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Setup Ubuntu
+        run: |
+          sudo apt update -y
+          sudo apt install -y cmake gcc g++ python3 python3-dev \
+                   openmpi-bin libopenmpi-dev
+
+      - name: Setup Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Build MFC
+        run: ./mfc.sh build -j 4
+
+      - name: Run convergence tests
+        run: ./mfc.sh test --only Convergence --no-build -j 1