Fix clang-tidy complaints
[alexxy/gromacs.git] / src / gromacs / listed_forces / pairs.cpp
index e910e592275cb495faadd988ba2de3c7b8e45a48..064c837756c639b5c7da10f8d9c8566a9717cc69 100644 (file)
@@ -1,7 +1,8 @@
 /*
  * This file is part of the GROMACS molecular simulation package.
  *
- * Copyright (c) 2014,2015,2016,2017,2018,2019, by the GROMACS development team, led by
+ * Copyright (c) 2014,2015,2016,2017,2018 by the GROMACS development team.
+ * Copyright (c) 2019,2020,2021, by the GROMACS development team, led by
  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
  * and including many others, as listed in the AUTHORS file in the
  * top-level source directory and at http://www.gromacs.org.
 
 #include <cmath>
 
+#include "gromacs/listed_forces/bonded.h"
 #include "gromacs/math/functions.h"
 #include "gromacs/math/vec.h"
+#include "gromacs/mdtypes/forcerec.h"
 #include "gromacs/mdtypes/group.h"
+#include "gromacs/mdtypes/interaction_const.h"
 #include "gromacs/mdtypes/md_enums.h"
 #include "gromacs/mdtypes/nblist.h"
+#include "gromacs/mdtypes/simulation_workload.h"
 #include "gromacs/pbcutil/ishift.h"
-#include "gromacs/pbcutil/mshift.h"
 #include "gromacs/pbcutil/pbc.h"
-#include "gromacs/pbcutil/pbc-simd.h"
+#include "gromacs/pbcutil/pbc_simd.h"
 #include "gromacs/simd/simd.h"
 #include "gromacs/simd/simd_math.h"
 #include "gromacs/simd/vector_operations.h"
 #include "gromacs/utility/fatalerror.h"
 #include "gromacs/utility/gmxassert.h"
 
-#include "listed-internal.h"
+#include "listed_internal.h"
 
 using namespace gmx; // TODO: Remove when this file is moved into gmx namespace
 
 /*! \brief Issue a warning if a listed interaction is beyond a table limit */
-static void
-warning_rlimit(const rvec *x, int ai, int aj, int * global_atom_index, real r, real rlimit)
+static void warning_rlimit(const rvec* x, int ai, int aj, int* global_atom_index, real r, real rlimit)
 {
-    gmx_warning("Listed nonbonded interaction between particles %d and %d\n"
-                "at distance %.3f which is larger than the table limit %.3f nm.\n\n"
-                "This is likely either a 1,4 interaction, or a listed interaction inside\n"
-                "a smaller molecule you are decoupling during a free energy calculation.\n"
-                "Since interactions at distances beyond the table cannot be computed,\n"
-                "they are skipped until they are inside the table limit again. You will\n"
-                "only see this message once, even if it occurs for several interactions.\n\n"
-                "IMPORTANT: This should not happen in a stable simulation, so there is\n"
-                "probably something wrong with your system. Only change the table-extension\n"
-                "distance in the mdp file if you are really sure that is the reason.\n",
-                glatnr(global_atom_index, ai), glatnr(global_atom_index, aj), r, rlimit);
+    gmx_warning(
+            "Listed nonbonded interaction between particles %d and %d\n"
+            "at distance %.3f which is larger than the table limit %.3f nm.\n\n"
+            "This is likely either a 1,4 interaction, or a listed interaction inside\n"
+            "a smaller molecule you are decoupling during a free energy calculation.\n"
+            "Since interactions at distances beyond the table cannot be computed,\n"
+            "they are skipped until they are inside the table limit again. You will\n"
+            "only see this message once, even if it occurs for several interactions.\n\n"
+            "IMPORTANT: This should not happen in a stable simulation, so there is\n"
+            "probably something wrong with your system. Only change the table-extension\n"
+            "distance in the mdp file if you are really sure that is the reason.\n",
+            glatnr(global_atom_index, ai),
+            glatnr(global_atom_index, aj),
+            r,
+            rlimit);
 
     if (debug)
     {
         fprintf(debug,
-                "%8f %8f %8f\n%8f %8f %8f\n1-4 (%d,%d) interaction not within cut-off! r=%g. Ignored\n",
-                x[ai][XX], x[ai][YY], x[ai][ZZ], x[aj][XX], x[aj][YY], x[aj][ZZ],
-                glatnr(global_atom_index, ai), glatnr(global_atom_index, aj), r);
+                "%8f %8f %8f\n%8f %8f %8f\n1-4 (%d,%d) interaction not within cut-off! r=%g. "
+                "Ignored\n",
+                x[ai][XX],
+                x[ai][YY],
+                x[ai][ZZ],
+                x[aj][XX],
+                x[aj][YY],
+                x[aj][ZZ],
+                glatnr(global_atom_index, ai),
+                glatnr(global_atom_index, aj),
+                r);
     }
 }
 
 /*! \brief Compute the energy and force for a single pair interaction */
-static real
-evaluate_single(real r2, real tabscale, const real *vftab, real tableStride,
-                real qq, real c6, real c12, real *velec, real *vvdw)
+static real evaluate_single(real        r2,
+                            real        tabscale,
+                            const real* vftab,
+                            real        tableStride,
+                            real        qq,
+                            real        c6,
+                            real        c12,
+                            real*       velec,
+                            real*       vvdw)
 {
-    real       rinv, r, rtab, eps, eps2, Y, F, Geps, Heps2, Fp, VVe, FFe, VVd, FFd, VVr, FFr, fscal;
-    int        ntab;
+    real rinv, r, rtab, eps, eps2, Y, F, Geps, Heps2, Fp, VVe, FFe, VVd, FFd, VVr, FFr, fscal;
+    int  ntab;
 
     /* Do the tabulated interactions - first table lookup */
-    rinv             = gmx::invsqrt(r2);
-    r                = r2*rinv;
-    rtab             = r*tabscale;
-    ntab             = static_cast<int>(rtab);
-    eps              = rtab-ntab;
-    eps2             = eps*eps;
-    ntab             = static_cast<int>(tableStride*ntab);
+    rinv = gmx::invsqrt(r2);
+    r    = r2 * rinv;
+    rtab = r * tabscale;
+    ntab = static_cast<int>(rtab);
+    eps  = rtab - ntab;
+    eps2 = eps * eps;
+    ntab = static_cast<int>(tableStride * ntab);
     /* Electrostatics */
-    Y                = vftab[ntab];
-    F                = vftab[ntab+1];
-    Geps             = eps*vftab[ntab+2];
-    Heps2            = eps2*vftab[ntab+3];
-    Fp               = F+Geps+Heps2;
-    VVe              = Y+eps*Fp;
-    FFe              = Fp+Geps+2.0*Heps2;
+    Y     = vftab[ntab];
+    F     = vftab[ntab + 1];
+    Geps  = eps * vftab[ntab + 2];
+    Heps2 = eps2 * vftab[ntab + 3];
+    Fp    = F + Geps + Heps2;
+    VVe   = Y + eps * Fp;
+    FFe   = Fp + Geps + 2.0 * Heps2;
     /* Dispersion */
-    Y                = vftab[ntab+4];
-    F                = vftab[ntab+5];
-    Geps             = eps*vftab[ntab+6];
-    Heps2            = eps2*vftab[ntab+7];
-    Fp               = F+Geps+Heps2;
-    VVd              = Y+eps*Fp;
-    FFd              = Fp+Geps+2.0*Heps2;
+    Y     = vftab[ntab + 4];
+    F     = vftab[ntab + 5];
+    Geps  = eps * vftab[ntab + 6];
+    Heps2 = eps2 * vftab[ntab + 7];
+    Fp    = F + Geps + Heps2;
+    VVd   = Y + eps * Fp;
+    FFd   = Fp + Geps + 2.0 * Heps2;
     /* Repulsion */
-    Y                = vftab[ntab+8];
-    F                = vftab[ntab+9];
-    Geps             = eps*vftab[ntab+10];
-    Heps2            = eps2*vftab[ntab+11];
-    Fp               = F+Geps+Heps2;
-    VVr              = Y+eps*Fp;
-    FFr              = Fp+Geps+2.0*Heps2;
+    Y     = vftab[ntab + 8];
+    F     = vftab[ntab + 9];
+    Geps  = eps * vftab[ntab + 10];
+    Heps2 = eps2 * vftab[ntab + 11];
+    Fp    = F + Geps + Heps2;
+    VVr   = Y + eps * Fp;
+    FFr   = Fp + Geps + 2.0 * Heps2;
 
-    *velec           = qq*VVe;
-    *vvdw            = c6*VVd+c12*VVr;
+    *velec = qq * VVe;
+    *vvdw  = c6 * VVd + c12 * VVr;
 
-    fscal            = -(qq*FFe+c6*FFd+c12*FFr)*tabscale*rinv;
+    fscal = -(qq * FFe + c6 * FFd + c12 * FFr) * tabscale * rinv;
 
     return fscal;
 }
 
+static inline real sixthRoot(const real r)
+{
+    return gmx::invsqrt(std::cbrt(r));
+}
+
 /*! \brief Compute the energy and force for a single pair interaction under FEP */
-static real
-free_energy_evaluate_single(real r2, real sc_r_power, real alpha_coul,
-                            real alpha_vdw, real tabscale, const real *vftab, real tableStride,
-                            real qqA, real c6A, real c12A, real qqB,
-                            real c6B, real c12B, const real LFC[2], const real LFV[2], const real DLF[2],
-                            const real lfac_coul[2], const real lfac_vdw[2], const real dlfac_coul[2],
-                            const real dlfac_vdw[2], real sigma6_def, real sigma6_min,
-                            real sigma2_def, real sigma2_min,
-                            real *velectot, real *vvdwtot, real *dvdl)
+template<KernelSoftcoreType softcoreType>
+static real free_energy_evaluate_single(real                                           r2,
+                                        real                                           rCoulCutoff,
+                                        const interaction_const_t::SoftCoreParameters& scParams,
+                                        real                                           tabscale,
+                                        const real*                                    vftab,
+                                        real                                           tableStride,
+                                        real                                           qqA,
+                                        real                                           c6A,
+                                        real                                           c12A,
+                                        real                                           qqB,
+                                        real                                           c6B,
+                                        real                                           c12B,
+                                        real                                           facel,
+                                        const real                                     LFC[2],
+                                        const real                                     LFV[2],
+                                        const real                                     DLF[2],
+                                        const real                                     lfac_coul[2],
+                                        const real                                     lfac_vdw[2],
+                                        const real dlfac_coul[2],
+                                        const real dlfac_vdw[2],
+                                        real*      velectot,
+                                        real*      vvdwtot,
+                                        real*      dvdl)
 {
-    real       rp, rpm2, rtab, eps, eps2, Y, F, Geps, Heps2, Fp, VV, FF, fscal;
-    real       qq[2], c6[2], c12[2], sigma6[2], sigma2[2], sigma_pow[2];
-    real       alpha_coul_eff, alpha_vdw_eff, dvdl_coul, dvdl_vdw;
+    real       rtab, eps, eps2, Y, F, Geps, Heps2, Fp, VV, FF, fscal;
+    real       qq[2], c6[2], c12[2], sigma6[2], sigma_pow[2];
+    real       alpha_coul_eff, alpha_vdw_eff, dvdl_coul_sum, dvdl_vdw_sum;
     real       rpinv, r_coul, r_vdw, velecsum, vvdwsum;
     real       fscal_vdw[2], fscal_elec[2];
     real       velec[2], vvdw[2];
+    real       dvdl_elec[2], dvdl_vdw[2];
+    real       gapsysScaleLinpointCoul, gapsysScaleLinpointVdW, gapsysSigma6VdW[2];
+    real       rQ, rLJ;
+    real       scaleDvdlRCoul;
     int        i, ntab;
-    const real half        = 0.5;
-    const real minusOne    = -1.0;
-    const real one         = 1.0;
-    const real two         = 2.0;
-    const real six         = 6.0;
-    const real fourtyeight = 48.0;
-
-    qq[0]    = qqA;
-    qq[1]    = qqB;
-    c6[0]    = c6A;
-    c6[1]    = c6B;
-    c12[0]   = c12A;
-    c12[1]   = c12B;
-
-    if (sc_r_power == six)
-    {
-        rpm2             = r2*r2;   /* r4 */
-        rp               = rpm2*r2; /* r6 */
-    }
-    else if (sc_r_power == fourtyeight)
-    {
-        rp               = r2*r2*r2; /* r6 */
-        rp               = rp*rp;    /* r12 */
-        rp               = rp*rp;    /* r24 */
-        rp               = rp*rp;    /* r48 */
-        rpm2             = rp/r2;    /* r46 */
-    }
-    else
-    {
-        rp             = std::pow(r2, half * sc_r_power);  /* not currently supported as input, but can handle it */
-        rpm2           = rp/r2;
-    }
+    const real half = 0.5_real;
+    const real one  = 1.0_real;
+    const real two  = 2.0_real;
+
+    qq[0]  = qqA;
+    qq[1]  = qqB;
+    c6[0]  = c6A;
+    c6[1]  = c6B;
+    c12[0] = c12A;
+    c12[1] = c12B;
+
+    const real rpm2 = r2 * r2;   /* r4 */
+    const real rp   = rpm2 * r2; /* r6 */
+    const real r    = sqrt(r2);  /* r1 */
 
     /* Loop over state A(0) and B(1) */
     for (i = 0; i < 2; i++)
     {
-        if ((c6[i] > 0) && (c12[i] > 0))
+        if constexpr (softcoreType == KernelSoftcoreType::Beutler)
         {
-            /* The c6 & c12 coefficients now contain the constants 6.0 and 12.0, respectively.
-             * Correct for this by multiplying with (1/12.0)/(1/6.0)=6.0/12.0=0.5.
-             */
-            sigma6[i]       = half*c12[i]/c6[i];
-            sigma2[i]       = std::cbrt(half*c12[i]/c6[i]);
-            /* should be able to get rid of this ^^^ internal pow call eventually.  Will require agreement on
-               what data to store externally.  Can't be fixed without larger scale changes, so not 5.0 */
-            if (sigma6[i] < sigma6_min)   /* for disappearing coul and vdw with soft core at the same time */
+            if ((c6[i] > 0) && (c12[i] > 0))
             {
-                sigma6[i] = sigma6_min;
-                sigma2[i] = sigma2_min;
+                /* The c6 & c12 coefficients now contain the constants 6.0 and 12.0, respectively.
+                 * Correct for this by multiplying with (1/12.0)/(1/6.0)=6.0/12.0=0.5.
+                 */
+                sigma6[i] = half * c12[i] / c6[i];
+                if (sigma6[i] < scParams.sigma6Minimum) /* for disappearing coul and vdw with soft core at the same time */
+                {
+                    sigma6[i] = scParams.sigma6Minimum;
+                }
             }
+            else
+            {
+                sigma6[i] = scParams.sigma6WithInvalidSigma;
+            }
+            sigma_pow[i] = sigma6[i];
         }
-        else
-        {
-            sigma6[i]       = sigma6_def;
-            sigma2[i]       = sigma2_def;
-        }
-        if (sc_r_power == six)
+        // NOLINTNEXTLINE(readability-misleading-indentation) remove when clang-tidy-13 is required
+        if constexpr (softcoreType == KernelSoftcoreType::Gapsys)
         {
-            sigma_pow[i]    = sigma6[i];
+            if ((c6[i] > 0) && (c12[i] > 0))
+            {
+                /* The c6 & c12 coefficients now contain the constants 6.0 and 12.0, respectively.
+                 * Correct for this by multiplying with (1/12.0)/(1/6.0)=6.0/12.0=0.5.
+                 */
+                gapsysSigma6VdW[i] = half * c12[i] / c6[i];
+            }
+            else
+            {
+                gapsysSigma6VdW[i] = scParams.gapsysSigma6VdW;
+            }
         }
-        else if (sc_r_power == fourtyeight)
+    }
+
+    // NOLINTNEXTLINE(readability-misleading-indentation) remove when clang-tidy-13 is required
+    if constexpr (softcoreType == KernelSoftcoreType::Beutler)
+    {
+        /* only use softcore if one of the states has a zero endstate - softcore is for avoiding infinities!*/
+        if ((c12[0] > 0) && (c12[1] > 0))
         {
-            sigma_pow[i]    = sigma6[i]*sigma6[i];       /* sigma^12 */
-            sigma_pow[i]    = sigma_pow[i]*sigma_pow[i]; /* sigma^24 */
-            sigma_pow[i]    = sigma_pow[i]*sigma_pow[i]; /* sigma^48 */
+            alpha_vdw_eff  = 0;
+            alpha_coul_eff = 0;
         }
         else
-        {       /* not really supported as input, but in here for testing the general case*/
-            sigma_pow[i]    = std::pow(sigma2[i], sc_r_power/2);
+        {
+            alpha_vdw_eff  = scParams.alphaVdw;
+            alpha_coul_eff = scParams.alphaCoulomb;
         }
     }
-
-    /* only use softcore if one of the states has a zero endstate - softcore is for avoiding infinities!*/
-    if ((c12[0] > 0) && (c12[1] > 0))
-    {
-        alpha_vdw_eff    = 0;
-        alpha_coul_eff   = 0;
-    }
-    else
+    // NOLINTNEXTLINE(readability-misleading-indentation) remove when clang-tidy-13 is required
+    if constexpr (softcoreType == KernelSoftcoreType::Gapsys)
     {
-        alpha_vdw_eff    = alpha_vdw;
-        alpha_coul_eff   = alpha_coul;
+        /* only use softcore if one of the states has a zero endstate - softcore is for avoiding infinities!*/
+        if ((c12[0] > 0) && (c12[1] > 0))
+        {
+            gapsysScaleLinpointVdW  = 0;
+            gapsysScaleLinpointCoul = 0;
+        }
+        else
+        {
+            gapsysScaleLinpointVdW  = scParams.gapsysScaleLinpointVdW;
+            gapsysScaleLinpointCoul = scParams.gapsysScaleLinpointCoul;
+        }
     }
 
     /* Loop over A and B states again */
+    // NOLINTNEXTLINE(readability-misleading-indentation) remove when clang-tidy-13 is required
     for (i = 0; i < 2; i++)
     {
         fscal_elec[i] = 0;
         fscal_vdw[i]  = 0;
         velec[i]      = 0;
         vvdw[i]       = 0;
+        dvdl_elec[i]  = 0;
+        dvdl_vdw[i]   = 0;
+        rQ            = 0;
+        rLJ           = 0;
 
         /* Only spend time on A or B state if it is non-zero */
-        if ( (qq[i] != 0) || (c6[i] != 0) || (c12[i] != 0) )
+        if ((qq[i] != 0) || (c6[i] != 0) || (c12[i] != 0))
         {
             /* Coulomb */
-            rpinv            = one/(alpha_coul_eff*lfac_coul[i]*sigma_pow[i]+rp);
-            r_coul           = std::pow(rpinv, minusOne / sc_r_power);
-
-            /* Electrostatics table lookup data */
-            rtab             = r_coul*tabscale;
-            ntab             = static_cast<int>(rtab);
-            eps              = rtab-ntab;
-            eps2             = eps*eps;
-            ntab             = static_cast<int>(tableStride*ntab);
-            /* Electrostatics */
-            Y                = vftab[ntab];
-            F                = vftab[ntab+1];
-            Geps             = eps*vftab[ntab+2];
-            Heps2            = eps2*vftab[ntab+3];
-            Fp               = F+Geps+Heps2;
-            VV               = Y+eps*Fp;
-            FF               = Fp+Geps+two*Heps2;
-            velec[i]         = qq[i]*VV;
-            fscal_elec[i]    = -qq[i]*FF*r_coul*rpinv*tabscale;
+            if constexpr (softcoreType == KernelSoftcoreType::Beutler)
+            {
+                rpinv  = one / (alpha_coul_eff * lfac_coul[i] * sigma_pow[i] + rp);
+                r_coul = sixthRoot(rpinv);
+            }
+            else
+            {
+                rpinv  = one / rp;
+                r_coul = r;
+            }
+
+            // NOLINTNEXTLINE(readability-misleading-indentation) remove when clang-tidy-13 is required
+            if constexpr (softcoreType == KernelSoftcoreType::Gapsys)
+            {
+                if ((facel != 0) && (LFC[i] < 1))
+                {
+                    rQ = gmx::sixthroot(one - LFC[i]) * (one + std::fabs(qq[i] / facel));
+                    rQ *= gapsysScaleLinpointCoul;
+                }
+                else
+                {
+                    rQ = 0;
+                }
+                scaleDvdlRCoul = 1;
+                if (rQ > rCoulCutoff)
+                {
+                    rQ             = rCoulCutoff;
+                    scaleDvdlRCoul = 0;
+                }
+            }
+
+            // NOLINTNEXTLINE(readability-misleading-indentation) remove when clang-tidy-13 is required
+            if ((softcoreType == KernelSoftcoreType::Gapsys) && (r < rQ))
+            {
+                real rInvQ    = one / rQ;
+                real constFac = qq[i] * rInvQ;
+                real linFac   = constFac * r * rInvQ;
+                real quadrFac = linFac * r * rInvQ;
+
+                /* Computing Coulomb force and potential energy */
+                fscal_elec[i] = -2 * quadrFac + 3 * linFac;
+                fscal_elec[i] *= rpinv;
+
+                velec[i] = quadrFac - 3 * (linFac - constFac);
+
+                dvdl_elec[i] += scaleDvdlRCoul * DLF[i] * half * (LFC[i] / (1 - LFC[i]))
+                                * (quadrFac - 2 * linFac + constFac);
+            }
+            else // Beutler, resp. hardcore
+            {
+                /* Electrostatics table lookup data */
+                rtab = r_coul * tabscale;
+                ntab = static_cast<int>(rtab);
+                eps  = rtab - ntab;
+                eps2 = eps * eps;
+                ntab = static_cast<int>(tableStride * ntab);
+                /* Electrostatics */
+                Y             = vftab[ntab];
+                F             = vftab[ntab + 1];
+                Geps          = eps * vftab[ntab + 2];
+                Heps2         = eps2 * vftab[ntab + 3];
+                Fp            = F + Geps + Heps2;
+                VV            = Y + eps * Fp;
+                FF            = Fp + Geps + two * Heps2;
+                velec[i]      = qq[i] * VV;
+                fscal_elec[i] = -qq[i] * FF * r_coul * rpinv * tabscale;
+            }
 
             /* Vdw */
-            rpinv            = one/(alpha_vdw_eff*lfac_vdw[i]*sigma_pow[i]+rp);
-            r_vdw            = std::pow(rpinv, minusOne / sc_r_power);
-            /* Vdw table lookup data */
-            rtab             = r_vdw*tabscale;
-            ntab             = static_cast<int>(rtab);
-            eps              = rtab-ntab;
-            eps2             = eps*eps;
-            ntab             = 12*ntab;
-            /* Dispersion */
-            Y                = vftab[ntab+4];
-            F                = vftab[ntab+5];
-            Geps             = eps*vftab[ntab+6];
-            Heps2            = eps2*vftab[ntab+7];
-            Fp               = F+Geps+Heps2;
-            VV               = Y+eps*Fp;
-            FF               = Fp+Geps+two*Heps2;
-            vvdw[i]          = c6[i]*VV;
-            fscal_vdw[i]     = -c6[i]*FF;
-
-            /* Repulsion */
-            Y                = vftab[ntab+8];
-            F                = vftab[ntab+9];
-            Geps             = eps*vftab[ntab+10];
-            Heps2            = eps2*vftab[ntab+11];
-            Fp               = F+Geps+Heps2;
-            VV               = Y+eps*Fp;
-            FF               = Fp+Geps+two*Heps2;
-            vvdw[i]         += c12[i]*VV;
-            fscal_vdw[i]    -= c12[i]*FF;
-            fscal_vdw[i]    *= r_vdw*rpinv*tabscale;
+            if constexpr (softcoreType == KernelSoftcoreType::Beutler)
+            {
+                rpinv = one / (alpha_vdw_eff * lfac_vdw[i] * sigma_pow[i] + rp);
+                r_vdw = sixthRoot(rpinv);
+            }
+            else
+            {
+                rpinv = one / rp;
+                r_vdw = r;
+            }
+
+            if constexpr (softcoreType == KernelSoftcoreType::Gapsys)
+            {
+                constexpr real c_twentySixSeventh = 26.0_real / 7.0_real;
+
+                if (LFV[i] < 1)
+                {
+
+                    rLJ = gmx::sixthroot(c_twentySixSeventh * gapsysSigma6VdW[i] * (one - LFV[i]));
+                    rLJ *= gapsysScaleLinpointVdW;
+                }
+                else
+                {
+                    rLJ = 0;
+                }
+            }
+
+            // NOLINTNEXTLINE(readability-misleading-indentation) remove when clang-tidy-13 is required
+            if ((softcoreType == KernelSoftcoreType::Gapsys) && (r < rLJ))
+            {
+                // scaled values for c6 and c12
+                real c6s, c12s;
+                c6s  = c6[i] / 6.0_real;
+                c12s = c12[i] / 12.0_real;
+
+                /* Temporary variables for inverted values */
+                real rInvLJ = one / rLJ;
+                real rInv14, rInv13, rInv12;
+                real rInv8, rInv7, rInv6;
+                rInv6 = rInvLJ * rInvLJ * rInvLJ;
+                rInv6 *= rInv6;
+                rInv7  = rInv6 * rInvLJ;
+                rInv8  = rInv7 * rInvLJ;
+                rInv14 = c12s * rInv7 * rInv7 * r2;
+                rInv13 = c12s * rInv7 * rInv6 * r;
+                rInv12 = c12s * rInv6 * rInv6;
+                rInv8 *= c6s * r2;
+                rInv7 *= c6s * r;
+                rInv6 *= c6s;
+
+                /* Temporary variables for A and B */
+                real quadrFac, linearFac, constFac;
+                quadrFac  = 156 * rInv14 - 42 * rInv8;
+                linearFac = 168 * rInv13 - 48 * rInv7;
+                constFac  = 91 * rInv12 - 28 * rInv6;
+
+                /* Computing LJ force and potential energy*/
+                fscal_vdw[i] = -quadrFac + linearFac;
+                fscal_vdw[i] *= rpinv;
+
+                vvdw[i] = 0.5_real * quadrFac - linearFac + constFac;
+
+                dvdl_vdw[i] += DLF[i] * 28 * (LFV[i] / (one - LFV[i]))
+                               * ((6.5_real * rInv14 - rInv8) - (13 * rInv13 - 2 * rInv7)
+                                  + (6.5_real * rInv12 - rInv6));
+            }
+            else // Beutler, resp. hardcore
+            {
+                /* Vdw table lookup data */
+                rtab = r_vdw * tabscale;
+                ntab = static_cast<int>(rtab);
+                eps  = rtab - ntab;
+                eps2 = eps * eps;
+                ntab = 12 * ntab;
+                /* Dispersion */
+                Y            = vftab[ntab + 4];
+                F            = vftab[ntab + 5];
+                Geps         = eps * vftab[ntab + 6];
+                Heps2        = eps2 * vftab[ntab + 7];
+                Fp           = F + Geps + Heps2;
+                VV           = Y + eps * Fp;
+                FF           = Fp + Geps + two * Heps2;
+                vvdw[i]      = c6[i] * VV;
+                fscal_vdw[i] = -c6[i] * FF;
+
+                /* Repulsion */
+                Y     = vftab[ntab + 8];
+                F     = vftab[ntab + 9];
+                Geps  = eps * vftab[ntab + 10];
+                Heps2 = eps2 * vftab[ntab + 11];
+                Fp    = F + Geps + Heps2;
+                VV    = Y + eps * Fp;
+                FF    = Fp + Geps + two * Heps2;
+                vvdw[i] += c12[i] * VV;
+                fscal_vdw[i] -= c12[i] * FF;
+                fscal_vdw[i] *= r_vdw * rpinv * tabscale;
+            }
         }
     }
     /* Now we have velec[i], vvdw[i], and fscal[i] for both states */
     /* Assemble A and B states */
-    velecsum  = 0;
-    vvdwsum   = 0;
-    dvdl_coul = 0;
-    dvdl_vdw  = 0;
-    fscal     = 0;
+    velecsum      = 0;
+    vvdwsum       = 0;
+    dvdl_coul_sum = 0;
+    dvdl_vdw_sum  = 0;
+    fscal         = 0;
     for (i = 0; i < 2; i++)
     {
-        velecsum      += LFC[i]*velec[i];
-        vvdwsum       += LFV[i]*vvdw[i];
+        velecsum += LFC[i] * velec[i];
+        vvdwsum += LFV[i] * vvdw[i];
 
-        fscal         += (LFC[i]*fscal_elec[i]+LFV[i]*fscal_vdw[i])*rpm2;
+        fscal += (LFC[i] * fscal_elec[i] + LFV[i] * fscal_vdw[i]) * rpm2;
 
-        dvdl_coul     += velec[i]*DLF[i] + LFC[i]*alpha_coul_eff*dlfac_coul[i]*fscal_elec[i]*sigma_pow[i];
-        dvdl_vdw      += vvdw[i]*DLF[i] + LFV[i]*alpha_vdw_eff*dlfac_vdw[i]*fscal_vdw[i]*sigma_pow[i];
+        if constexpr (softcoreType == KernelSoftcoreType::Gapsys)
+        {
+            dvdl_coul_sum += dvdl_elec[i];
+            dvdl_vdw_sum += dvdl_vdw[i];
+        }
+        // NOLINTNEXTLINE(readability-misleading-indentation) remove when clang-tidy-13 is required
+        dvdl_coul_sum += velec[i] * DLF[i];
+        dvdl_vdw_sum += vvdw[i] * DLF[i];
+        if constexpr (softcoreType == KernelSoftcoreType::Beutler)
+        {
+            dvdl_coul_sum += LFC[i] * alpha_coul_eff * dlfac_coul[i] * fscal_elec[i] * sigma_pow[i];
+            dvdl_vdw_sum += LFV[i] * alpha_vdw_eff * dlfac_vdw[i] * fscal_vdw[i] * sigma_pow[i];
+        }
     }
 
-    dvdl[efptCOUL]     += dvdl_coul;
-    dvdl[efptVDW]      += dvdl_vdw;
+    dvdl[static_cast<int>(FreeEnergyPerturbationCouplingType::Coul)] += dvdl_coul_sum;
+    dvdl[static_cast<int>(FreeEnergyPerturbationCouplingType::Vdw)] += dvdl_vdw_sum;
 
-    *velectot           = velecsum;
-    *vvdwtot            = vvdwsum;
+    *velectot = velecsum;
+    *vvdwtot  = vvdwsum;
 
     return fscal;
 }
 
 /*! \brief Calculate pair interactions, supports all types and conditions. */
-static real
-do_pairs_general(int ftype, int nbonds,
-                 const t_iatom iatoms[], const t_iparams iparams[],
-                 const rvec x[], rvec4 f[], rvec fshift[],
-                 const struct t_pbc *pbc, const struct t_graph *g,
-                 const real *lambda, real *dvdl,
-                 const t_mdatoms *md,
-                 const t_forcerec *fr, gmx_grppairener_t *grppener,
-                 int *global_atom_index)
+template<BondedKernelFlavor flavor>
+static real do_pairs_general(int                           ftype,
+                             int                           nbonds,
+                             const t_iatom                 iatoms[],
+                             const t_iparams               iparams[],
+                             const rvec                    x[],
+                             rvec4                         f[],
+                             rvec                          fshift[],
+                             const struct t_pbc*           pbc,
+                             const real*                   lambda,
+                             real*                         dvdl,
+                             gmx::ArrayRef<real>           chargeA,
+                             gmx::ArrayRef<real>           chargeB,
+                             gmx::ArrayRef<bool>           atomIsPerturbed,
+                             gmx::ArrayRef<unsigned short> cENER,
+                             int                           numEnergyGroups,
+                             const t_forcerec*             fr,
+                             gmx_grppairener_t*            grppener,
+                             int*                          global_atom_index)
 {
-    real             qq, c6, c12;
-    rvec             dx;
-    ivec             dt;
-    int              i, itype, ai, aj, gid;
-    int              fshift_index;
-    real             r2;
-    real             fscal, velec, vvdw;
-    real *           energygrp_elec;
-    real *           energygrp_vdw;
-    static gmx_bool  warned_rlimit = FALSE;
+    real            qq, c6, c12;
+    rvec            dx;
+    int             i, itype, ai, aj, gid;
+    int             fshift_index;
+    real            r2;
+    real            fscal, velec, vvdw;
+    real*           energygrp_elec;
+    real*           energygrp_vdw;
+    static gmx_bool warned_rlimit = FALSE;
     /* Free energy stuff */
-    gmx_bool         bFreeEnergy;
-    real             LFC[2], LFV[2], DLF[2], lfac_coul[2], lfac_vdw[2], dlfac_coul[2], dlfac_vdw[2];
-    real             qqB, c6B, c12B, sigma2_def, sigma2_min;
+    gmx_bool   bFreeEnergy;
+    real       LFC[2], LFV[2], DLF[2], lfac_coul[2], lfac_vdw[2], dlfac_coul[2], dlfac_vdw[2];
+    real       qqB, c6B, c12B;
+    const real oneSixth = 1.0_real / 6.0_real;
 
     switch (ftype)
     {
         case F_LJ14:
         case F_LJC14_Q:
-            energygrp_elec = grppener->ener[egCOUL14];
-            energygrp_vdw  = grppener->ener[egLJ14];
+            energygrp_elec = grppener->energyGroupPairTerms[NonBondedEnergyTerms::Coulomb14].data();
+            energygrp_vdw  = grppener->energyGroupPairTerms[NonBondedEnergyTerms::LJ14].data();
             break;
         case F_LJC_PAIRS_NB:
-            energygrp_elec = grppener->ener[egCOULSR];
-            energygrp_vdw  = grppener->ener[egLJSR];
+            energygrp_elec = grppener->energyGroupPairTerms[NonBondedEnergyTerms::CoulombSR].data();
+            energygrp_vdw  = grppener->energyGroupPairTerms[NonBondedEnergyTerms::LJSR].data();
             break;
         default:
             energygrp_elec = nullptr; /* Keep compiler happy */
@@ -379,74 +566,75 @@ do_pairs_general(int ftype, int nbonds,
             gmx_fatal(FARGS, "Unknown function type %d in do_nonbonded14", ftype);
     }
 
-    if (fr->efep != efepNO)
+    if (fr->efep != FreeEnergyPerturbationType::No)
     {
         /* Lambda factor for state A=1-lambda and B=lambda */
-        LFC[0] = 1.0 - lambda[efptCOUL];
-        LFV[0] = 1.0 - lambda[efptVDW];
-        LFC[1] = lambda[efptCOUL];
-        LFV[1] = lambda[efptVDW];
+        LFC[0] = 1.0 - lambda[static_cast<int>(FreeEnergyPerturbationCouplingType::Coul)];
+        LFV[0] = 1.0 - lambda[static_cast<int>(FreeEnergyPerturbationCouplingType::Vdw)];
+        LFC[1] = lambda[static_cast<int>(FreeEnergyPerturbationCouplingType::Coul)];
+        LFV[1] = lambda[static_cast<int>(FreeEnergyPerturbationCouplingType::Vdw)];
 
         /*derivative of the lambda factor for state A and B */
         DLF[0] = -1;
         DLF[1] = 1;
 
-        /* precalculate */
-        sigma2_def = std::cbrt(fr->sc_sigma6_def);
-        sigma2_min = std::cbrt(fr->sc_sigma6_min);
+        GMX_ASSERT(fr->ic->softCoreParameters, "We need soft-core parameters");
+        const auto& scParams = *fr->ic->softCoreParameters;
 
         for (i = 0; i < 2; i++)
         {
-            lfac_coul[i]  = (fr->sc_power == 2 ? (1-LFC[i])*(1-LFC[i]) : (1-LFC[i]));
-            dlfac_coul[i] = DLF[i]*fr->sc_power/fr->sc_r_power*(fr->sc_power == 2 ? (1-LFC[i]) : 1);
-            lfac_vdw[i]   = (fr->sc_power == 2 ? (1-LFV[i])*(1-LFV[i]) : (1-LFV[i]));
-            dlfac_vdw[i]  = DLF[i]*fr->sc_power/fr->sc_r_power*(fr->sc_power == 2 ? (1-LFV[i]) : 1);
+            lfac_coul[i] = (scParams.lambdaPower == 2 ? (1 - LFC[i]) * (1 - LFC[i]) : (1 - LFC[i]));
+            dlfac_coul[i] = DLF[i] * scParams.lambdaPower * oneSixth
+                            * (scParams.lambdaPower == 2 ? (1 - LFC[i]) : 1);
+            lfac_vdw[i]  = (scParams.lambdaPower == 2 ? (1 - LFV[i]) * (1 - LFV[i]) : (1 - LFV[i]));
+            dlfac_vdw[i] = DLF[i] * scParams.lambdaPower * oneSixth
+                           * (scParams.lambdaPower == 2 ? (1 - LFV[i]) : 1);
         }
     }
-    else
-    {
-        sigma2_min = sigma2_def = 0;
-    }
 
     /* TODO This code depends on the logic in tables.c that constructs
        the table layout, which should be made explicit in future
        cleanup. */
-    GMX_ASSERT(etiNR == 3, "Pair-interaction code that uses GROMACS interaction tables supports exactly 3 tables");
-    GMX_ASSERT(fr->pairsTable->interaction == GMX_TABLE_INTERACTION_ELEC_VDWREP_VDWDISP,
-               "Pair interaction kernels need a table with Coulomb, repulsion and dispersion entries");
+    GMX_ASSERT(
+            etiNR == 3,
+            "Pair-interaction code that uses GROMACS interaction tables supports exactly 3 tables");
+    GMX_ASSERT(
+            fr->pairsTable->interaction == TableInteraction::ElectrostaticVdwRepulsionVdwDispersion,
+            "Pair interaction kernels need a table with Coulomb, repulsion and dispersion entries");
 
     const real epsfac = fr->ic->epsfac;
 
     bFreeEnergy = FALSE;
-    for (i = 0; (i < nbonds); )
+    for (i = 0; (i < nbonds);)
     {
         itype = iatoms[i++];
         ai    = iatoms[i++];
         aj    = iatoms[i++];
-        gid   = GID(md->cENER[ai], md->cENER[aj], md->nenergrp);
+        gid   = GID(cENER[ai], cENER[aj], numEnergyGroups);
 
         /* Get parameters */
         switch (ftype)
         {
             case F_LJ14:
                 bFreeEnergy =
-                    (fr->efep != efepNO &&
-                     (((md->nPerturbed != 0) && (md->bPerturbed[ai] || md->bPerturbed[aj])) ||
-                      iparams[itype].lj14.c6A != iparams[itype].lj14.c6B ||
-                      iparams[itype].lj14.c12A != iparams[itype].lj14.c12B));
-                qq               = md->chargeA[ai]*md->chargeA[aj]*epsfac*fr->fudgeQQ;
-                c6               = iparams[itype].lj14.c6A;
-                c12              = iparams[itype].lj14.c12A;
+                        (fr->efep != FreeEnergyPerturbationType::No
+                         && ((!atomIsPerturbed.empty() && (atomIsPerturbed[ai] || atomIsPerturbed[aj]))
+                             || iparams[itype].lj14.c6A != iparams[itype].lj14.c6B
+                             || iparams[itype].lj14.c12A != iparams[itype].lj14.c12B));
+                qq  = chargeA[ai] * chargeA[aj] * epsfac * fr->fudgeQQ;
+                c6  = iparams[itype].lj14.c6A;
+                c12 = iparams[itype].lj14.c12A;
                 break;
             case F_LJC14_Q:
-                qq               = iparams[itype].ljc14.qi*iparams[itype].ljc14.qj*epsfac*iparams[itype].ljc14.fqq;
-                c6               = iparams[itype].ljc14.c6;
-                c12              = iparams[itype].ljc14.c12;
+                qq = iparams[itype].ljc14.qi * iparams[itype].ljc14.qj * epsfac
+                     * iparams[itype].ljc14.fqq;
+                c6  = iparams[itype].ljc14.c6;
+                c12 = iparams[itype].ljc14.c12;
                 break;
             case F_LJC_PAIRS_NB:
-                qq               = iparams[itype].ljcnb.qi*iparams[itype].ljcnb.qj*epsfac;
-                c6               = iparams[itype].ljcnb.c6;
-                c12              = iparams[itype].ljcnb.c12;
+                qq  = iparams[itype].ljcnb.qi * iparams[itype].ljcnb.qj * epsfac;
+                c6  = iparams[itype].ljcnb.c6;
+                c12 = iparams[itype].ljcnb.c12;
                 break;
             default:
                 /* Cannot happen since we called gmx_fatal() above in this case */
@@ -459,7 +647,7 @@ do_pairs_general(int ftype, int nbonds,
          * same factor, so when we use the original c6/c12 parameters from iparams[] they must
          * be scaled up.
          */
-        c6  *= 6.0;
+        c6 *= 6.0;
         c12 *= 12.0;
 
         /* Do we need to apply full periodic boundary conditions? */
@@ -469,18 +657,18 @@ do_pairs_general(int ftype, int nbonds,
         }
         else
         {
-            fshift_index = CENTRAL;
+            fshift_index = c_centralShiftIndex;
             rvec_sub(x[ai], x[aj], dx);
         }
-        r2           = norm2(dx);
+        r2 = norm2(dx);
 
-        if (r2 >= fr->pairsTable->r*fr->pairsTable->r)
+        if (r2 >= fr->pairsTable->interactionRange * fr->pairsTable->interactionRange)
         {
             /* This check isn't race free. But it doesn't matter because if a race occurs the only
              * disadvantage is that the warning is printed twice */
             if (!warned_rlimit)
             {
-                warning_rlimit(x, ai, aj, global_atom_index, sqrt(r2), fr->pairsTable->r);
+                warning_rlimit(x, ai, aj, global_atom_index, sqrt(r2), fr->pairsTable->interactionRange);
                 warned_rlimit = TRUE;
             }
             continue;
@@ -489,41 +677,155 @@ do_pairs_general(int ftype, int nbonds,
         if (bFreeEnergy)
         {
             /* Currently free energy is only supported for F_LJ14, so no need to check for that if we got here */
-            qqB              = md->chargeB[ai]*md->chargeB[aj]*epsfac*fr->fudgeQQ;
-            c6B              = iparams[itype].lj14.c6B*6.0;
-            c12B             = iparams[itype].lj14.c12B*12.0;
-
-            fscal            = free_energy_evaluate_single(r2, fr->sc_r_power, fr->sc_alphacoul, fr->sc_alphavdw,
-                                                           fr->pairsTable->scale, fr->pairsTable->data, fr->pairsTable->stride,
-                                                           qq, c6, c12, qqB, c6B, c12B,
-                                                           LFC, LFV, DLF, lfac_coul, lfac_vdw, dlfac_coul, dlfac_vdw,
-                                                           fr->sc_sigma6_def, fr->sc_sigma6_min, sigma2_def, sigma2_min, &velec, &vvdw, dvdl);
+            qqB  = chargeB[ai] * chargeB[aj] * epsfac * fr->fudgeQQ;
+            c6B  = iparams[itype].lj14.c6B * 6.0;
+            c12B = iparams[itype].lj14.c12B * 12.0;
+
+            const auto& scParams = *fr->ic->softCoreParameters;
+            if (scParams.softcoreType == SoftcoreType::Beutler)
+            {
+                if (scParams.alphaCoulomb == 0 && scParams.alphaVdw == 0)
+                {
+                    fscal = free_energy_evaluate_single<KernelSoftcoreType::None>(
+                            r2,
+                            fr->ic->rcoulomb,
+                            *fr->ic->softCoreParameters,
+                            fr->pairsTable->scale,
+                            fr->pairsTable->data.data(),
+                            fr->pairsTable->stride,
+                            qq,
+                            c6,
+                            c12,
+                            qqB,
+                            c6B,
+                            c12B,
+                            epsfac,
+                            LFC,
+                            LFV,
+                            DLF,
+                            lfac_coul,
+                            lfac_vdw,
+                            dlfac_coul,
+                            dlfac_vdw,
+                            &velec,
+                            &vvdw,
+                            dvdl);
+                }
+                else
+                {
+                    fscal = free_energy_evaluate_single<KernelSoftcoreType::Beutler>(
+                            r2,
+                            fr->ic->rcoulomb,
+                            *fr->ic->softCoreParameters,
+                            fr->pairsTable->scale,
+                            fr->pairsTable->data.data(),
+                            fr->pairsTable->stride,
+                            qq,
+                            c6,
+                            c12,
+                            qqB,
+                            c6B,
+                            c12B,
+                            epsfac,
+                            LFC,
+                            LFV,
+                            DLF,
+                            lfac_coul,
+                            lfac_vdw,
+                            dlfac_coul,
+                            dlfac_vdw,
+                            &velec,
+                            &vvdw,
+                            dvdl);
+                }
+            }
+            else // Gapsys
+            {
+                if (scParams.gapsysScaleLinpointCoul == 0 && scParams.gapsysScaleLinpointVdW == 0)
+                {
+                    fscal = free_energy_evaluate_single<KernelSoftcoreType::None>(
+                            r2,
+                            fr->ic->rcoulomb,
+                            *fr->ic->softCoreParameters,
+                            fr->pairsTable->scale,
+                            fr->pairsTable->data.data(),
+                            fr->pairsTable->stride,
+                            qq,
+                            c6,
+                            c12,
+                            qqB,
+                            c6B,
+                            c12B,
+                            epsfac,
+                            LFC,
+                            LFV,
+                            DLF,
+                            lfac_coul,
+                            lfac_vdw,
+                            dlfac_coul,
+                            dlfac_vdw,
+                            &velec,
+                            &vvdw,
+                            dvdl);
+                }
+                else
+                {
+                    fscal = free_energy_evaluate_single<KernelSoftcoreType::Gapsys>(
+                            r2,
+                            fr->ic->rcoulomb,
+                            *fr->ic->softCoreParameters,
+                            fr->pairsTable->scale,
+                            fr->pairsTable->data.data(),
+                            fr->pairsTable->stride,
+                            qq,
+                            c6,
+                            c12,
+                            qqB,
+                            c6B,
+                            c12B,
+                            epsfac,
+                            LFC,
+                            LFV,
+                            DLF,
+                            lfac_coul,
+                            lfac_vdw,
+                            dlfac_coul,
+                            dlfac_vdw,
+                            &velec,
+                            &vvdw,
+                            dvdl);
+                }
+            }
         }
         else
         {
             /* Evaluate tabulated interaction without free energy */
-            fscal            = evaluate_single(r2, fr->pairsTable->scale, fr->pairsTable->data, fr->pairsTable->stride,
-                                               qq, c6, c12, &velec, &vvdw);
+            fscal = evaluate_single(r2,
+                                    fr->pairsTable->scale,
+                                    fr->pairsTable->data.data(),
+                                    fr->pairsTable->stride,
+                                    qq,
+                                    c6,
+                                    c12,
+                                    &velec,
+                                    &vvdw);
         }
 
-        energygrp_elec[gid]  += velec;
-        energygrp_vdw[gid]   += vvdw;
+        energygrp_elec[gid] += velec;
+        energygrp_vdw[gid] += vvdw;
         svmul(fscal, dx, dx);
 
         /* Add the forces */
         rvec_inc(f[ai], dx);
         rvec_dec(f[aj], dx);
 
-        if (g)
-        {
-            /* Correct the shift forces using the graph */
-            ivec_sub(SHIFT_IVEC(g, ai), SHIFT_IVEC(g, aj), dt);
-            fshift_index = IVEC2IS(dt);
-        }
-        if (fshift_index != CENTRAL)
+        if (computeVirial(flavor))
         {
-            rvec_inc(fshift[fshift_index], dx);
-            rvec_dec(fshift[CENTRAL], dx);
+            if (fshift_index != c_centralShiftIndex)
+            {
+                rvec_inc(fshift[fshift_index], dx);
+                rvec_dec(fshift[c_centralShiftIndex], dx);
+            }
         }
     }
     return 0.0;
@@ -533,34 +835,34 @@ do_pairs_general(int ftype, int nbonds,
  *
  * This function is templated for real/SimdReal and for optimization.
  */
-template<typename T, int pack_size,
-         typename pbc_type>
-static void
-do_pairs_simple(int nbonds,
-                const t_iatom iatoms[], const t_iparams iparams[],
-                const rvec x[], rvec4 f[],
-                const pbc_type pbc,
-                const t_mdatoms *md,
-                const real scale_factor)
+template<typename T, int pack_size, typename pbc_type>
+static void do_pairs_simple(int                 nbonds,
+                            const t_iatom       iatoms[],
+                            const t_iparams     iparams[],
+                            const rvec          x[],
+                            rvec4               f[],
+                            const pbc_type      pbc,
+                            gmx::ArrayRef<real> charge,
+                            const real          scale_factor)
 {
     const int nfa1 = 1 + 2;
 
-    T         six(6);
-    T         twelve(12);
-    T         ef(scale_factor);
+    T six(6);
+    T twelve(12);
+    T ef(scale_factor);
 
 #if GMX_SIMD_HAVE_REAL
-    alignas(GMX_SIMD_ALIGNMENT) std::int32_t  ai[pack_size];
-    alignas(GMX_SIMD_ALIGNMENT) std::int32_t  aj[pack_size];
-    alignas(GMX_SIMD_ALIGNMENT) real          coeff[3*pack_size];
+    alignas(GMX_SIMD_ALIGNMENT) std::int32_t ai[pack_size];
+    alignas(GMX_SIMD_ALIGNMENT) std::int32_t aj[pack_size];
+    alignas(GMX_SIMD_ALIGNMENT) real         coeff[3 * pack_size];
 #else
-    std::int32_t   ai[pack_size];
-    std::int32_t   aj[pack_size];
-    real           coeff[3*pack_size];
+    std::int32_t ai[pack_size];
+    std::int32_t aj[pack_size];
+    real         coeff[3 * pack_size];
 #endif
 
     /* nbonds is #pairs*nfa1, here we step pack_size pairs */
-    for (int i = 0; i < nbonds; i += pack_size*nfa1)
+    for (int i = 0; i < nbonds; i += pack_size * nfa1)
     {
         /* Collect atoms for pack_size pairs.
          * iu indexes into iatoms, we should not let iu go beyond nbonds.
@@ -572,11 +874,11 @@ do_pairs_simple(int nbonds,
             ai[s]     = iatoms[iu + 1];
             aj[s]     = iatoms[iu + 2];
 
-            if (i + s*nfa1 < nbonds)
+            if (i + s * nfa1 < nbonds)
             {
-                coeff[0*pack_size + s] = iparams[itype].lj14.c6A;
-                coeff[1*pack_size + s] = iparams[itype].lj14.c12A;
-                coeff[2*pack_size + s] = md->chargeA[ai[s]]*md->chargeA[aj[s]];
+                coeff[0 * pack_size + s] = iparams[itype].lj14.c6A;
+                coeff[1 * pack_size + s] = iparams[itype].lj14.c12A;
+                coeff[2 * pack_size + s] = charge[ai[s]] * charge[aj[s]];
 
                 /* Avoid indexing the iatoms array out of bounds.
                  * We pad the coordinate indices with the last atom pair.
@@ -589,69 +891,77 @@ do_pairs_simple(int nbonds,
             else
             {
                 /* Pad the coefficient arrays with zeros to get zero forces */
-                coeff[0*pack_size + s] = 0;
-                coeff[1*pack_size + s] = 0;
-                coeff[2*pack_size + s] = 0;
+                coeff[0 * pack_size + s] = 0;
+                coeff[1 * pack_size + s] = 0;
+                coeff[2 * pack_size + s] = 0;
             }
         }
 
         /* Load the coordinates */
         T xi[DIM], xj[DIM];
-        gatherLoadUTranspose<3>(reinterpret_cast<const real *>(x), ai, &xi[XX], &xi[YY], &xi[ZZ]);
-        gatherLoadUTranspose<3>(reinterpret_cast<const real *>(x), aj, &xj[XX], &xj[YY], &xj[ZZ]);
+        gatherLoadUTranspose<3>(reinterpret_cast<const real*>(x), ai, &xi[XX], &xi[YY], &xi[ZZ]);
+        gatherLoadUTranspose<3>(reinterpret_cast<const real*>(x), aj, &xj[XX], &xj[YY], &xj[ZZ]);
 
-        T c6    = load<T>(coeff + 0*pack_size);
-        T c12   = load<T>(coeff + 1*pack_size);
-        T qq    = load<T>(coeff + 2*pack_size);
+        T c6  = load<T>(coeff + 0 * pack_size);
+        T c12 = load<T>(coeff + 1 * pack_size);
+        T qq  = load<T>(coeff + 2 * pack_size);
 
         /* We could save these operations by storing 6*C6,12*C12 */
-        c6             = six*c6;
-        c12            = twelve*c12;
+        c6  = six * c6;
+        c12 = twelve * c12;
 
         T dr[DIM];
         pbc_dx_aiuc(pbc, xi, xj, dr);
 
-        T rsq   = dr[XX]*dr[XX] + dr[YY]*dr[YY] + dr[ZZ]*dr[ZZ];
+        T rsq   = dr[XX] * dr[XX] + dr[YY] * dr[YY] + dr[ZZ] * dr[ZZ];
         T rinv  = gmx::invsqrt(rsq);
-        T rinv2 = rinv*rinv;
-        T rinv6 = rinv2*rinv2*rinv2;
+        T rinv2 = rinv * rinv;
+        T rinv6 = rinv2 * rinv2 * rinv2;
 
         /* Calculate the Coulomb force * r */
-        T cfr   = ef*qq*rinv;
+        T cfr = ef * qq * rinv;
 
         /* Calculate the LJ force * r and add it to the Coulomb part */
-        T fr    = gmx::fma(fms(c12, rinv6, c6), rinv6, cfr);
+        T fr = gmx::fma(fms(c12, rinv6, c6), rinv6, cfr);
 
-        T finvr = fr*rinv2;
-        T fx    = finvr*dr[XX];
-        T fy    = finvr*dr[YY];
-        T fz    = finvr*dr[ZZ];
+        T finvr = fr * rinv2;
+        T fx    = finvr * dr[XX];
+        T fy    = finvr * dr[YY];
+        T fz    = finvr * dr[ZZ];
 
         /* Add the pair forces to the force array.
          * Note that here we might add multiple force components for some atoms
          * due to the SIMD padding. But the extra force components are zero.
          */
-        transposeScatterIncrU<4>(reinterpret_cast<real *>(f), ai, fx, fy, fz);
-        transposeScatterDecrU<4>(reinterpret_cast<real *>(f), aj, fx, fy, fz);
+        transposeScatterIncrU<4>(reinterpret_cast<real*>(f), ai, fx, fy, fz);
+        transposeScatterDecrU<4>(reinterpret_cast<real*>(f), aj, fx, fy, fz);
     }
 }
 
 /*! \brief Calculate all listed pair interactions */
-void
-do_pairs(int ftype, int nbonds,
-         const t_iatom iatoms[], const t_iparams iparams[],
-         const rvec x[], rvec4 f[], rvec fshift[],
-         const struct t_pbc *pbc, const struct t_graph *g,
-         const real *lambda, real *dvdl,
-         const t_mdatoms *md,
-         const t_forcerec *fr,
-         const gmx_bool computeForcesOnly,
-         gmx_grppairener_t *grppener,
-         int *global_atom_index)
+void do_pairs(int                           ftype,
+              int                           nbonds,
+              const t_iatom                 iatoms[],
+              const t_iparams               iparams[],
+              const rvec                    x[],
+              rvec4                         f[],
+              rvec                          fshift[],
+              const struct t_pbc*           pbc,
+              const real*                   lambda,
+              real*                         dvdl,
+              gmx::ArrayRef<real>           chargeA,
+              gmx::ArrayRef<real>           chargeB,
+              gmx::ArrayRef<bool>           atomIsPerturbed,
+              gmx::ArrayRef<unsigned short> cENER,
+              const int                     numEnergyGroups,
+              const t_forcerec*             fr,
+              const bool                    havePerturbedInteractions,
+              const gmx::StepWorkload&      stepWork,
+              gmx_grppairener_t*            grppener,
+              int*                          global_atom_index)
 {
-    if (ftype == F_LJ14 &&
-        fr->ic->vdwtype != evdwUSER && !EEL_USER(fr->ic->eeltype) &&
-        computeForcesOnly)
+    if (ftype == F_LJ14 && fr->ic->vdwtype != VanDerWaalsType::User && !EEL_USER(fr->ic->eeltype)
+        && !havePerturbedInteractions && (!stepWork.computeVirial && !stepWork.computeEnergy))
     {
         /* We use a fast code-path for plain LJ 1-4 without FEP.
          *
@@ -661,40 +971,76 @@ do_pairs(int ftype, int nbonds,
          * and sum the virial for the shifts. But we should do this
          * at once for the angles and dihedrals as well.
          */
-#if GMX_SIMD
-        alignas(GMX_SIMD_ALIGNMENT) real pbc_simd[9*GMX_SIMD_REAL_WIDTH];
-        set_pbc_simd(pbc, pbc_simd);
-
-        do_pairs_simple<SimdReal, GMX_SIMD_REAL_WIDTH,
-                        const real *>(nbonds, iatoms, iparams,
-                                      x, f, pbc_simd,
-                                      md, fr->ic->epsfac*fr->fudgeQQ);
-#else
-        /* This construct is needed because pbc_dx_aiuc doesn't accept pbc=NULL */
-        t_pbc        pbc_no;
-        const t_pbc *pbc_nonnull;
-
-        if (pbc != nullptr)
+#if GMX_SIMD_HAVE_REAL
+        if (fr->use_simd_kernels)
         {
-            pbc_nonnull   = pbc;
+            alignas(GMX_SIMD_ALIGNMENT) real pbc_simd[9 * GMX_SIMD_REAL_WIDTH];
+            set_pbc_simd(pbc, pbc_simd);
+
+            do_pairs_simple<SimdReal, GMX_SIMD_REAL_WIDTH, const real*>(
+                    nbonds, iatoms, iparams, x, f, pbc_simd, chargeA, fr->ic->epsfac * fr->fudgeQQ);
         }
         else
+#endif
         {
-            set_pbc(&pbc_no, epbcNONE, nullptr);
-            pbc_nonnull   = &pbc_no;
-        }
+            /* This construct is needed because pbc_dx_aiuc doesn't accept pbc=NULL */
+            t_pbc        pbc_no;
+            const t_pbc* pbc_nonnull;
 
-        do_pairs_simple<real, 1,
-                        const t_pbc *>(nbonds, iatoms, iparams,
-                                       x, f, pbc_nonnull,
-                                       md, fr->ic->epsfac*fr->fudgeQQ);
-#endif
+            if (pbc != nullptr)
+            {
+                pbc_nonnull = pbc;
+            }
+            else
+            {
+                set_pbc(&pbc_no, PbcType::No, nullptr);
+                pbc_nonnull = &pbc_no;
+            }
+
+            do_pairs_simple<real, 1, const t_pbc*>(
+                    nbonds, iatoms, iparams, x, f, pbc_nonnull, chargeA, fr->ic->epsfac * fr->fudgeQQ);
+        }
+    }
+    else if (stepWork.computeVirial)
+    {
+        do_pairs_general<BondedKernelFlavor::ForcesAndVirialAndEnergy>(ftype,
+                                                                       nbonds,
+                                                                       iatoms,
+                                                                       iparams,
+                                                                       x,
+                                                                       f,
+                                                                       fshift,
+                                                                       pbc,
+                                                                       lambda,
+                                                                       dvdl,
+                                                                       chargeA,
+                                                                       chargeB,
+                                                                       atomIsPerturbed,
+                                                                       cENER,
+                                                                       numEnergyGroups,
+                                                                       fr,
+                                                                       grppener,
+                                                                       global_atom_index);
     }
     else
     {
-        do_pairs_general(ftype, nbonds, iatoms, iparams,
-                         x, f, fshift, pbc, g,
-                         lambda, dvdl,
-                         md, fr, grppener, global_atom_index);
+        do_pairs_general<BondedKernelFlavor::ForcesAndEnergy>(ftype,
+                                                              nbonds,
+                                                              iatoms,
+                                                              iparams,
+                                                              x,
+                                                              f,
+                                                              fshift,
+                                                              pbc,
+                                                              lambda,
+                                                              dvdl,
+                                                              chargeA,
+                                                              chargeB,
+                                                              atomIsPerturbed,
+                                                              cENER,
+                                                              numEnergyGroups,
+                                                              fr,
+                                                              grppener,
+                                                              global_atom_index);
     }
 }