diff options
Diffstat (limited to 'ML_BSSN_MP_O8/src/ML_BSSN_MP_O8_convertFromADMBase.cc')
| -rw-r--r-- | ML_BSSN_MP_O8/src/ML_BSSN_MP_O8_convertFromADMBase.cc | 175 |
1 files changed, 54 insertions, 121 deletions
diff --git a/ML_BSSN_MP_O8/src/ML_BSSN_MP_O8_convertFromADMBase.cc b/ML_BSSN_MP_O8/src/ML_BSSN_MP_O8_convertFromADMBase.cc index 667a19d..2cf4397 100644 --- a/ML_BSSN_MP_O8/src/ML_BSSN_MP_O8_convertFromADMBase.cc +++ b/ML_BSSN_MP_O8/src/ML_BSSN_MP_O8_convertFromADMBase.cc @@ -12,6 +12,7 @@ #include "cctk_Parameters.h" #include "GenericFD.h" #include "Differencing.h" +#include "cctk_Loop.h" #include "loopcontrol.h" #include "vectors.h" @@ -28,8 +29,6 @@ static void ML_BSSN_MP_O8_convertFromADMBase_Body(cGH const * restrict const cct DECLARE_CCTK_PARAMETERS; - /* Declare finite differencing variables */ - /* Include user-supplied include files */ /* Initialise finite differencing variables */ @@ -66,9 +65,9 @@ static void ML_BSSN_MP_O8_convertFromADMBase_Body(cGH const * restrict const cct CCTK_REAL_VEC const p1o12dx = kmul(INV(dx),ToReal(0.0833333333333333333333333333333)); CCTK_REAL_VEC const p1o12dy = kmul(INV(dy),ToReal(0.0833333333333333333333333333333)); CCTK_REAL_VEC const p1o12dz = kmul(INV(dz),ToReal(0.0833333333333333333333333333333)); - CCTK_REAL_VEC const p1o144dxdy = kmul(INV(dx),kmul(INV(dy),ToReal(0.00694444444444444444444444444444))); - CCTK_REAL_VEC const p1o144dxdz = kmul(INV(dx),kmul(INV(dz),ToReal(0.00694444444444444444444444444444))); - CCTK_REAL_VEC const p1o144dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.00694444444444444444444444444444))); + CCTK_REAL_VEC const p1o144dxdy = kmul(INV(kmul(dx,dy)),ToReal(0.00694444444444444444444444444444)); + CCTK_REAL_VEC const p1o144dxdz = kmul(INV(kmul(dx,dz)),ToReal(0.00694444444444444444444444444444)); + CCTK_REAL_VEC const p1o144dydz = kmul(INV(kmul(dy,dz)),ToReal(0.00694444444444444444444444444444)); CCTK_REAL_VEC const p1o1680dx = kmul(INV(dx),ToReal(0.000595238095238095238095238095238)); CCTK_REAL_VEC const p1o1680dy = kmul(INV(dy),ToReal(0.000595238095238095238095238095238)); CCTK_REAL_VEC const p1o1680dz = kmul(INV(dz),ToReal(0.000595238095238095238095238095238)); @@ -87,14 +86,14 @@ static void ML_BSSN_MP_O8_convertFromADMBase_Body(cGH const * restrict const cct CCTK_REAL_VEC const p1o2dx = kmul(INV(dx),ToReal(0.5)); CCTK_REAL_VEC const p1o2dy = kmul(INV(dy),ToReal(0.5)); CCTK_REAL_VEC const p1o2dz = kmul(INV(dz),ToReal(0.5)); - CCTK_REAL_VEC const p1o3600dxdy = kmul(INV(dx),kmul(INV(dy),ToReal(0.000277777777777777777777777777778))); - CCTK_REAL_VEC const p1o3600dxdz = kmul(INV(dx),kmul(INV(dz),ToReal(0.000277777777777777777777777777778))); - CCTK_REAL_VEC const p1o3600dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.000277777777777777777777777777778))); + CCTK_REAL_VEC const p1o3600dxdy = kmul(INV(kmul(dx,dy)),ToReal(0.000277777777777777777777777777778)); + CCTK_REAL_VEC const p1o3600dxdz = kmul(INV(kmul(dx,dz)),ToReal(0.000277777777777777777777777777778)); + CCTK_REAL_VEC const p1o3600dydz = kmul(INV(kmul(dy,dz)),ToReal(0.000277777777777777777777777777778)); CCTK_REAL_VEC const p1o4dx = kmul(INV(dx),ToReal(0.25)); - CCTK_REAL_VEC const p1o4dxdy = kmul(INV(dx),kmul(INV(dy),ToReal(0.25))); - CCTK_REAL_VEC const p1o4dxdz = kmul(INV(dx),kmul(INV(dz),ToReal(0.25))); + CCTK_REAL_VEC const p1o4dxdy = kmul(INV(kmul(dx,dy)),ToReal(0.25)); + CCTK_REAL_VEC const p1o4dxdz = kmul(INV(kmul(dx,dz)),ToReal(0.25)); CCTK_REAL_VEC const p1o4dy = kmul(INV(dy),ToReal(0.25)); - CCTK_REAL_VEC const p1o4dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.25))); + CCTK_REAL_VEC const p1o4dydz = kmul(INV(kmul(dy,dz)),ToReal(0.25)); CCTK_REAL_VEC const p1o4dz = kmul(INV(dz),ToReal(0.25)); CCTK_REAL_VEC const p1o5040dx2 = kmul(INV(SQR(dx)),ToReal(0.000198412698412698412698412698413)); CCTK_REAL_VEC const p1o5040dy2 = kmul(INV(SQR(dy)),ToReal(0.000198412698412698412698412698413)); @@ -108,9 +107,9 @@ static void ML_BSSN_MP_O8_convertFromADMBase_Body(cGH const * restrict const cct CCTK_REAL_VEC const p1o64dx = kmul(INV(dx),ToReal(0.015625)); CCTK_REAL_VEC const p1o64dy = kmul(INV(dy),ToReal(0.015625)); CCTK_REAL_VEC const p1o64dz = kmul(INV(dz),ToReal(0.015625)); - CCTK_REAL_VEC const p1o705600dxdy = kmul(INV(dx),kmul(INV(dy),ToReal(1.41723356009070294784580498866e-6))); - CCTK_REAL_VEC const p1o705600dxdz = kmul(INV(dx),kmul(INV(dz),ToReal(1.41723356009070294784580498866e-6))); - CCTK_REAL_VEC const p1o705600dydz = kmul(INV(dy),kmul(INV(dz),ToReal(1.41723356009070294784580498866e-6))); + CCTK_REAL_VEC const p1o705600dxdy = kmul(INV(kmul(dx,dy)),ToReal(1.41723356009070294784580498866e-6)); + CCTK_REAL_VEC const p1o705600dxdz = kmul(INV(kmul(dx,dz)),ToReal(1.41723356009070294784580498866e-6)); + CCTK_REAL_VEC const p1o705600dydz = kmul(INV(kmul(dy,dz)),ToReal(1.41723356009070294784580498866e-6)); CCTK_REAL_VEC const p1o840dx = kmul(INV(dx),ToReal(0.00119047619047619047619047619048)); CCTK_REAL_VEC const p1o840dy = kmul(INV(dy),ToReal(0.00119047619047619047619047619048)); CCTK_REAL_VEC const p1o840dz = kmul(INV(dz),ToReal(0.00119047619047619047619047619048)); @@ -191,7 +190,7 @@ static void ML_BSSN_MP_O8_convertFromADMBase_Body(cGH const * restrict const cct /* Loop over the grid points */ #pragma omp parallel - LC_LOOP3VEC (ML_BSSN_MP_O8_convertFromADMBase, + LC_LOOP3VEC(ML_BSSN_MP_O8_convertFromADMBase, i,j,k, imin[0],imin[1],imin[2], imax[0],imax[1],imax[2], cctk_lsh[0],cctk_lsh[1],cctk_lsh[2], CCTK_REAL_VEC_SIZE) @@ -296,25 +295,25 @@ static void ML_BSSN_MP_O8_convertFromADMBase_Body(cGH const * restrict const cct CCTK_REAL_VEC gt33L = kmul(em4phi,g33); trKL = - kmadd(gu11,kxxL,kmadd(gu22,kyyL,kmadd(gu33,kzzL,kmul(kmadd(gu12,kxyL,kmadd(gu13,kxzL,kmul(gu23,kyzL))),ToReal(2))))); + kmadd(kxxL,gu11,kmadd(kyyL,gu22,kmadd(kzzL,gu33,kmul(kmadd(kxyL,gu12,kmadd(kxzL,gu13,kmul(kyzL,gu23))),ToReal(2))))); CCTK_REAL_VEC At11L = - kmul(em4phi,kmadd(g11,kmul(trKL,ToReal(-0.333333333333333333333333333333)),kxxL)); + kmul(em4phi,kmadd(trKL,kmul(g11,ToReal(-0.333333333333333333333333333333)),kxxL)); CCTK_REAL_VEC At12L = - kmul(em4phi,kmadd(g12,kmul(trKL,ToReal(-0.333333333333333333333333333333)),kxyL)); + kmul(em4phi,kmadd(trKL,kmul(g12,ToReal(-0.333333333333333333333333333333)),kxyL)); CCTK_REAL_VEC At13L = - kmul(em4phi,kmadd(g13,kmul(trKL,ToReal(-0.333333333333333333333333333333)),kxzL)); + kmul(em4phi,kmadd(trKL,kmul(g13,ToReal(-0.333333333333333333333333333333)),kxzL)); CCTK_REAL_VEC At22L = - kmul(em4phi,kmadd(g22,kmul(trKL,ToReal(-0.333333333333333333333333333333)),kyyL)); + kmul(em4phi,kmadd(trKL,kmul(g22,ToReal(-0.333333333333333333333333333333)),kyyL)); CCTK_REAL_VEC At23L = - kmul(em4phi,kmadd(g23,kmul(trKL,ToReal(-0.333333333333333333333333333333)),kyzL)); + kmul(em4phi,kmadd(trKL,kmul(g23,ToReal(-0.333333333333333333333333333333)),kyzL)); CCTK_REAL_VEC At33L = - kmul(em4phi,kmadd(g33,kmul(trKL,ToReal(-0.333333333333333333333333333333)),kzzL)); + kmul(em4phi,kmadd(trKL,kmul(g33,ToReal(-0.333333333333333333333333333333)),kzzL)); CCTK_REAL_VEC alphaL = alpL; @@ -324,104 +323,28 @@ static void ML_BSSN_MP_O8_convertFromADMBase_Body(cGH const * restrict const cct CCTK_REAL_VEC beta3L = betazL; - /* If necessary, store only partial vectors after the first iteration */ - - if (CCTK_REAL_VEC_SIZE > 2 && CCTK_BUILTIN_EXPECT(i < lc_imin && i+CCTK_REAL_VEC_SIZE > lc_imax, 0)) - { - ptrdiff_t const elt_count_lo = lc_imin-i; - ptrdiff_t const elt_count_hi = lc_imax-i; - vec_store_nta_partial_mid(alpha[index],alphaL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(At11[index],At11L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(At12[index],At12L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(At13[index],At13L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(At22[index],At22L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(At23[index],At23L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(At33[index],At33L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(beta1[index],beta1L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(beta2[index],beta2L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(beta3[index],beta3L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gt11[index],gt11L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gt12[index],gt12L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gt13[index],gt13L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gt22[index],gt22L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gt23[index],gt23L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gt33[index],gt33L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(phi[index],phiL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(trK[index],trKL,elt_count_lo,elt_count_hi); - break; - } - - /* If necessary, store only partial vectors after the first iteration */ - - if (CCTK_REAL_VEC_SIZE > 1 && CCTK_BUILTIN_EXPECT(i < lc_imin, 0)) - { - ptrdiff_t const elt_count = lc_imin-i; - vec_store_nta_partial_hi(alpha[index],alphaL,elt_count); - vec_store_nta_partial_hi(At11[index],At11L,elt_count); - vec_store_nta_partial_hi(At12[index],At12L,elt_count); - vec_store_nta_partial_hi(At13[index],At13L,elt_count); - vec_store_nta_partial_hi(At22[index],At22L,elt_count); - vec_store_nta_partial_hi(At23[index],At23L,elt_count); - vec_store_nta_partial_hi(At33[index],At33L,elt_count); - vec_store_nta_partial_hi(beta1[index],beta1L,elt_count); - vec_store_nta_partial_hi(beta2[index],beta2L,elt_count); - vec_store_nta_partial_hi(beta3[index],beta3L,elt_count); - vec_store_nta_partial_hi(gt11[index],gt11L,elt_count); - vec_store_nta_partial_hi(gt12[index],gt12L,elt_count); - vec_store_nta_partial_hi(gt13[index],gt13L,elt_count); - vec_store_nta_partial_hi(gt22[index],gt22L,elt_count); - vec_store_nta_partial_hi(gt23[index],gt23L,elt_count); - vec_store_nta_partial_hi(gt33[index],gt33L,elt_count); - vec_store_nta_partial_hi(phi[index],phiL,elt_count); - vec_store_nta_partial_hi(trK[index],trKL,elt_count); - continue; - } - - /* If necessary, store only partial vectors after the last iteration */ - - if (CCTK_REAL_VEC_SIZE > 1 && CCTK_BUILTIN_EXPECT(i+CCTK_REAL_VEC_SIZE > lc_imax, 0)) - { - ptrdiff_t const elt_count = lc_imax-i; - vec_store_nta_partial_lo(alpha[index],alphaL,elt_count); - vec_store_nta_partial_lo(At11[index],At11L,elt_count); - vec_store_nta_partial_lo(At12[index],At12L,elt_count); - vec_store_nta_partial_lo(At13[index],At13L,elt_count); - vec_store_nta_partial_lo(At22[index],At22L,elt_count); - vec_store_nta_partial_lo(At23[index],At23L,elt_count); - vec_store_nta_partial_lo(At33[index],At33L,elt_count); - vec_store_nta_partial_lo(beta1[index],beta1L,elt_count); - vec_store_nta_partial_lo(beta2[index],beta2L,elt_count); - vec_store_nta_partial_lo(beta3[index],beta3L,elt_count); - vec_store_nta_partial_lo(gt11[index],gt11L,elt_count); - vec_store_nta_partial_lo(gt12[index],gt12L,elt_count); - vec_store_nta_partial_lo(gt13[index],gt13L,elt_count); - vec_store_nta_partial_lo(gt22[index],gt22L,elt_count); - vec_store_nta_partial_lo(gt23[index],gt23L,elt_count); - vec_store_nta_partial_lo(gt33[index],gt33L,elt_count); - vec_store_nta_partial_lo(phi[index],phiL,elt_count); - vec_store_nta_partial_lo(trK[index],trKL,elt_count); - break; - } - vec_store_nta(alpha[index],alphaL); - vec_store_nta(At11[index],At11L); - vec_store_nta(At12[index],At12L); - vec_store_nta(At13[index],At13L); - vec_store_nta(At22[index],At22L); - vec_store_nta(At23[index],At23L); - vec_store_nta(At33[index],At33L); - vec_store_nta(beta1[index],beta1L); - vec_store_nta(beta2[index],beta2L); - vec_store_nta(beta3[index],beta3L); - vec_store_nta(gt11[index],gt11L); - vec_store_nta(gt12[index],gt12L); - vec_store_nta(gt13[index],gt13L); - vec_store_nta(gt22[index],gt22L); - vec_store_nta(gt23[index],gt23L); - vec_store_nta(gt33[index],gt33L); - vec_store_nta(phi[index],phiL); - vec_store_nta(trK[index],trKL); + /* Copy local copies back to grid functions */ + vec_store_partial_prepare(i,lc_imin,lc_imax); + vec_store_nta_partial(alpha[index],alphaL); + vec_store_nta_partial(At11[index],At11L); + vec_store_nta_partial(At12[index],At12L); + vec_store_nta_partial(At13[index],At13L); + vec_store_nta_partial(At22[index],At22L); + vec_store_nta_partial(At23[index],At23L); + vec_store_nta_partial(At33[index],At33L); + vec_store_nta_partial(beta1[index],beta1L); + vec_store_nta_partial(beta2[index],beta2L); + vec_store_nta_partial(beta3[index],beta3L); + vec_store_nta_partial(gt11[index],gt11L); + vec_store_nta_partial(gt12[index],gt12L); + vec_store_nta_partial(gt13[index],gt13L); + vec_store_nta_partial(gt22[index],gt22L); + vec_store_nta_partial(gt23[index],gt23L); + vec_store_nta_partial(gt33[index],gt33L); + vec_store_nta_partial(phi[index],phiL); + vec_store_nta_partial(trK[index],trKL); } - LC_ENDLOOP3VEC (ML_BSSN_MP_O8_convertFromADMBase); + LC_ENDLOOP3VEC(ML_BSSN_MP_O8_convertFromADMBase); } extern "C" void ML_BSSN_MP_O8_convertFromADMBase(CCTK_ARGUMENTS) @@ -440,7 +363,17 @@ extern "C" void ML_BSSN_MP_O8_convertFromADMBase(CCTK_ARGUMENTS) return; } - const char *groups[] = {"ADMBase::curv","ADMBase::lapse","ADMBase::metric","ADMBase::shift","ML_BSSN_MP_O8::ML_curv","ML_BSSN_MP_O8::ML_lapse","ML_BSSN_MP_O8::ML_log_confac","ML_BSSN_MP_O8::ML_metric","ML_BSSN_MP_O8::ML_shift","ML_BSSN_MP_O8::ML_trace_curv"}; + const char *const groups[] = { + "ADMBase::curv", + "ADMBase::lapse", + "ADMBase::metric", + "ADMBase::shift", + "ML_BSSN_MP_O8::ML_curv", + "ML_BSSN_MP_O8::ML_lapse", + "ML_BSSN_MP_O8::ML_log_confac", + "ML_BSSN_MP_O8::ML_metric", + "ML_BSSN_MP_O8::ML_shift", + "ML_BSSN_MP_O8::ML_trace_curv"}; GenericFD_AssertGroupStorage(cctkGH, "ML_BSSN_MP_O8_convertFromADMBase", 10, groups); switch(fdOrder) @@ -458,7 +391,7 @@ extern "C" void ML_BSSN_MP_O8_convertFromADMBase(CCTK_ARGUMENTS) break; } - GenericFD_LoopOverEverything(cctkGH, &ML_BSSN_MP_O8_convertFromADMBase_Body); + GenericFD_LoopOverEverything(cctkGH, ML_BSSN_MP_O8_convertFromADMBase_Body); if (verbose > 1) { |