64-BitBrainstorm_2026/AMSS-NCKU
8
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add safe BSSN-EScalar kernel and transfer toggles

Browse Source
This commit is contained in:
CGH0S7
2026-04-25 01:41:55 +08:00
parent 0f1d0de1e7
commit 0cf58176d9
6 changed files with 379 additions and 413 deletions
Download Patch File Download Diff File Expand all files Collapse all files

109
AMSS_NCKU_source/bssnEScalar_class.C
View File

@@ -23,8 +23,14 @@ using namespace std;
#include "rungekutta4_rout.h"
#include "sommerfeld_rout.h"
#include "getnp4.h"
#include "shellfunctions.h"
#include "parameters.h"
#include "shellfunctions.h"
#include "parameters.h"
#if BSSN_USE_ESCALAR_C_KERNEL
#define BSSN_ESCALAR_RHS f_compute_rhs_bssn_escalar_c
#else
#define BSSN_ESCALAR_RHS f_compute_rhs_bssn_escalar
#endif
#ifdef With_AHF
#include "derivatives.h"
@@ -169,13 +175,7 @@ void bssnEScalar_class::Initialize()
Setup_Black_Hole_position();
}
// BSSN-EScalar currently uses the uncached communication fallback paths.
sync_cache_pre = 0;
sync_cache_cor = 0;
sync_cache_rp_coarse = 0;
sync_cache_rp_fine = 0;
sync_cache_restrict = 0;
sync_cache_outbd = 0;
setup_transfer_caches();
}
//================================================================================================
@@ -345,11 +345,13 @@ void bssnEScalar_class::Read_Ansorg()
}
inf.close();
}
int order = 6;
Ansorg read_ansorg("Ansorg.psid", order);
// set initial data
for (int lev = 0; lev < GH->levels; lev++)
{
int order = 6;
Ansorg read_ansorg("Ansorg.psid", order);
if (myrank == 0)
cout << "[debug] Read_Ansorg: Ansorg object ready" << endl;
// set initial data
for (int lev = 0; lev < GH->levels; lev++)
{
MyList<Patch> *Pp = GH->PatL[lev];
while (Pp)
{
@@ -381,12 +383,14 @@ void bssnEScalar_class::Read_Ansorg()
if (BL == Pp->data->ble)
break;
BL = BL->next;
}
Pp = Pp->next;
}
}
#ifdef WithShell
// ShellPatch part
}
Pp = Pp->next;
}
if (myrank == 0)
cout << "[debug] Read_Ansorg: finished level " << lev << " patch init" << endl;
}
#ifdef WithShell
// ShellPatch part
MyList<ss_patch> *Pp = SH->PatL;
while (Pp)
{
@@ -423,15 +427,19 @@ void bssnEScalar_class::Read_Ansorg()
if (BL == Pp->data->ble)
break;
BL = BL->next;
}
Pp = Pp->next;
}
#endif
}
Pp = Pp->next;
}
if (myrank == 0)
cout << "[debug] Read_Ansorg: finished shell init" << endl;
#endif
delete[] Porg_here;
delete[] pmom_local;
delete[] spin_local;
delete[] mass_local;
if (myrank == 0)
cout << "[debug] Read_Ansorg: finished local cleanup" << endl;
// dump read_in initial data
// for(int lev=0;lev<GH->levels;lev++) Parallel::Dump_Data(GH->PatL[lev],StateList,0,PhysTime,dT);
}
@@ -762,7 +770,7 @@ void bssnEScalar_class::Step(int lev, int YN)
cg->fgfs[Ayy0->sgfn], cg->fgfs[Ayz0->sgfn], cg->fgfs[Azz0->sgfn]);
#endif
if (f_compute_rhs_bssn_escalar_c(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
if (BSSN_ESCALAR_RHS(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn],
cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],
@@ -1016,11 +1024,12 @@ void bssnEScalar_class::Step(int lev, int YN)
}
#endif
Parallel::Sync(GH->PatL[lev], SynchList_pre, Symmetry);
Parallel::AsyncSyncState async_pre;
sync_predictor_start(lev, SynchList_pre, async_pre);
#ifdef WithShell
if (lev == 0)
{
#ifdef WithShell
if (lev == 0)
{
clock_t prev_clock, curr_clock;
if (myrank == 0)
curr_clock = clock();
@@ -1032,9 +1041,10 @@ void bssnEScalar_class::Step(int lev, int YN)
cout << " Shell stuff synchronization used "
<< (double)(curr_clock - prev_clock) / ((double)CLOCKS_PER_SEC)
<< " seconds! " << endl;
}
}
#endif
}
}
#endif
sync_predictor_finish(lev, async_pre, SynchList_pre);
// for black hole position
if (BH_num > 0 && lev == GH->levels - 1)
@@ -1104,7 +1114,7 @@ void bssnEScalar_class::Step(int lev, int YN)
cg->fgfs[Ayy->sgfn], cg->fgfs[Ayz->sgfn], cg->fgfs[Azz->sgfn]);
#endif
if (f_compute_rhs_bssn_escalar_c(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
if (BSSN_ESCALAR_RHS(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
cg->fgfs[phi->sgfn], cg->fgfs[trK->sgfn],
cg->fgfs[gxx->sgfn], cg->fgfs[gxy->sgfn], cg->fgfs[gxz->sgfn],
cg->fgfs[gyy->sgfn], cg->fgfs[gyz->sgfn], cg->fgfs[gzz->sgfn],
@@ -1372,11 +1382,12 @@ void bssnEScalar_class::Step(int lev, int YN)
}
#endif
Parallel::Sync(GH->PatL[lev], SynchList_cor, Symmetry);
Parallel::AsyncSyncState async_cor;
sync_corrector_start(lev, SynchList_cor, async_cor);
#ifdef WithShell
if (lev == 0)
{
#ifdef WithShell
if (lev == 0)
{
clock_t prev_clock, curr_clock;
if (myrank == 0)
curr_clock = clock();
@@ -1388,9 +1399,10 @@ void bssnEScalar_class::Step(int lev, int YN)
cout << " Shell stuff synchronization used "
<< (double)(curr_clock - prev_clock) / ((double)CLOCKS_PER_SEC)
<< " seconds! " << endl;
}
}
#endif
}
}
#endif
sync_corrector_finish(lev, async_cor, SynchList_cor);
// for black hole position
if (BH_num > 0 && lev == GH->levels - 1)
{
@@ -1858,11 +1870,14 @@ void bssnEScalar_class::AnalysisStuff_EScalar(int lev, double dT_lev)
//================================================================================================
void bssnEScalar_class::Interp_Constraint()
{
// we do not support a_lev != 0 yet.
if (a_lev > 0)
return;
void bssnEScalar_class::Interp_Constraint(bool infg)
{
if (!infg)
return;
// we do not support a_lev != 0 yet.
if (a_lev > 0)
return;
for (int lev = 0; lev < GH->levels; lev++)
{
@@ -1881,7 +1896,7 @@ void bssnEScalar_class::Interp_Constraint()
if (myrank == cg->rank)
{
if (lev > 0)
f_compute_rhs_bssn_escalar_c(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
BSSN_ESCALAR_RHS(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn],
cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],
@@ -2101,7 +2116,7 @@ void bssnEScalar_class::Constraint_Out()
if (myrank == cg->rank)
{
if (lev > 0)
f_compute_rhs_bssn_escalar_c(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
BSSN_ESCALAR_RHS(cg->shape, TRK4, cg->X[0], cg->X[1], cg->X[2],
cg->fgfs[phi0->sgfn], cg->fgfs[trK0->sgfn],
cg->fgfs[gxx0->sgfn], cg->fgfs[gxy0->sgfn], cg->fgfs[gxz0->sgfn],
cg->fgfs[gyy0->sgfn], cg->fgfs[gyz0->sgfn], cg->fgfs[gzz0->sgfn],