Swapnil/eos table #99

shankar-1729 · 2024-09-12T20:14:16Z

Merge the EOS table branch with the development branch.

…ipt in anticipation of changing them based on opacities. Set Vphase to not evolve in time.

…file

…ro meaning its time derivative its zero after it is used in collision term

…test

…e hamiltonians

…ution

…collision_CFL / ( c * max_IMFP )

…e code do collisions.

…ntials

…even if max_adaptive_speedup>1 is not satisfy

…wapnil/eos_table Jenkins changes by Sherwood

…/eos_table Fix initial huge memory allocation issue

…U-arrays and sets them to corresponding MultiFabs

The function does either outer_boundary or inner_boundary depending on which template parameter is true and which is false

We set (x,y,z) coordinates face-centered based on direction, and also set Vphase based on direction.

…conflicts resolution The merge conflict for Source/Evolve.cpp needs to be fixed. WARNING: code won't compile in the current state.

The conflict occurred in the assignment of opacities into the opacity matrices.

…on with the electron neutrino and antineutrino values interpolated from the equation of state table.

…5_CUDA Use of NUM_FLAVORS = 3 currently fails

…stability and Collisions to equilibrium test Resolve merge conflict in sample_inputs/inputs_coll_equi_test

srichers · 2024-09-27T18:53:25Z

Source/Evolve.cpp

+        //printf("(Evolve.cpp) number of particles: np = %d \n", np);
+        FlavoredNeutrinoContainer::ParticleType* pstruct = &(pti.GetArrayOfStructs()[0]);
+
+        amrex::ParallelFor (np, [=] AMREX_GPU_DEVICE (int i) {


This loop looks like it does nothing. Should we remove it?

I have cleaned this loop and commented it out. But I would like to keep the commented code since it can be used for printing values for each particle in case particle debug is needed in the future.

srichers · 2024-09-27T18:54:17Z

Source/Evolve.cpp

+            EOS_tabulated_obj.get_mue_muhat(rho, temperature, Ye, mue_out, muhat_out, keyerr, anyerr);
+            if (anyerr) assert(0); //If there is an error in interpolation call, stop execution. 
+
+//#define DEBUG_INTERPOLATION_TABLES


Are we certain the tables interpolate correctly now? Should we remove this debug output?

I would keep it. It is not being compiled, so it won't affect the performance in any way. The debug output can be enabled by uncommenting //#define DEBUG_INTERPOLATION_TABLES, which can be very useful when debugging later on. But let me know if you would like this to be removed.

srichers · 2024-09-27T18:55:13Z

Source/FlavoredNeutrinoContainerInit.cpp

@@ -290,6 +290,7 @@ InitParticles(const TestParams* parms)

 	    if(parms->IMFP_method == 1){
 			p.rdata(PIdx::Vphase) = dx[0]*dx[1]*dx[2]*4*MathConst::pi*(pow(p.rdata(PIdx::pupt)+parms->delta_E/2,3)-pow(p.rdata(PIdx::pupt)-parms->delta_E/2,3))/(3*ndirs_per_loc*parms->nppc[0]*parms->nppc[1]*parms->nppc[2]);
+			//printf("(Inside FlavoredNeutrinoContainerInit.cpp) Vphase = %g \n", p.rdata(PIdx::Vphase));


this should be removed

srichers · 2024-09-27T18:55:39Z

Source/FlavoredNeutrinoContainerInit.cpp

+      const auto lo = amrex::lbound(tile_box);
+      const auto hi = amrex::ubound(tile_box);
+
+	  //printf("tile_box = [%d, %d, %d] x [%d, %d, %d] \n", lo.x, lo.y, lo.z, hi.x, hi.y, hi.z);


remove debugging lines

srichers · 2024-09-27T18:56:26Z

Source/FlavoredNeutrinoContainerInit.cpp

+		  }
+
+		  if (!create_particle_this_cell) continue;
+		  //printf("CREATE PARTRICLE AT: i = %d, j = %d, k = %d \n", i, j, k);


remove debugging comment

srichers · 2024-09-27T19:03:42Z

Source/ReadNuLibTable.cpp

+
+    //TODO: 
+    int my_reader_process = 0; //reader_process;
+    /*if (my_reader_process < 0 || my_reader_process >= CCTK_nProcs(cctkGH))


remove debugging comment

srichers · 2024-09-27T19:04:00Z

Source/ReadNuLibTable.cpp

+              for(int k = 0; k<nye_;k++) 
+                for(int j = 0; j<ntemp_; j++) 
+        	      for(int i = 0; i<nrho_; i++) {
+        	        //int indold = i + nrho_*(j + ntemp_*(k + nye_*iv));


remove debugging comment

srichers · 2024-09-27T19:04:10Z

Source/ReadNuLibTable.cpp

+    myManagedArena.deallocate(energy_top, ngroup_);
+    //----------------------------------------------------------------------------------------------
+
+    // convert any other quantities to log. 


remove debugging comment

srichers · 2024-09-27T19:04:25Z

Source/main.cpp

@@ -72,24 +107,48 @@ void evolve_flavor(const TestParams* parms)
    const IntVect ngrow(1 + (1+shape_factor_order_vec)/2);
    for(int i=0; i<AMREX_SPACEDIM; i++) AMREX_ASSERT(parms->ncell[i] >= ngrow[i]);

+    //printf("ngrow = [%d, %d, %d] \n", ngrow[0], ngrow[1], ngrow[2]);


remove debugging comment

srichers · 2024-09-27T19:04:28Z

Source/main.cpp

    Vector<int> domain_lo_bc_types(AMREX_SPACEDIM, BCType::int_dir);
    Vector<int> domain_hi_bc_types(AMREX_SPACEDIM, BCType::int_dir);
+    //Vector<int> domain_lo_bc_types(AMREX_SPACEDIM, BCType::foextrap);


remove debugging comment

Sherwood Richers and others added 30 commits January 17, 2024 20:40

Particles store the phase space volume they occupy

56f5a20

set evolution equations for N and L to be zero in code generation scr…

f777df1

…ipt in anticipation of changing them based on opacities. Set Vphase to not evolve in time.

whitespace

f042f18

infrastructure to set IMFPs and chemical potentials in the parameter …

9251ec1

…file

redefining f --> Nrho (before was f ---> Nrho)

d5a1f63

computing volume of phase space

ce3e01f

deleting lines that convert twice mu to erg

f1adb1b

adding delta E as an input parameter to compute phase space volume

d0a0641

adding new input parameters to input files to run collitions

3baca49

adding collition term in QKE

764e827

moving line that compute derivates, now it set the phase volume to ze…

3f9ef87

…ro meaning its time derivative its zero after it is used in collision term

considering the particle that carry antineutrinos in the fast flavor …

f093668

…test

Fixing an error in the collision term

275f5a0

including adaptative time step for absortion and emission

2614233

creating vacuum paticles initial condition script

8f7b928

adding attenuation parameter to the time derivative of N_Ab due to th…

fd6964a

…e hamiltonians

solving mistake to compute time step due to collisions

00f8307

adding collision test input file

19916f7

testing if the particles converge to the expected equilibrium distrib…

8105f03

…ution

solving issue to copy GNUmakefile_jenkins to Exec directory

2507f0f

doing fast compilation in collision test

0c1708d

clarifying better the collision time steps, now delta t collision is …

95fdc84

…collision_CFL / ( c * max_IMFP )

deletting renomarlization function. It will not be used anymore

1b8a938

solving mistake in compilation comand to compile the code for 3 flavors

a6d3255

Setting default collision time step to infinite. It will change if th…

5b12e39

…e code do collisions.

setting right indentation

b24f1f5

including attenuation parameter to compute the time steps due to pote…

f8338bd

…ntials

computing minimum between time steps due to potentials and collision …

1d925c3

…even if max_adaptive_speedup>1 is not satisfy

solve mistake in chemical potential units in coment

618e685

add coment on the energy units

a804c0a

shankar-1729 and others added 22 commits June 17, 2024 13:39

Merge branch 'swapnil/eos_table' of github.com:AMReX-Astro/Emu into s…

432f894

…wapnil/eos_table Jenkins changes by Sherwood

Merge branch 'development' of github.com:AMReX-Astro/Emu into swapnil…

c3d0057

…/eos_table Fix initial huge memory allocation issue

Implement support for NuLib HDF5 tables

6d5f016

Determine energy bin automatically from given neutrino energy

fd61790

add mue,muhat interpolation function + general cleanup

1947ec1

Read EoS and NuLib table names from parfile

4b57f8b

Finish application of EoS/NuLib tables in evolve.cpp

cf78f14

Add support routine set_rho_T_Ye that reads values from HDF5 table GP…

8937f11

…U-arrays and sets them to corresponding MultiFabs

Update Jenkinsfile

482da20

Read table names only when IMFP_method is 2

f5bc8b0

Add new makefile GNUmakefile_jenkins_HDF5_CUDA and update Jenkinsfile

96dff14

Implement function to continuously add particles at outer boundary

9276ece

Make the function to add particles at boundary templated

c043dd5

The function does either outer_boundary or inner_boundary depending on which template parameter is true and which is false

Reimplement particle creation at outer boundary

8f842f1

We set (x,y,z) coordinates face-centered based on direction, and also set Vphase based on direction.

Pull changes from development branch to eos_table branch, with merge …

ec7e4ce

…conflicts resolution The merge conflict for Source/Evolve.cpp needs to be fixed. WARNING: code won't compile in the current state.

Resolving conflict between development and Swapnil's EOS branch.

f694f71

The conflict occurred in the assignment of opacities into the opacity matrices.

Clean up the code and add comments

cb7512c

Filling the chemical potential matrix to be used in the QKE calculati…

dd85884

…on with the electron neutrino and antineutrino values interpolated from the equation of state table.

Do not create particles at outer boundary when periodic BC is used

8741901

Use NUM_FLAVORS = 2 instead of 3 in makefiles/GNUmakefile_jenkins_HDF…

b683cc1

…5_CUDA Use of NUM_FLAVORS = 3 currently fails

Remove ../Scripts/babysitting/avgfee_HDF5.py test from Jenkinsfile

6c967a8

use GNUmakefile_jenkins_HDF5_CUDA in Jenkinsfile Collisions flavor in…

9f77126

…stability and Collisions to equilibrium test Resolve merge conflict in sample_inputs/inputs_coll_equi_test

shankar-1729 requested a review from erickurquilla1999 September 12, 2024 20:15

shankar-1729 added 3 commits September 16, 2024 16:22

Use actual pup* and dt for Vphase calculation

9cccd62

Fix particle creation at boundary and calculate total Vphase

6db027b

Set correct BC_type (should be moved to parameter file)

0aa2559

erickurquilla1999 approved these changes Sep 27, 2024

View reviewed changes

erickurquilla1999 merged commit cdab6c5 into development Sep 27, 2024
2 checks passed

erickurquilla1999 deleted the swapnil/eos_table branch September 27, 2024 18:50

srichers reviewed Sep 27, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Swapnil/eos table #99

Swapnil/eos table #99

shankar-1729 commented Sep 12, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

srichers Sep 27, 2024

shankar-1729 Sep 30, 2024

Swapnil/eos table #99

Swapnil/eos table #99

Conversation

shankar-1729 commented Sep 12, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment