Move responsibility for bonded threading decomposition

[alexxy/gromacs.git] / src / gromacs / listed-forces / manage-threading.cpp

/*
 * This file is part of the GROMACS molecular simulation package.
 *
 * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
 * Copyright (c) 2001-2004, The GROMACS development team.
 * Copyright (c) 2013,2014,2015,2016,2017,2018, 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.
 *
 * GROMACS is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public License
 * as published by the Free Software Foundation; either version 2.1
 * of the License, or (at your option) any later version.
 *
 * GROMACS is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with GROMACS; if not, see
 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
 *
 * If you want to redistribute modifications to GROMACS, please
 * consider that scientific software is very special. Version
 * control is crucial - bugs must be traceable. We will be happy to
 * consider code for inclusion in the official distribution, but
 * derived work must not be called official GROMACS. Details are found
 * in the README & COPYING files - if they are missing, get the
 * official version at http://www.gromacs.org.
 *
 * To help us fund GROMACS development, we humbly ask that you cite
 * the research papers on the package. Check out http://www.gromacs.org.
 */
/*! \internal \file
 * \brief This file defines functions for managing threading of listed
 * interactions.
 *
 * \author Mark Abraham <mark.j.abraham@gmail.com>
 * \ingroup module_listed-forces
 */
#include "gmxpre.h"

#include "manage-threading.h"

#include "config.h"

#include <assert.h>
#include <limits.h>
#include <stdlib.h>

#include <algorithm>

#include "gromacs/listed-forces/listed-forces.h"
#include "gromacs/mdlib/gmx_omp_nthreads.h"
#include "gromacs/pbcutil/ishift.h"
#include "gromacs/topology/ifunc.h"
#include "gromacs/utility/exceptions.h"
#include "gromacs/utility/fatalerror.h"
#include "gromacs/utility/smalloc.h"
#include "gromacs/utility/stringutil.h"

#include "listed-internal.h"

/*! \brief struct for passing all data required for a function type */
typedef struct {
    int            ftype; /**< the function type index */
    const t_ilist *il;    /**< pointer to t_ilist entry corresponding to ftype */
    int            nat;   /**< nr of atoms involved in a single ftype interaction */
} ilist_data_t;

/*! \brief Divides listed interactions over threads
 *
 * This routine attempts to divide all interactions of the ntype bondeds
 * types stored in ild over the threads such that each thread has roughly
 * equal load and different threads avoid touching the same atoms as much
 * as possible.
 */
static void divide_bondeds_by_locality(bonded_threading_t *bt,
                                       int                 ntype,
                                       const ilist_data_t *ild)
{
    int nat_tot, nat_sum;
    int ind[F_NRE];    /* index into the ild[].il->iatoms */
    int at_ind[F_NRE]; /* index of the first atom of the interaction at ind */
    int f, t;

    assert(ntype <= F_NRE);

    nat_tot = 0;
    for (f = 0; f < ntype; f++)
    {
        /* Sum #bondeds*#atoms_per_bond over all bonded types */
        nat_tot  += ild[f].il->nr/(ild[f].nat + 1)*ild[f].nat;
        /* The start bound for thread 0 is 0 for all interactions */
        ind[f]    = 0;
        /* Initialize the next atom index array */
        assert(ild[f].il->nr > 0);
        at_ind[f] = ild[f].il->iatoms[1];
    }

    nat_sum = 0;
    /* Loop over the end bounds of the nthreads threads to determine
     * which interactions threads 0 to nthreads shall calculate.
     *
     * NOTE: The cost of these combined loops is #interactions*ntype.
     * This code is running single threaded (difficult to parallelize
     * over threads). So the relative cost of this function increases
     * linearly with the number of threads. Since the inner-most loop
     * is cheap and this is done only at DD repartitioning, the cost should
     * be negligble. At high thread count many other parts of the code
     * scale the same way, so it's (currently) not worth improving this.
     */
    for (t = 1; t <= bt->nthreads; t++)
    {
        int nat_thread;

        /* Here we assume that the computational cost is proportional
         * to the number of atoms in the interaction. This is a rough
         * measure, but roughly correct. Usually there are very few
         * interactions anyhow and there are distributed relatively
         * uniformly. Proper and RB dihedrals are often distributed
         * non-uniformly, but their cost is roughly equal.
         */
        nat_thread = (nat_tot*t)/bt->nthreads;

        while (nat_sum < nat_thread)
        {
            /* To divide bonds based on atom order, we compare
             * the index of the first atom in the bonded interaction.
             * This works well, since the domain decomposition generates
             * bondeds in order of the atoms by looking up interactions
             * which are linked to the first atom in each interaction.
             * It usually also works well without DD, since than the atoms
             * in bonded interactions are usually in increasing order.
             * If they are not assigned in increasing order, the balancing
             * is still good, but the memory access and reduction cost will
             * be higher.
             */
            int f_min;

            /* Find out which of the types has the lowest atom index */
            f_min = 0;
            for (f = 1; f < ntype; f++)
            {
                if (at_ind[f] < at_ind[f_min])
                {
                    f_min = f;
                }
            }
            assert(f_min >= 0 && f_min < ntype);

            /* Assign the interaction with the lowest atom index (of type
             * index f_min) to thread t-1 by increasing ind.
             */
            ind[f_min] += ild[f_min].nat + 1;
            nat_sum    += ild[f_min].nat;

            /* Update the first unassigned atom index for this type */
            if (ind[f_min] < ild[f_min].il->nr)
            {
                at_ind[f_min] = ild[f_min].il->iatoms[ind[f_min] + 1];
            }
            else
            {
                /* We have assigned all interactions of this type.
                 * Setting at_ind to INT_MAX ensures this type will not be
                 * chosen in the for loop above during next iterations.
                 */
                at_ind[f_min] = INT_MAX;
            }
        }

        /* Store the bonded end boundaries (at index t) for thread t-1 */
        for (f = 0; f < ntype; f++)
        {
            bt->il_thread_division[ild[f].ftype*(bt->nthreads + 1) + t] = ind[f];
        }
    }

    for (f = 0; f < ntype; f++)
    {
        assert(ind[f] == ild[f].il->nr);
    }
}

//! Divides bonded interactions over threads
static void divide_bondeds_over_threads(bonded_threading_t *bt,
                                        const t_idef       *idef)
{
    ilist_data_t ild[F_NRE];
    int          ntype;
    int          f;

    assert(bt->nthreads > 0);

    if (F_NRE*(bt->nthreads + 1) > bt->il_thread_division_nalloc)
    {
        bt->il_thread_division_nalloc = F_NRE*(bt->nthreads + 1);
        srenew(bt->il_thread_division, bt->il_thread_division_nalloc);
    }

    bt->haveBondeds = false;
    ntype           = 0;
    for (f = 0; f < F_NRE; f++)
    {
        if (!ftype_is_bonded_potential(f))
        {
            continue;
        }

        if (idef->il[f].nr > 0)
        {
            bt->haveBondeds = true;
        }

        if (idef->il[f].nr == 0)
        {
            /* No interactions, avoid all the integer math below */
            int t;
            for (t = 0; t <= bt->nthreads; t++)
            {
                bt->il_thread_division[f*(bt->nthreads + 1) + t] = 0;
            }
        }
        else if (bt->nthreads <= bt->max_nthread_uniform || f == F_DISRES)
        {
            /* On up to 4 threads, load balancing the bonded work
             * is more important than minimizing the reduction cost.
             */
            int nat1, t;

            /* nat1 = 1 + #atoms(ftype) which is the stride use for iatoms */
            nat1 = 1 + NRAL(f);

            for (t = 0; t <= bt->nthreads; t++)
            {
                int nr_t;

                /* Divide equally over the threads */
                nr_t = (((idef->il[f].nr/nat1)*t)/bt->nthreads)*nat1;

                if (f == F_DISRES)
                {
                    /* Ensure that distance restraint pairs with the same label
                     * end up on the same thread.
                     */
                    while (nr_t > 0 && nr_t < idef->il[f].nr &&
                           idef->iparams[idef->il[f].iatoms[nr_t]].disres.label ==
                           idef->iparams[idef->il[f].iatoms[nr_t-nat1]].disres.label)
                    {
                        nr_t += nat1;
                    }
                }

                bt->il_thread_division[f*(bt->nthreads + 1) + t] = nr_t;
            }
        }
        else
        {
            /* Add this ftype to the list to be distributed */
            int nat;

            nat              = NRAL(f);
            ild[ntype].ftype = f;
            ild[ntype].il    = &idef->il[f];
            ild[ntype].nat   = nat;

            /* The first index for the thread division is always 0 */
            bt->il_thread_division[f*(bt->nthreads + 1)] = 0;

            ntype++;
        }
    }

    if (ntype > 0)
    {
        divide_bondeds_by_locality(bt, ntype, ild);
    }

    if (debug)
    {
        int f;

        fprintf(debug, "Division of bondeds over threads:\n");
        for (f = 0; f < F_NRE; f++)
        {
            if (ftype_is_bonded_potential(f) && idef->il[f].nr > 0)
            {
                int t;

                fprintf(debug, "%16s", interaction_function[f].name);
                for (t = 0; t < bt->nthreads; t++)
                {
                    fprintf(debug, " %4d",
                            (bt->il_thread_division[f*(bt->nthreads + 1) + t + 1] -
                             bt->il_thread_division[f*(bt->nthreads + 1) + t])/
                            (1 + NRAL(f)));
                }
                fprintf(debug, "\n");
            }
        }
    }
}

//! Construct a reduction mask for which parts (blocks) of the force array are touched on which thread task
static void
calc_bonded_reduction_mask(int                       natoms,
                           f_thread_t               *f_thread,
                           const t_idef             *idef,
                           int                       thread,
                           const bonded_threading_t &bondedThreading)
{
    static_assert(BITMASK_SIZE == GMX_OPENMP_MAX_THREADS, "For the error message below we assume these two are equal.");

    if (bondedThreading.nthreads > BITMASK_SIZE)
    {
#pragma omp master
        gmx_fatal(FARGS, "You are using %d OpenMP threads, which is larger than GMX_OPENMP_MAX_THREADS (%d). Decrease the number of OpenMP threads or rebuild GROMACS with a larger value for GMX_OPENMP_MAX_THREADS.",
                  bondedThreading.nthreads, GMX_OPENMP_MAX_THREADS);
#pragma omp barrier
    }
    GMX_ASSERT(bondedThreading.nthreads <= BITMASK_SIZE, "We need at least nthreads bits in the mask");

    int nblock = (natoms + reduction_block_size - 1) >> reduction_block_bits;

    if (nblock > f_thread->block_nalloc)
    {
        f_thread->block_nalloc = over_alloc_large(nblock);
        srenew(f_thread->mask,        f_thread->block_nalloc);
        srenew(f_thread->block_index, f_thread->block_nalloc);
        sfree_aligned(f_thread->f);
        snew_aligned(f_thread->f,     f_thread->block_nalloc*reduction_block_size, 128);
    }

    gmx_bitmask_t *mask = f_thread->mask;

    for (int b = 0; b < nblock; b++)
    {
        bitmask_clear(&mask[b]);
    }

    for (int ftype = 0; ftype < F_NRE; ftype++)
    {
        if (ftype_is_bonded_potential(ftype))
        {
            int nb = idef->il[ftype].nr;
            if (nb > 0)
            {
                int nat1 = interaction_function[ftype].nratoms + 1;

                int nb0 = bondedThreading.il_thread_division[ftype*(bondedThreading.nthreads + 1) + thread];
                int nb1 = bondedThreading.il_thread_division[ftype*(bondedThreading.nthreads + 1) + thread + 1];

                for (int i = nb0; i < nb1; i += nat1)
                {
                    for (int a = 1; a < nat1; a++)
                    {
                        bitmask_set_bit(&mask[idef->il[ftype].iatoms[i+a] >> reduction_block_bits], thread);
                    }
                }
            }
        }
    }

    /* Make an index of the blocks our thread touches, so we can do fast
     * force buffer clearing.
     */
    f_thread->nblock_used = 0;
    for (int b = 0; b < nblock; b++)
    {
        if (bitmask_is_set(mask[b], thread))
        {
            f_thread->block_index[f_thread->nblock_used++] = b;
        }
    }
}

void setup_bonded_threading(bonded_threading_t *bt,
                            int                 numAtoms,
                            const t_idef       *idef)
{
    int                 ctot = 0;

    assert(bt->nthreads >= 1);

    /* Divide the bonded interaction over the threads */
    divide_bondeds_over_threads(bt, idef);

    if (!bt->haveBondeds)
    {
        /* We don't have bondeds, so there is nothing to reduce */
        return;
    }

    /* Determine to which blocks each thread's bonded force calculation
     * contributes. Store this as a mask for each thread.
     */
#pragma omp parallel for num_threads(bt->nthreads) schedule(static)
    for (int t = 0; t < bt->nthreads; t++)
    {
        try
        {
            calc_bonded_reduction_mask(numAtoms, &bt->f_t[t],
                                       idef, t, *bt);
        }
        GMX_CATCH_ALL_AND_EXIT_WITH_FATAL_ERROR;
    }

    /* Reduce the masks over the threads and determine which blocks
     * we need to reduce over.
     */
    int nblock_tot = (numAtoms + reduction_block_size - 1) >> reduction_block_bits;
    if (nblock_tot > bt->block_nalloc)
    {
        bt->block_nalloc = over_alloc_large(nblock_tot);
        srenew(bt->block_index, bt->block_nalloc);
        srenew(bt->mask,        bt->block_nalloc);
    }
    bt->nblock_used = 0;
    for (int b = 0; b < nblock_tot; b++)
    {
        gmx_bitmask_t *mask = &bt->mask[b];

        /* Generate the union over the threads of the bitmask */
        bitmask_clear(mask);
        for (int t = 0; t < bt->nthreads; t++)
        {
            bitmask_union(mask, bt->f_t[t].mask[b]);
        }
        if (!bitmask_is_zero(*mask))
        {
            bt->block_index[bt->nblock_used++] = b;
        }

        if (debug)
        {
            int c = 0;
            for (int t = 0; t < bt->nthreads; t++)
            {
                if (bitmask_is_set(*mask, t))
                {
                    c++;
                }
            }
            ctot += c;

            if (gmx_debug_at)
            {
#if BITMASK_SIZE <= 64 //move into bitmask when it is C++
                std::string flags = gmx::formatString("%x", *mask);
#else
                std::string flags = gmx::formatAndJoin(*mask,
                                                       *mask+BITMASK_ALEN,
                                                       "", gmx::StringFormatter("%x"));
#endif
                fprintf(debug, "block %d flags %s count %d\n",
                        b, flags.c_str(), c);
            }
        }
    }
    if (debug)
    {
        fprintf(debug, "Number of %d atom blocks to reduce: %d\n",
                reduction_block_size, bt->nblock_used);
        fprintf(debug, "Reduction density %.2f for touched blocks only %.2f\n",
                ctot*reduction_block_size/(double)numAtoms,
                ctot/(double)bt->nblock_used);
    }
}

void tear_down_bonded_threading(bonded_threading_t *bt)
{
    for (int th = 0; th < bt->nthreads; th++)
    {
        sfree(bt->f_t[th].fshift);
        for (int i = 0; i < egNR; i++)
        {
            sfree(bt->f_t[th].grpp.ener[i]);
        }
    }
    sfree(bt->f_t);
    sfree(bt->il_thread_division);
    sfree(bt);
}

void init_bonded_threading(FILE *fplog, int nenergrp,
                           struct bonded_threading_t **bt_ptr)
{
    bonded_threading_t *bt;

    snew(bt, 1);

    /* These thread local data structures are used for bondeds only.
     *
     * Note that we also use there structures when running single-threaded.
     * This is because the bonded force buffer uses type rvec4, whereas
     * the normal force buffer is uses type rvec. This leads to a little
     * reduction overhead, but the speed gain in the bonded calculations
     * of doing transposeScatterIncr/DecrU with aligment 4 instead of 3
     * is much larger than the reduction overhead.
     */
    bt->nthreads = gmx_omp_nthreads_get(emntBonded);

    snew(bt->f_t, bt->nthreads);
#pragma omp parallel for num_threads(bt->nthreads) schedule(static)
    for (int t = 0; t < bt->nthreads; t++)
    {
        try
        {
            /* Note that thread 0 uses the global fshift and energy arrays,
             * but to keep the code simple, we initialize all data here.
             */
            bt->f_t[t].f        = nullptr;
            bt->f_t[t].f_nalloc = 0;
            snew(bt->f_t[t].fshift, SHIFTS);
            bt->f_t[t].grpp.nener = nenergrp*nenergrp;
            for (int i = 0; i < egNR; i++)
            {
                snew(bt->f_t[t].grpp.ener[i], bt->f_t[t].grpp.nener);
            }
        }
        GMX_CATCH_ALL_AND_EXIT_WITH_FATAL_ERROR;
    }

    bt->nblock_used  = 0;
    bt->block_index  = nullptr;
    bt->mask         = nullptr;
    bt->block_nalloc = 0;

    /* The optimal value after which to switch from uniform to localized
     * bonded interaction distribution is 3, 4 or 5 depending on the system
     * and hardware.
     */
    const int max_nthread_uniform = 4;
    char *    ptr;

    if ((ptr = getenv("GMX_BONDED_NTHREAD_UNIFORM")) != nullptr)
    {
        sscanf(ptr, "%d", &bt->max_nthread_uniform);
        if (fplog != nullptr)
        {
            fprintf(fplog, "\nMax threads for uniform bonded distribution set to %d by env.var.\n",
                    bt->max_nthread_uniform);
        }
    }
    else
    {
        bt->max_nthread_uniform = max_nthread_uniform;
    }

    *bt_ptr = bt;
}