From f87a82b998a429bcc88c928a06e175b633a916cc Mon Sep 17 00:00:00 2001 From: Barry Wardell Date: Fri, 27 Apr 2012 12:33:29 +0100 Subject: Regenerate thorns. --- ML_ADM/schedule.ccl | 60 +++++++++-- ML_ADM/src/ML_ADM_Minkowski.cc | 144 +++++++------------------ ML_ADM/src/ML_ADM_RHS.cc | 148 ++++++++------------------ ML_ADM/src/ML_ADM_boundary.cc | 144 +++++++------------------ ML_ADM/src/ML_ADM_constraints.cc | 84 +++++---------- ML_ADM/src/ML_ADM_constraints_boundary.cc | 82 +++++--------- ML_ADM/src/ML_ADM_convertFromADMBase.cc | 148 ++++++++------------------ ML_ADM/src/ML_ADM_convertToADMBase.cc | 170 +++++++++--------------------- ML_ADM/src/make.code.defn | 2 +- 9 files changed, 313 insertions(+), 669 deletions(-) (limited to 'ML_ADM') diff --git a/ML_ADM/schedule.ccl b/ML_ADM/schedule.ccl index 1745ad3..c5c8a2b 100644 --- a/ML_ADM/schedule.ccl +++ b/ML_ADM/schedule.ccl @@ -121,12 +121,6 @@ schedule ML_ADM_Startup at STARTUP OPTIONS: meta } "create banner" -schedule ML_ADM_RegisterVars in MoL_Register -{ - LANG: C - OPTIONS: meta -} "Register Variables for MoL" - schedule ML_ADM_RegisterSymmetries in SymmetryRegister { LANG: C @@ -139,6 +133,10 @@ if (CCTK_EQUALS(my_initial_data, "Minkowski")) schedule ML_ADM_Minkowski IN ADMBase_InitialData { LANG: C + WRITES: ML_ADM::ML_curv + WRITES: ML_ADM::ML_lapse + WRITES: ML_ADM::ML_metric + WRITES: ML_ADM::ML_shift } "ML_ADM_Minkowski" } @@ -148,12 +146,28 @@ if (CCTK_EQUALS(my_initial_data, "ADMBase")) schedule ML_ADM_convertFromADMBase AT initial AFTER ADMBase_PostInitial { LANG: C + READS: ADMBase::curv + READS: ADMBase::lapse + READS: ADMBase::metric + READS: ADMBase::shift + WRITES: ML_ADM::ML_curv + WRITES: ML_ADM::ML_lapse + WRITES: ML_ADM::ML_metric + WRITES: ML_ADM::ML_shift } "ML_ADM_convertFromADMBase" } schedule ML_ADM_RHS IN MoL_CalcRHS { LANG: C + READS: ML_ADM::ML_curv + READS: ML_ADM::ML_lapse + READS: ML_ADM::ML_metric + READS: ML_ADM::ML_shift + WRITES: ML_ADM::ML_curvrhs + WRITES: ML_ADM::ML_lapserhs + WRITES: ML_ADM::ML_metricrhs + WRITES: ML_ADM::ML_shiftrhs } "ML_ADM_RHS" schedule ML_ADM_RHS AT analysis @@ -163,6 +177,14 @@ schedule ML_ADM_RHS AT analysis SYNC: ML_lapserhs SYNC: ML_metricrhs SYNC: ML_shiftrhs + READS: ML_ADM::ML_curv + READS: ML_ADM::ML_lapse + READS: ML_ADM::ML_metric + READS: ML_ADM::ML_shift + WRITES: ML_ADM::ML_curvrhs + WRITES: ML_ADM::ML_lapserhs + WRITES: ML_ADM::ML_metricrhs + WRITES: ML_ADM::ML_shiftrhs } "ML_ADM_RHS" @@ -171,12 +193,26 @@ if (CCTK_EQUALS(my_boundary_condition, "Minkowski")) schedule ML_ADM_boundary IN MoL_PostStep { LANG: C + WRITES: ML_ADM::ML_curv + WRITES: ML_ADM::ML_lapse + WRITES: ML_ADM::ML_metric + WRITES: ML_ADM::ML_shift } "ML_ADM_boundary" } schedule ML_ADM_convertToADMBase IN MoL_PostStep AFTER ML_ADM_ApplyBCs { LANG: C + READS: ML_ADM::ML_curv + READS: ML_ADM::ML_lapse + READS: ML_ADM::ML_metric + READS: ML_ADM::ML_shift + WRITES: ADMBase::curv + WRITES: ADMBase::dtlapse + WRITES: ADMBase::dtshift + WRITES: ADMBase::lapse + WRITES: ADMBase::metric + WRITES: ADMBase::shift } "ML_ADM_convertToADMBase" schedule ML_ADM_constraints AT analysis @@ -186,11 +222,17 @@ schedule ML_ADM_constraints AT analysis SYNC: ML_mom TRIGGERS: ML_Ham TRIGGERS: ML_mom + READS: ML_ADM::ML_curv + READS: ML_ADM::ML_metric + WRITES: ML_ADM::ML_Ham + WRITES: ML_ADM::ML_mom } "ML_ADM_constraints" schedule ML_ADM_constraints_boundary AT analysis AFTER ML_ADM_constraints { LANG: C + WRITES: ML_ADM::ML_Ham + WRITES: ML_ADM::ML_mom } "ML_ADM_constraints_boundary" schedule ML_ADM_SelectBoundConds in MoL_PostStep @@ -209,6 +251,12 @@ schedule ML_ADM_CheckBoundaries at BASEGRID OPTIONS: meta } "check boundaries treatment" +schedule ML_ADM_RegisterVars in MoL_Register +{ + LANG: C + OPTIONS: meta +} "Register Variables for MoL" + schedule group ApplyBCs as ML_ADM_ApplyBCs in MoL_PostStep after ML_ADM_SelectBoundConds { # no language specified diff --git a/ML_ADM/src/ML_ADM_Minkowski.cc b/ML_ADM/src/ML_ADM_Minkowski.cc index f220dd2..b201596 100644 --- a/ML_ADM/src/ML_ADM_Minkowski.cc +++ b/ML_ADM/src/ML_ADM_Minkowski.cc @@ -29,8 +29,6 @@ static void ML_ADM_Minkowski_Body(cGH const * restrict const cctkGH, int const d DECLARE_CCTK_PARAMETERS; - /* Declare finite differencing variables */ - /* Include user-supplied include files */ /* Initialise finite differencing variables */ @@ -61,30 +59,30 @@ static void ML_ADM_Minkowski_Body(cGH const * restrict const cctkGH, int const d 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 p1o180dx2 = kmul(INV(SQR(dx)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dy2 = kmul(INV(SQR(dy)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dz2 = kmul(INV(SQR(dz)),ToReal(0.00555555555555555555555555555556)); 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 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 p1o4dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.25))); + 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 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 p1o4dydz = kmul(INV(kmul(dy,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)); CCTK_REAL_VEC const p1o5040dz2 = kmul(INV(SQR(dz)),ToReal(0.000198412698412698412698412698413)); CCTK_REAL_VEC const p1o60dx = kmul(INV(dx),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dy = kmul(INV(dy),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dz = kmul(INV(dz),ToReal(0.0166666666666666666666666666667)); - 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)); @@ -150,7 +148,7 @@ static void ML_ADM_Minkowski_Body(cGH const * restrict const cctkGH, int const d /* Loop over the grid points */ #pragma omp parallel - LC_LOOP3VEC (ML_ADM_Minkowski, + LC_LOOP3VEC(ML_ADM_Minkowski, 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) @@ -214,96 +212,26 @@ static void ML_ADM_Minkowski_Body(cGH const * restrict const cctkGH, int const d CCTK_REAL_VEC beta3L = ToReal(0); - /* 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(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(g11[index],g11L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g12[index],g12L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g13[index],g13L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g22[index],g22L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g23[index],g23L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g33[index],g33L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K11[index],K11L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K12[index],K12L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K13[index],K13L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K22[index],K22L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K23[index],K23L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K33[index],K33L,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(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(g11[index],g11L,elt_count); - vec_store_nta_partial_hi(g12[index],g12L,elt_count); - vec_store_nta_partial_hi(g13[index],g13L,elt_count); - vec_store_nta_partial_hi(g22[index],g22L,elt_count); - vec_store_nta_partial_hi(g23[index],g23L,elt_count); - vec_store_nta_partial_hi(g33[index],g33L,elt_count); - vec_store_nta_partial_hi(K11[index],K11L,elt_count); - vec_store_nta_partial_hi(K12[index],K12L,elt_count); - vec_store_nta_partial_hi(K13[index],K13L,elt_count); - vec_store_nta_partial_hi(K22[index],K22L,elt_count); - vec_store_nta_partial_hi(K23[index],K23L,elt_count); - vec_store_nta_partial_hi(K33[index],K33L,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(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(g11[index],g11L,elt_count); - vec_store_nta_partial_lo(g12[index],g12L,elt_count); - vec_store_nta_partial_lo(g13[index],g13L,elt_count); - vec_store_nta_partial_lo(g22[index],g22L,elt_count); - vec_store_nta_partial_lo(g23[index],g23L,elt_count); - vec_store_nta_partial_lo(g33[index],g33L,elt_count); - vec_store_nta_partial_lo(K11[index],K11L,elt_count); - vec_store_nta_partial_lo(K12[index],K12L,elt_count); - vec_store_nta_partial_lo(K13[index],K13L,elt_count); - vec_store_nta_partial_lo(K22[index],K22L,elt_count); - vec_store_nta_partial_lo(K23[index],K23L,elt_count); - vec_store_nta_partial_lo(K33[index],K33L,elt_count); - break; - } - vec_store_nta(alpha[index],alphaL); - vec_store_nta(beta1[index],beta1L); - vec_store_nta(beta2[index],beta2L); - vec_store_nta(beta3[index],beta3L); - vec_store_nta(g11[index],g11L); - vec_store_nta(g12[index],g12L); - vec_store_nta(g13[index],g13L); - vec_store_nta(g22[index],g22L); - vec_store_nta(g23[index],g23L); - vec_store_nta(g33[index],g33L); - vec_store_nta(K11[index],K11L); - vec_store_nta(K12[index],K12L); - vec_store_nta(K13[index],K13L); - vec_store_nta(K22[index],K22L); - vec_store_nta(K23[index],K23L); - vec_store_nta(K33[index],K33L); + /* 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(beta1[index],beta1L); + vec_store_nta_partial(beta2[index],beta2L); + vec_store_nta_partial(beta3[index],beta3L); + vec_store_nta_partial(g11[index],g11L); + vec_store_nta_partial(g12[index],g12L); + vec_store_nta_partial(g13[index],g13L); + vec_store_nta_partial(g22[index],g22L); + vec_store_nta_partial(g23[index],g23L); + vec_store_nta_partial(g33[index],g33L); + vec_store_nta_partial(K11[index],K11L); + vec_store_nta_partial(K12[index],K12L); + vec_store_nta_partial(K13[index],K13L); + vec_store_nta_partial(K22[index],K22L); + vec_store_nta_partial(K23[index],K23L); + vec_store_nta_partial(K33[index],K33L); } - LC_ENDLOOP3VEC (ML_ADM_Minkowski); + LC_ENDLOOP3VEC(ML_ADM_Minkowski); } extern "C" void ML_ADM_Minkowski(CCTK_ARGUMENTS) @@ -322,7 +250,11 @@ extern "C" void ML_ADM_Minkowski(CCTK_ARGUMENTS) return; } - const char *groups[] = {"ML_ADM::ML_curv","ML_ADM::ML_lapse","ML_ADM::ML_metric","ML_ADM::ML_shift"}; + const char *const groups[] = { + "ML_ADM::ML_curv", + "ML_ADM::ML_lapse", + "ML_ADM::ML_metric", + "ML_ADM::ML_shift"}; GenericFD_AssertGroupStorage(cctkGH, "ML_ADM_Minkowski", 4, groups); switch(fdOrder) @@ -340,7 +272,7 @@ extern "C" void ML_ADM_Minkowski(CCTK_ARGUMENTS) break; } - GenericFD_LoopOverEverything(cctkGH, &ML_ADM_Minkowski_Body); + GenericFD_LoopOverEverything(cctkGH, ML_ADM_Minkowski_Body); if (verbose > 1) { diff --git a/ML_ADM/src/ML_ADM_RHS.cc b/ML_ADM/src/ML_ADM_RHS.cc index 4aa5793..b24ebe3 100644 --- a/ML_ADM/src/ML_ADM_RHS.cc +++ b/ML_ADM/src/ML_ADM_RHS.cc @@ -50,8 +50,6 @@ static void ML_ADM_RHS_Body(cGH const * restrict const cctkGH, int const dir, in DECLARE_CCTK_PARAMETERS; - /* Declare finite differencing variables */ - /* Include user-supplied include files */ /* Initialise finite differencing variables */ @@ -82,30 +80,30 @@ static void ML_ADM_RHS_Body(cGH const * restrict const cctkGH, int const dir, in 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 p1o180dx2 = kmul(INV(SQR(dx)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dy2 = kmul(INV(SQR(dy)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dz2 = kmul(INV(SQR(dz)),ToReal(0.00555555555555555555555555555556)); 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 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 p1o4dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.25))); + 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 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 p1o4dydz = kmul(INV(kmul(dy,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)); CCTK_REAL_VEC const p1o5040dz2 = kmul(INV(SQR(dz)),ToReal(0.000198412698412698412698412698413)); CCTK_REAL_VEC const p1o60dx = kmul(INV(dx),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dy = kmul(INV(dy),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dz = kmul(INV(dz),ToReal(0.0166666666666666666666666666667)); - 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)); @@ -171,7 +169,7 @@ static void ML_ADM_RHS_Body(cGH const * restrict const cctkGH, int const dir, in /* Loop over the grid points */ #pragma omp parallel - LC_LOOP3VEC (ML_ADM_RHS, + LC_LOOP3VEC(ML_ADM_RHS, 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) @@ -1473,96 +1471,26 @@ static void ML_ADM_RHS_Body(cGH const * restrict const cctkGH, int const dir, in CCTK_REAL_VEC beta3rhsL = ToReal(0); - /* 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(alpharhs[index],alpharhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(beta1rhs[index],beta1rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(beta2rhs[index],beta2rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(beta3rhs[index],beta3rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g11rhs[index],g11rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g12rhs[index],g12rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g13rhs[index],g13rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g22rhs[index],g22rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g23rhs[index],g23rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g33rhs[index],g33rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K11rhs[index],K11rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K12rhs[index],K12rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K13rhs[index],K13rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K22rhs[index],K22rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K23rhs[index],K23rhsL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K33rhs[index],K33rhsL,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(alpharhs[index],alpharhsL,elt_count); - vec_store_nta_partial_hi(beta1rhs[index],beta1rhsL,elt_count); - vec_store_nta_partial_hi(beta2rhs[index],beta2rhsL,elt_count); - vec_store_nta_partial_hi(beta3rhs[index],beta3rhsL,elt_count); - vec_store_nta_partial_hi(g11rhs[index],g11rhsL,elt_count); - vec_store_nta_partial_hi(g12rhs[index],g12rhsL,elt_count); - vec_store_nta_partial_hi(g13rhs[index],g13rhsL,elt_count); - vec_store_nta_partial_hi(g22rhs[index],g22rhsL,elt_count); - vec_store_nta_partial_hi(g23rhs[index],g23rhsL,elt_count); - vec_store_nta_partial_hi(g33rhs[index],g33rhsL,elt_count); - vec_store_nta_partial_hi(K11rhs[index],K11rhsL,elt_count); - vec_store_nta_partial_hi(K12rhs[index],K12rhsL,elt_count); - vec_store_nta_partial_hi(K13rhs[index],K13rhsL,elt_count); - vec_store_nta_partial_hi(K22rhs[index],K22rhsL,elt_count); - vec_store_nta_partial_hi(K23rhs[index],K23rhsL,elt_count); - vec_store_nta_partial_hi(K33rhs[index],K33rhsL,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(alpharhs[index],alpharhsL,elt_count); - vec_store_nta_partial_lo(beta1rhs[index],beta1rhsL,elt_count); - vec_store_nta_partial_lo(beta2rhs[index],beta2rhsL,elt_count); - vec_store_nta_partial_lo(beta3rhs[index],beta3rhsL,elt_count); - vec_store_nta_partial_lo(g11rhs[index],g11rhsL,elt_count); - vec_store_nta_partial_lo(g12rhs[index],g12rhsL,elt_count); - vec_store_nta_partial_lo(g13rhs[index],g13rhsL,elt_count); - vec_store_nta_partial_lo(g22rhs[index],g22rhsL,elt_count); - vec_store_nta_partial_lo(g23rhs[index],g23rhsL,elt_count); - vec_store_nta_partial_lo(g33rhs[index],g33rhsL,elt_count); - vec_store_nta_partial_lo(K11rhs[index],K11rhsL,elt_count); - vec_store_nta_partial_lo(K12rhs[index],K12rhsL,elt_count); - vec_store_nta_partial_lo(K13rhs[index],K13rhsL,elt_count); - vec_store_nta_partial_lo(K22rhs[index],K22rhsL,elt_count); - vec_store_nta_partial_lo(K23rhs[index],K23rhsL,elt_count); - vec_store_nta_partial_lo(K33rhs[index],K33rhsL,elt_count); - break; - } - vec_store_nta(alpharhs[index],alpharhsL); - vec_store_nta(beta1rhs[index],beta1rhsL); - vec_store_nta(beta2rhs[index],beta2rhsL); - vec_store_nta(beta3rhs[index],beta3rhsL); - vec_store_nta(g11rhs[index],g11rhsL); - vec_store_nta(g12rhs[index],g12rhsL); - vec_store_nta(g13rhs[index],g13rhsL); - vec_store_nta(g22rhs[index],g22rhsL); - vec_store_nta(g23rhs[index],g23rhsL); - vec_store_nta(g33rhs[index],g33rhsL); - vec_store_nta(K11rhs[index],K11rhsL); - vec_store_nta(K12rhs[index],K12rhsL); - vec_store_nta(K13rhs[index],K13rhsL); - vec_store_nta(K22rhs[index],K22rhsL); - vec_store_nta(K23rhs[index],K23rhsL); - vec_store_nta(K33rhs[index],K33rhsL); + /* Copy local copies back to grid functions */ + vec_store_partial_prepare(i,lc_imin,lc_imax); + vec_store_nta_partial(alpharhs[index],alpharhsL); + vec_store_nta_partial(beta1rhs[index],beta1rhsL); + vec_store_nta_partial(beta2rhs[index],beta2rhsL); + vec_store_nta_partial(beta3rhs[index],beta3rhsL); + vec_store_nta_partial(g11rhs[index],g11rhsL); + vec_store_nta_partial(g12rhs[index],g12rhsL); + vec_store_nta_partial(g13rhs[index],g13rhsL); + vec_store_nta_partial(g22rhs[index],g22rhsL); + vec_store_nta_partial(g23rhs[index],g23rhsL); + vec_store_nta_partial(g33rhs[index],g33rhsL); + vec_store_nta_partial(K11rhs[index],K11rhsL); + vec_store_nta_partial(K12rhs[index],K12rhsL); + vec_store_nta_partial(K13rhs[index],K13rhsL); + vec_store_nta_partial(K22rhs[index],K22rhsL); + vec_store_nta_partial(K23rhs[index],K23rhsL); + vec_store_nta_partial(K33rhs[index],K33rhsL); } - LC_ENDLOOP3VEC (ML_ADM_RHS); + LC_ENDLOOP3VEC(ML_ADM_RHS); } extern "C" void ML_ADM_RHS(CCTK_ARGUMENTS) @@ -1581,7 +1509,15 @@ extern "C" void ML_ADM_RHS(CCTK_ARGUMENTS) return; } - const char *groups[] = {"ML_ADM::ML_curv","ML_ADM::ML_curvrhs","ML_ADM::ML_lapse","ML_ADM::ML_lapserhs","ML_ADM::ML_metric","ML_ADM::ML_metricrhs","ML_ADM::ML_shift","ML_ADM::ML_shiftrhs"}; + const char *const groups[] = { + "ML_ADM::ML_curv", + "ML_ADM::ML_curvrhs", + "ML_ADM::ML_lapse", + "ML_ADM::ML_lapserhs", + "ML_ADM::ML_metric", + "ML_ADM::ML_metricrhs", + "ML_ADM::ML_shift", + "ML_ADM::ML_shiftrhs"}; GenericFD_AssertGroupStorage(cctkGH, "ML_ADM_RHS", 8, groups); switch(fdOrder) @@ -1603,7 +1539,7 @@ extern "C" void ML_ADM_RHS(CCTK_ARGUMENTS) break; } - GenericFD_LoopOverInterior(cctkGH, &ML_ADM_RHS_Body); + GenericFD_LoopOverInterior(cctkGH, ML_ADM_RHS_Body); if (verbose > 1) { diff --git a/ML_ADM/src/ML_ADM_boundary.cc b/ML_ADM/src/ML_ADM_boundary.cc index d4d81cf..6f0ef79 100644 --- a/ML_ADM/src/ML_ADM_boundary.cc +++ b/ML_ADM/src/ML_ADM_boundary.cc @@ -50,8 +50,6 @@ static void ML_ADM_boundary_Body(cGH const * restrict const cctkGH, int const di DECLARE_CCTK_PARAMETERS; - /* Declare finite differencing variables */ - /* Include user-supplied include files */ /* Initialise finite differencing variables */ @@ -82,30 +80,30 @@ static void ML_ADM_boundary_Body(cGH const * restrict const cctkGH, int const di 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 p1o180dx2 = kmul(INV(SQR(dx)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dy2 = kmul(INV(SQR(dy)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dz2 = kmul(INV(SQR(dz)),ToReal(0.00555555555555555555555555555556)); 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 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 p1o4dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.25))); + 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 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 p1o4dydz = kmul(INV(kmul(dy,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)); CCTK_REAL_VEC const p1o5040dz2 = kmul(INV(SQR(dz)),ToReal(0.000198412698412698412698412698413)); CCTK_REAL_VEC const p1o60dx = kmul(INV(dx),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dy = kmul(INV(dy),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dz = kmul(INV(dz),ToReal(0.0166666666666666666666666666667)); - 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)); @@ -171,7 +169,7 @@ static void ML_ADM_boundary_Body(cGH const * restrict const cctkGH, int const di /* Loop over the grid points */ #pragma omp parallel - LC_LOOP3VEC (ML_ADM_boundary, + LC_LOOP3VEC(ML_ADM_boundary, 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) @@ -235,96 +233,26 @@ static void ML_ADM_boundary_Body(cGH const * restrict const cctkGH, int const di CCTK_REAL_VEC beta3L = ToReal(0); - /* 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(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(g11[index],g11L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g12[index],g12L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g13[index],g13L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g22[index],g22L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g23[index],g23L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g33[index],g33L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K11[index],K11L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K12[index],K12L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K13[index],K13L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K22[index],K22L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K23[index],K23L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K33[index],K33L,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(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(g11[index],g11L,elt_count); - vec_store_nta_partial_hi(g12[index],g12L,elt_count); - vec_store_nta_partial_hi(g13[index],g13L,elt_count); - vec_store_nta_partial_hi(g22[index],g22L,elt_count); - vec_store_nta_partial_hi(g23[index],g23L,elt_count); - vec_store_nta_partial_hi(g33[index],g33L,elt_count); - vec_store_nta_partial_hi(K11[index],K11L,elt_count); - vec_store_nta_partial_hi(K12[index],K12L,elt_count); - vec_store_nta_partial_hi(K13[index],K13L,elt_count); - vec_store_nta_partial_hi(K22[index],K22L,elt_count); - vec_store_nta_partial_hi(K23[index],K23L,elt_count); - vec_store_nta_partial_hi(K33[index],K33L,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(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(g11[index],g11L,elt_count); - vec_store_nta_partial_lo(g12[index],g12L,elt_count); - vec_store_nta_partial_lo(g13[index],g13L,elt_count); - vec_store_nta_partial_lo(g22[index],g22L,elt_count); - vec_store_nta_partial_lo(g23[index],g23L,elt_count); - vec_store_nta_partial_lo(g33[index],g33L,elt_count); - vec_store_nta_partial_lo(K11[index],K11L,elt_count); - vec_store_nta_partial_lo(K12[index],K12L,elt_count); - vec_store_nta_partial_lo(K13[index],K13L,elt_count); - vec_store_nta_partial_lo(K22[index],K22L,elt_count); - vec_store_nta_partial_lo(K23[index],K23L,elt_count); - vec_store_nta_partial_lo(K33[index],K33L,elt_count); - break; - } - vec_store_nta(alpha[index],alphaL); - vec_store_nta(beta1[index],beta1L); - vec_store_nta(beta2[index],beta2L); - vec_store_nta(beta3[index],beta3L); - vec_store_nta(g11[index],g11L); - vec_store_nta(g12[index],g12L); - vec_store_nta(g13[index],g13L); - vec_store_nta(g22[index],g22L); - vec_store_nta(g23[index],g23L); - vec_store_nta(g33[index],g33L); - vec_store_nta(K11[index],K11L); - vec_store_nta(K12[index],K12L); - vec_store_nta(K13[index],K13L); - vec_store_nta(K22[index],K22L); - vec_store_nta(K23[index],K23L); - vec_store_nta(K33[index],K33L); + /* 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(beta1[index],beta1L); + vec_store_nta_partial(beta2[index],beta2L); + vec_store_nta_partial(beta3[index],beta3L); + vec_store_nta_partial(g11[index],g11L); + vec_store_nta_partial(g12[index],g12L); + vec_store_nta_partial(g13[index],g13L); + vec_store_nta_partial(g22[index],g22L); + vec_store_nta_partial(g23[index],g23L); + vec_store_nta_partial(g33[index],g33L); + vec_store_nta_partial(K11[index],K11L); + vec_store_nta_partial(K12[index],K12L); + vec_store_nta_partial(K13[index],K13L); + vec_store_nta_partial(K22[index],K22L); + vec_store_nta_partial(K23[index],K23L); + vec_store_nta_partial(K33[index],K33L); } - LC_ENDLOOP3VEC (ML_ADM_boundary); + LC_ENDLOOP3VEC(ML_ADM_boundary); } extern "C" void ML_ADM_boundary(CCTK_ARGUMENTS) @@ -343,7 +271,11 @@ extern "C" void ML_ADM_boundary(CCTK_ARGUMENTS) return; } - const char *groups[] = {"ML_ADM::ML_curv","ML_ADM::ML_lapse","ML_ADM::ML_metric","ML_ADM::ML_shift"}; + const char *const groups[] = { + "ML_ADM::ML_curv", + "ML_ADM::ML_lapse", + "ML_ADM::ML_metric", + "ML_ADM::ML_shift"}; GenericFD_AssertGroupStorage(cctkGH, "ML_ADM_boundary", 4, groups); switch(fdOrder) @@ -361,7 +293,7 @@ extern "C" void ML_ADM_boundary(CCTK_ARGUMENTS) break; } - GenericFD_LoopOverBoundaryWithGhosts(cctkGH, &ML_ADM_boundary_Body); + GenericFD_LoopOverBoundaryWithGhosts(cctkGH, ML_ADM_boundary_Body); if (verbose > 1) { diff --git a/ML_ADM/src/ML_ADM_constraints.cc b/ML_ADM/src/ML_ADM_constraints.cc index 3175e41..ca349a9 100644 --- a/ML_ADM/src/ML_ADM_constraints.cc +++ b/ML_ADM/src/ML_ADM_constraints.cc @@ -44,8 +44,6 @@ static void ML_ADM_constraints_Body(cGH const * restrict const cctkGH, int const DECLARE_CCTK_PARAMETERS; - /* Declare finite differencing variables */ - /* Include user-supplied include files */ /* Initialise finite differencing variables */ @@ -76,30 +74,30 @@ static void ML_ADM_constraints_Body(cGH const * restrict const cctkGH, int const 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 p1o180dx2 = kmul(INV(SQR(dx)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dy2 = kmul(INV(SQR(dy)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dz2 = kmul(INV(SQR(dz)),ToReal(0.00555555555555555555555555555556)); 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 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 p1o4dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.25))); + 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 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 p1o4dydz = kmul(INV(kmul(dy,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)); CCTK_REAL_VEC const p1o5040dz2 = kmul(INV(SQR(dz)),ToReal(0.000198412698412698412698412698413)); CCTK_REAL_VEC const p1o60dx = kmul(INV(dx),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dy = kmul(INV(dy),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dz = kmul(INV(dz),ToReal(0.0166666666666666666666666666667)); - 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)); @@ -165,7 +163,7 @@ static void ML_ADM_constraints_Body(cGH const * restrict const cctkGH, int const /* Loop over the grid points */ #pragma omp parallel - LC_LOOP3VEC (ML_ADM_constraints, + LC_LOOP3VEC(ML_ADM_constraints, 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) @@ -1218,48 +1216,14 @@ static void ML_ADM_constraints_Body(cGH const * restrict const cctkGH, int const CCTK_REAL_VEC M3L = kmadd(gu11,kmadd(K11L,G113,kmadd(K12L,G213,kadd(JacPDstandardNth1K13,knmsub(K23L,G211,knmsub(K33L,G311,kmsub(K13L,ksub(G313,G111),JacPDstandardNth3K11)))))),kmadd(gu12,kmadd(K11L,G123,kmadd(K12L,G223,kadd(JacPDstandardNth1K23,knmsub(K23L,G212,knmsub(K33L,G312,kmsub(K13L,ksub(G323,G112),JacPDstandardNth3K12)))))),kmadd(gu13,kmadd(K11L,G133,kmadd(K12L,G233,kadd(JacPDstandardNth1K33,knmsub(K23L,G213,knmsub(K33L,G313,kmsub(K13L,ksub(G333,G113),JacPDstandardNth3K13)))))),kmadd(gu21,kmadd(K12L,G113,kmadd(K22L,G213,kadd(JacPDstandardNth2K13,knmsub(K13L,G112,knmsub(K33L,G312,kmsub(K23L,ksub(G313,G212),JacPDstandardNth3K12)))))),kmadd(gu22,kmadd(K12L,G123,kmadd(K22L,G223,kadd(JacPDstandardNth2K23,knmsub(K13L,G122,knmsub(K33L,G322,kmsub(K23L,ksub(G323,G222),JacPDstandardNth3K22)))))),kmul(gu23,kmadd(K12L,G133,kmadd(K22L,G233,kadd(JacPDstandardNth2K33,knmsub(K13L,G123,knmsub(K33L,G323,kmsub(K23L,ksub(G333,G223),JacPDstandardNth3K23)))))))))))); - /* 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(H[index],HL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(M1[index],M1L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(M2[index],M2L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(M3[index],M3L,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(H[index],HL,elt_count); - vec_store_nta_partial_hi(M1[index],M1L,elt_count); - vec_store_nta_partial_hi(M2[index],M2L,elt_count); - vec_store_nta_partial_hi(M3[index],M3L,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(H[index],HL,elt_count); - vec_store_nta_partial_lo(M1[index],M1L,elt_count); - vec_store_nta_partial_lo(M2[index],M2L,elt_count); - vec_store_nta_partial_lo(M3[index],M3L,elt_count); - break; - } - vec_store_nta(H[index],HL); - vec_store_nta(M1[index],M1L); - vec_store_nta(M2[index],M2L); - vec_store_nta(M3[index],M3L); + /* Copy local copies back to grid functions */ + vec_store_partial_prepare(i,lc_imin,lc_imax); + vec_store_nta_partial(H[index],HL); + vec_store_nta_partial(M1[index],M1L); + vec_store_nta_partial(M2[index],M2L); + vec_store_nta_partial(M3[index],M3L); } - LC_ENDLOOP3VEC (ML_ADM_constraints); + LC_ENDLOOP3VEC(ML_ADM_constraints); } extern "C" void ML_ADM_constraints(CCTK_ARGUMENTS) @@ -1278,7 +1242,11 @@ extern "C" void ML_ADM_constraints(CCTK_ARGUMENTS) return; } - const char *groups[] = {"ML_ADM::ML_curv","ML_ADM::ML_Ham","ML_ADM::ML_metric","ML_ADM::ML_mom"}; + const char *const groups[] = { + "ML_ADM::ML_curv", + "ML_ADM::ML_Ham", + "ML_ADM::ML_metric", + "ML_ADM::ML_mom"}; GenericFD_AssertGroupStorage(cctkGH, "ML_ADM_constraints", 4, groups); switch(fdOrder) @@ -1300,7 +1268,7 @@ extern "C" void ML_ADM_constraints(CCTK_ARGUMENTS) break; } - GenericFD_LoopOverInterior(cctkGH, &ML_ADM_constraints_Body); + GenericFD_LoopOverInterior(cctkGH, ML_ADM_constraints_Body); if (verbose > 1) { diff --git a/ML_ADM/src/ML_ADM_constraints_boundary.cc b/ML_ADM/src/ML_ADM_constraints_boundary.cc index 7211b66..0749b4e 100644 --- a/ML_ADM/src/ML_ADM_constraints_boundary.cc +++ b/ML_ADM/src/ML_ADM_constraints_boundary.cc @@ -44,8 +44,6 @@ static void ML_ADM_constraints_boundary_Body(cGH const * restrict const cctkGH, DECLARE_CCTK_PARAMETERS; - /* Declare finite differencing variables */ - /* Include user-supplied include files */ /* Initialise finite differencing variables */ @@ -76,30 +74,30 @@ static void ML_ADM_constraints_boundary_Body(cGH const * restrict const cctkGH, 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 p1o180dx2 = kmul(INV(SQR(dx)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dy2 = kmul(INV(SQR(dy)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dz2 = kmul(INV(SQR(dz)),ToReal(0.00555555555555555555555555555556)); 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 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 p1o4dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.25))); + 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 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 p1o4dydz = kmul(INV(kmul(dy,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)); CCTK_REAL_VEC const p1o5040dz2 = kmul(INV(SQR(dz)),ToReal(0.000198412698412698412698412698413)); CCTK_REAL_VEC const p1o60dx = kmul(INV(dx),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dy = kmul(INV(dy),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dz = kmul(INV(dz),ToReal(0.0166666666666666666666666666667)); - 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)); @@ -165,7 +163,7 @@ static void ML_ADM_constraints_boundary_Body(cGH const * restrict const cctkGH, /* Loop over the grid points */ #pragma omp parallel - LC_LOOP3VEC (ML_ADM_constraints_boundary, + LC_LOOP3VEC(ML_ADM_constraints_boundary, 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) @@ -205,48 +203,14 @@ static void ML_ADM_constraints_boundary_Body(cGH const * restrict const cctkGH, CCTK_REAL_VEC M3L = ToReal(0); - /* 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(H[index],HL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(M1[index],M1L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(M2[index],M2L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(M3[index],M3L,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(H[index],HL,elt_count); - vec_store_nta_partial_hi(M1[index],M1L,elt_count); - vec_store_nta_partial_hi(M2[index],M2L,elt_count); - vec_store_nta_partial_hi(M3[index],M3L,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(H[index],HL,elt_count); - vec_store_nta_partial_lo(M1[index],M1L,elt_count); - vec_store_nta_partial_lo(M2[index],M2L,elt_count); - vec_store_nta_partial_lo(M3[index],M3L,elt_count); - break; - } - vec_store_nta(H[index],HL); - vec_store_nta(M1[index],M1L); - vec_store_nta(M2[index],M2L); - vec_store_nta(M3[index],M3L); + /* Copy local copies back to grid functions */ + vec_store_partial_prepare(i,lc_imin,lc_imax); + vec_store_nta_partial(H[index],HL); + vec_store_nta_partial(M1[index],M1L); + vec_store_nta_partial(M2[index],M2L); + vec_store_nta_partial(M3[index],M3L); } - LC_ENDLOOP3VEC (ML_ADM_constraints_boundary); + LC_ENDLOOP3VEC(ML_ADM_constraints_boundary); } extern "C" void ML_ADM_constraints_boundary(CCTK_ARGUMENTS) @@ -265,7 +229,9 @@ extern "C" void ML_ADM_constraints_boundary(CCTK_ARGUMENTS) return; } - const char *groups[] = {"ML_ADM::ML_Ham","ML_ADM::ML_mom"}; + const char *const groups[] = { + "ML_ADM::ML_Ham", + "ML_ADM::ML_mom"}; GenericFD_AssertGroupStorage(cctkGH, "ML_ADM_constraints_boundary", 2, groups); switch(fdOrder) @@ -283,7 +249,7 @@ extern "C" void ML_ADM_constraints_boundary(CCTK_ARGUMENTS) break; } - GenericFD_LoopOverBoundaryWithGhosts(cctkGH, &ML_ADM_constraints_boundary_Body); + GenericFD_LoopOverBoundaryWithGhosts(cctkGH, ML_ADM_constraints_boundary_Body); if (verbose > 1) { diff --git a/ML_ADM/src/ML_ADM_convertFromADMBase.cc b/ML_ADM/src/ML_ADM_convertFromADMBase.cc index ccf06b3..7ba54c9 100644 --- a/ML_ADM/src/ML_ADM_convertFromADMBase.cc +++ b/ML_ADM/src/ML_ADM_convertFromADMBase.cc @@ -29,8 +29,6 @@ static void ML_ADM_convertFromADMBase_Body(cGH const * restrict const cctkGH, in DECLARE_CCTK_PARAMETERS; - /* Declare finite differencing variables */ - /* Include user-supplied include files */ /* Initialise finite differencing variables */ @@ -61,30 +59,30 @@ static void ML_ADM_convertFromADMBase_Body(cGH const * restrict const cctkGH, in 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 p1o180dx2 = kmul(INV(SQR(dx)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dy2 = kmul(INV(SQR(dy)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dz2 = kmul(INV(SQR(dz)),ToReal(0.00555555555555555555555555555556)); 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 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 p1o4dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.25))); + 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 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 p1o4dydz = kmul(INV(kmul(dy,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)); CCTK_REAL_VEC const p1o5040dz2 = kmul(INV(SQR(dz)),ToReal(0.000198412698412698412698412698413)); CCTK_REAL_VEC const p1o60dx = kmul(INV(dx),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dy = kmul(INV(dy),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dz = kmul(INV(dz),ToReal(0.0166666666666666666666666666667)); - 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)); @@ -150,7 +148,7 @@ static void ML_ADM_convertFromADMBase_Body(cGH const * restrict const cctkGH, in /* Loop over the grid points */ #pragma omp parallel - LC_LOOP3VEC (ML_ADM_convertFromADMBase, + LC_LOOP3VEC(ML_ADM_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) @@ -230,96 +228,26 @@ static void ML_ADM_convertFromADMBase_Body(cGH const * restrict const cctkGH, in 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(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(g11[index],g11L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g12[index],g12L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g13[index],g13L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g22[index],g22L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g23[index],g23L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(g33[index],g33L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K11[index],K11L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K12[index],K12L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K13[index],K13L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K22[index],K22L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K23[index],K23L,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(K33[index],K33L,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(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(g11[index],g11L,elt_count); - vec_store_nta_partial_hi(g12[index],g12L,elt_count); - vec_store_nta_partial_hi(g13[index],g13L,elt_count); - vec_store_nta_partial_hi(g22[index],g22L,elt_count); - vec_store_nta_partial_hi(g23[index],g23L,elt_count); - vec_store_nta_partial_hi(g33[index],g33L,elt_count); - vec_store_nta_partial_hi(K11[index],K11L,elt_count); - vec_store_nta_partial_hi(K12[index],K12L,elt_count); - vec_store_nta_partial_hi(K13[index],K13L,elt_count); - vec_store_nta_partial_hi(K22[index],K22L,elt_count); - vec_store_nta_partial_hi(K23[index],K23L,elt_count); - vec_store_nta_partial_hi(K33[index],K33L,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(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(g11[index],g11L,elt_count); - vec_store_nta_partial_lo(g12[index],g12L,elt_count); - vec_store_nta_partial_lo(g13[index],g13L,elt_count); - vec_store_nta_partial_lo(g22[index],g22L,elt_count); - vec_store_nta_partial_lo(g23[index],g23L,elt_count); - vec_store_nta_partial_lo(g33[index],g33L,elt_count); - vec_store_nta_partial_lo(K11[index],K11L,elt_count); - vec_store_nta_partial_lo(K12[index],K12L,elt_count); - vec_store_nta_partial_lo(K13[index],K13L,elt_count); - vec_store_nta_partial_lo(K22[index],K22L,elt_count); - vec_store_nta_partial_lo(K23[index],K23L,elt_count); - vec_store_nta_partial_lo(K33[index],K33L,elt_count); - break; - } - vec_store_nta(alpha[index],alphaL); - vec_store_nta(beta1[index],beta1L); - vec_store_nta(beta2[index],beta2L); - vec_store_nta(beta3[index],beta3L); - vec_store_nta(g11[index],g11L); - vec_store_nta(g12[index],g12L); - vec_store_nta(g13[index],g13L); - vec_store_nta(g22[index],g22L); - vec_store_nta(g23[index],g23L); - vec_store_nta(g33[index],g33L); - vec_store_nta(K11[index],K11L); - vec_store_nta(K12[index],K12L); - vec_store_nta(K13[index],K13L); - vec_store_nta(K22[index],K22L); - vec_store_nta(K23[index],K23L); - vec_store_nta(K33[index],K33L); + /* 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(beta1[index],beta1L); + vec_store_nta_partial(beta2[index],beta2L); + vec_store_nta_partial(beta3[index],beta3L); + vec_store_nta_partial(g11[index],g11L); + vec_store_nta_partial(g12[index],g12L); + vec_store_nta_partial(g13[index],g13L); + vec_store_nta_partial(g22[index],g22L); + vec_store_nta_partial(g23[index],g23L); + vec_store_nta_partial(g33[index],g33L); + vec_store_nta_partial(K11[index],K11L); + vec_store_nta_partial(K12[index],K12L); + vec_store_nta_partial(K13[index],K13L); + vec_store_nta_partial(K22[index],K22L); + vec_store_nta_partial(K23[index],K23L); + vec_store_nta_partial(K33[index],K33L); } - LC_ENDLOOP3VEC (ML_ADM_convertFromADMBase); + LC_ENDLOOP3VEC(ML_ADM_convertFromADMBase); } extern "C" void ML_ADM_convertFromADMBase(CCTK_ARGUMENTS) @@ -338,7 +266,15 @@ extern "C" void ML_ADM_convertFromADMBase(CCTK_ARGUMENTS) return; } - const char *groups[] = {"ADMBase::curv","ADMBase::lapse","ADMBase::metric","ADMBase::shift","ML_ADM::ML_curv","ML_ADM::ML_lapse","ML_ADM::ML_metric","ML_ADM::ML_shift"}; + const char *const groups[] = { + "ADMBase::curv", + "ADMBase::lapse", + "ADMBase::metric", + "ADMBase::shift", + "ML_ADM::ML_curv", + "ML_ADM::ML_lapse", + "ML_ADM::ML_metric", + "ML_ADM::ML_shift"}; GenericFD_AssertGroupStorage(cctkGH, "ML_ADM_convertFromADMBase", 8, groups); switch(fdOrder) @@ -356,7 +292,7 @@ extern "C" void ML_ADM_convertFromADMBase(CCTK_ARGUMENTS) break; } - GenericFD_LoopOverEverything(cctkGH, &ML_ADM_convertFromADMBase_Body); + GenericFD_LoopOverEverything(cctkGH, ML_ADM_convertFromADMBase_Body); if (verbose > 1) { diff --git a/ML_ADM/src/ML_ADM_convertToADMBase.cc b/ML_ADM/src/ML_ADM_convertToADMBase.cc index 83608cf..1a75640 100644 --- a/ML_ADM/src/ML_ADM_convertToADMBase.cc +++ b/ML_ADM/src/ML_ADM_convertToADMBase.cc @@ -29,8 +29,6 @@ static void ML_ADM_convertToADMBase_Body(cGH const * restrict const cctkGH, int DECLARE_CCTK_PARAMETERS; - /* Declare finite differencing variables */ - /* Include user-supplied include files */ /* Initialise finite differencing variables */ @@ -61,30 +59,30 @@ static void ML_ADM_convertToADMBase_Body(cGH const * restrict const cctkGH, int 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 p1o180dx2 = kmul(INV(SQR(dx)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dy2 = kmul(INV(SQR(dy)),ToReal(0.00555555555555555555555555555556)); CCTK_REAL_VEC const p1o180dz2 = kmul(INV(SQR(dz)),ToReal(0.00555555555555555555555555555556)); 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 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 p1o4dydz = kmul(INV(dy),kmul(INV(dz),ToReal(0.25))); + 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 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 p1o4dydz = kmul(INV(kmul(dy,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)); CCTK_REAL_VEC const p1o5040dz2 = kmul(INV(SQR(dz)),ToReal(0.000198412698412698412698412698413)); CCTK_REAL_VEC const p1o60dx = kmul(INV(dx),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dy = kmul(INV(dy),ToReal(0.0166666666666666666666666666667)); CCTK_REAL_VEC const p1o60dz = kmul(INV(dz),ToReal(0.0166666666666666666666666666667)); - 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)); @@ -150,7 +148,7 @@ static void ML_ADM_convertToADMBase_Body(cGH const * restrict const cctkGH, int /* Loop over the grid points */ #pragma omp parallel - LC_LOOP3VEC (ML_ADM_convertToADMBase, + LC_LOOP3VEC(ML_ADM_convertToADMBase, 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) @@ -238,112 +236,30 @@ static void ML_ADM_convertToADMBase_Body(cGH const * restrict const cctkGH, int CCTK_REAL_VEC dtbetazL = ToReal(0); - /* 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(alp[index],alpL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(betax[index],betaxL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(betay[index],betayL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(betaz[index],betazL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(dtalp[index],dtalpL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(dtbetax[index],dtbetaxL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(dtbetay[index],dtbetayL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(dtbetaz[index],dtbetazL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gxx[index],gxxL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gxy[index],gxyL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gxz[index],gxzL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gyy[index],gyyL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gyz[index],gyzL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(gzz[index],gzzL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(kxx[index],kxxL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(kxy[index],kxyL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(kxz[index],kxzL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(kyy[index],kyyL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(kyz[index],kyzL,elt_count_lo,elt_count_hi); - vec_store_nta_partial_mid(kzz[index],kzzL,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(alp[index],alpL,elt_count); - vec_store_nta_partial_hi(betax[index],betaxL,elt_count); - vec_store_nta_partial_hi(betay[index],betayL,elt_count); - vec_store_nta_partial_hi(betaz[index],betazL,elt_count); - vec_store_nta_partial_hi(dtalp[index],dtalpL,elt_count); - vec_store_nta_partial_hi(dtbetax[index],dtbetaxL,elt_count); - vec_store_nta_partial_hi(dtbetay[index],dtbetayL,elt_count); - vec_store_nta_partial_hi(dtbetaz[index],dtbetazL,elt_count); - vec_store_nta_partial_hi(gxx[index],gxxL,elt_count); - vec_store_nta_partial_hi(gxy[index],gxyL,elt_count); - vec_store_nta_partial_hi(gxz[index],gxzL,elt_count); - vec_store_nta_partial_hi(gyy[index],gyyL,elt_count); - vec_store_nta_partial_hi(gyz[index],gyzL,elt_count); - vec_store_nta_partial_hi(gzz[index],gzzL,elt_count); - vec_store_nta_partial_hi(kxx[index],kxxL,elt_count); - vec_store_nta_partial_hi(kxy[index],kxyL,elt_count); - vec_store_nta_partial_hi(kxz[index],kxzL,elt_count); - vec_store_nta_partial_hi(kyy[index],kyyL,elt_count); - vec_store_nta_partial_hi(kyz[index],kyzL,elt_count); - vec_store_nta_partial_hi(kzz[index],kzzL,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(alp[index],alpL,elt_count); - vec_store_nta_partial_lo(betax[index],betaxL,elt_count); - vec_store_nta_partial_lo(betay[index],betayL,elt_count); - vec_store_nta_partial_lo(betaz[index],betazL,elt_count); - vec_store_nta_partial_lo(dtalp[index],dtalpL,elt_count); - vec_store_nta_partial_lo(dtbetax[index],dtbetaxL,elt_count); - vec_store_nta_partial_lo(dtbetay[index],dtbetayL,elt_count); - vec_store_nta_partial_lo(dtbetaz[index],dtbetazL,elt_count); - vec_store_nta_partial_lo(gxx[index],gxxL,elt_count); - vec_store_nta_partial_lo(gxy[index],gxyL,elt_count); - vec_store_nta_partial_lo(gxz[index],gxzL,elt_count); - vec_store_nta_partial_lo(gyy[index],gyyL,elt_count); - vec_store_nta_partial_lo(gyz[index],gyzL,elt_count); - vec_store_nta_partial_lo(gzz[index],gzzL,elt_count); - vec_store_nta_partial_lo(kxx[index],kxxL,elt_count); - vec_store_nta_partial_lo(kxy[index],kxyL,elt_count); - vec_store_nta_partial_lo(kxz[index],kxzL,elt_count); - vec_store_nta_partial_lo(kyy[index],kyyL,elt_count); - vec_store_nta_partial_lo(kyz[index],kyzL,elt_count); - vec_store_nta_partial_lo(kzz[index],kzzL,elt_count); - break; - } - vec_store_nta(alp[index],alpL); - vec_store_nta(betax[index],betaxL); - vec_store_nta(betay[index],betayL); - vec_store_nta(betaz[index],betazL); - vec_store_nta(dtalp[index],dtalpL); - vec_store_nta(dtbetax[index],dtbetaxL); - vec_store_nta(dtbetay[index],dtbetayL); - vec_store_nta(dtbetaz[index],dtbetazL); - vec_store_nta(gxx[index],gxxL); - vec_store_nta(gxy[index],gxyL); - vec_store_nta(gxz[index],gxzL); - vec_store_nta(gyy[index],gyyL); - vec_store_nta(gyz[index],gyzL); - vec_store_nta(gzz[index],gzzL); - vec_store_nta(kxx[index],kxxL); - vec_store_nta(kxy[index],kxyL); - vec_store_nta(kxz[index],kxzL); - vec_store_nta(kyy[index],kyyL); - vec_store_nta(kyz[index],kyzL); - vec_store_nta(kzz[index],kzzL); + /* Copy local copies back to grid functions */ + vec_store_partial_prepare(i,lc_imin,lc_imax); + vec_store_nta_partial(alp[index],alpL); + vec_store_nta_partial(betax[index],betaxL); + vec_store_nta_partial(betay[index],betayL); + vec_store_nta_partial(betaz[index],betazL); + vec_store_nta_partial(dtalp[index],dtalpL); + vec_store_nta_partial(dtbetax[index],dtbetaxL); + vec_store_nta_partial(dtbetay[index],dtbetayL); + vec_store_nta_partial(dtbetaz[index],dtbetazL); + vec_store_nta_partial(gxx[index],gxxL); + vec_store_nta_partial(gxy[index],gxyL); + vec_store_nta_partial(gxz[index],gxzL); + vec_store_nta_partial(gyy[index],gyyL); + vec_store_nta_partial(gyz[index],gyzL); + vec_store_nta_partial(gzz[index],gzzL); + vec_store_nta_partial(kxx[index],kxxL); + vec_store_nta_partial(kxy[index],kxyL); + vec_store_nta_partial(kxz[index],kxzL); + vec_store_nta_partial(kyy[index],kyyL); + vec_store_nta_partial(kyz[index],kyzL); + vec_store_nta_partial(kzz[index],kzzL); } - LC_ENDLOOP3VEC (ML_ADM_convertToADMBase); + LC_ENDLOOP3VEC(ML_ADM_convertToADMBase); } extern "C" void ML_ADM_convertToADMBase(CCTK_ARGUMENTS) @@ -362,7 +278,17 @@ extern "C" void ML_ADM_convertToADMBase(CCTK_ARGUMENTS) return; } - const char *groups[] = {"ADMBase::curv","ADMBase::dtlapse","ADMBase::dtshift","ADMBase::lapse","ADMBase::metric","ADMBase::shift","ML_ADM::ML_curv","ML_ADM::ML_lapse","ML_ADM::ML_metric","ML_ADM::ML_shift"}; + const char *const groups[] = { + "ADMBase::curv", + "ADMBase::dtlapse", + "ADMBase::dtshift", + "ADMBase::lapse", + "ADMBase::metric", + "ADMBase::shift", + "ML_ADM::ML_curv", + "ML_ADM::ML_lapse", + "ML_ADM::ML_metric", + "ML_ADM::ML_shift"}; GenericFD_AssertGroupStorage(cctkGH, "ML_ADM_convertToADMBase", 10, groups); switch(fdOrder) @@ -380,7 +306,7 @@ extern "C" void ML_ADM_convertToADMBase(CCTK_ARGUMENTS) break; } - GenericFD_LoopOverEverything(cctkGH, &ML_ADM_convertToADMBase_Body); + GenericFD_LoopOverEverything(cctkGH, ML_ADM_convertToADMBase_Body); if (verbose > 1) { diff --git a/ML_ADM/src/make.code.defn b/ML_ADM/src/make.code.defn index 06aad41..3a79c47 100644 --- a/ML_ADM/src/make.code.defn +++ b/ML_ADM/src/make.code.defn @@ -1,3 +1,3 @@ # File produced by Kranc -SRCS = Startup.cc RegisterMoL.cc RegisterSymmetries.cc ML_ADM_Minkowski.cc ML_ADM_convertFromADMBase.cc ML_ADM_RHS.cc ML_ADM_boundary.cc ML_ADM_convertToADMBase.cc ML_ADM_constraints.cc ML_ADM_constraints_boundary.cc Boundaries.cc +SRCS = Startup.cc RegisterSymmetries.cc RegisterMoL.cc ML_ADM_Minkowski.cc ML_ADM_convertFromADMBase.cc ML_ADM_RHS.cc ML_ADM_boundary.cc ML_ADM_convertToADMBase.cc ML_ADM_constraints.cc ML_ADM_constraints_boundary.cc Boundaries.cc -- cgit v1.2.3