Merge branch 'main' of github.com:amrex-astro/initial_models

.github/workflows/massive_star.yml

@@ -45,7 +45,7 @@ jobs:
       - name: Run with regular grid
         run: |
           cd massive_star
-          ./initialmodel1d.gnu.ex inputs
+          ./initialmodel1d.gnu.ex inputs problem.nx=16384
 
       - name: Compare to stored output
         run: |

.github/workflows/sub_chandra.yml

@@ -50,7 +50,7 @@ jobs:
       - name: Compare to stored output
         run: |
           cd sub_chandra
-          diff sub_chandra.M_WD-1.10.M_He-0.050.hse.CO.N14.10.00km ci-benchmarks/sub_chandra.M_WD-1.10.M_He-0.050.hse.CO.N14.10.00km
+          diff sub_chandra.M_WD-1.10.M_He-0.050.delta50.00km.hse.CO.N14.10.00km ci-benchmarks/sub_chandra.M_WD-1.10.M_He-0.050.delta50.00km.hse.CO.N14.10.00km
 
 
 

.gitignore

@@ -12,3 +12,5 @@ actual_network.F90
 build/
 
 tmp_build_dir
+
+nse*.tbl

massive_star/GNUmakefile

@@ -17,7 +17,7 @@ USE_CXX_REACTIONS = TRUE
 
 USE_FORT_MICROPHYSICS = FALSE
 
-MAX_ZONES := 16384
+MAX_ZONES := 65536
 
 DEFINES += -DNPTS_MODEL=$(MAX_ZONES)
 

massive_star/README.md

@@ -1,10 +1,41 @@
+# massive star initial model
+
 This is the initial model routine for the massive star problem.
 
-The file: 15m_500_sec.txt is a 15 solar mass MESA initial model, with
+The file: `15m_500_sec.txt` is a 15 solar mass MESA initial model, with
 an aprox21 composition.
 
 It can be converted into the subset of nuclei for aprox19 via
-convert_21_to_19.py
+`convert_21_to_19.py`
 
 This setup uses Ye as the primary composition variable from the initial
 model in regions that are in NSE.
+
+Note: you should ensure that the NSE conditions in the inputs file match
+those of your simulation, so the model will be properly in HSE.
+
+Also note that when running with > 16384 zones, you need to do:
+
+```
+ulimit -s 32768
+```
+
+since the arrays are put on the stack.
+
+## Overall algorithm
+
+The basic HSE algorithm proceeds as:
+
+* Map the raw MESA model onto a uniform grid
+
+* Reset the composition of any zones that are in NSE using the NSE
+  table
+
+* (optional) If the first MESA r is at a larger radius than the
+  innermost uniform model zone, integrate inward from the first MESA
+  point at constant entropy to find the central density in HSE at our
+  model resolution.  This is only necessary if the MESA model is very
+  low resolution.
+
+* Integrate outward from the center taking T and composition from the
+  MESA model and enforcing HSE and NSE with our EOS.