Ioss_ParallelUtils.h

Go to the documentation of this file.

// Copyright(C) 1999-2024 National Technology & Engineering Solutions
// of Sandia, LLC (NTESS).  Under the terms of Contract DE-NA0003525 with
// NTESS, the U.S. Government retains certain rights in this software.
//
// See packages/seacas/LICENSE for details
 
#pragma once
 
#include "Ioss_CodeTypes.h" // for Int64Vector, IntVector
#include "Ioss_Utils.h"
#include <cassert>
#include <cstddef> // for size_t
#include <stdint.h>
#include <string> // for string
#include <vector> // for vector
 
#include "ioss_export.h"
#if IOSS_DEBUG_OUTPUT
#include <fmt/format.h>
#include <fmt/ostream.h>
#endif
 
#ifdef SEACAS_HAVE_MPI
#include "Ioss_SerializeIO.h"
#endif
 
namespace Ioss {
  class PropertyManager;
 
  class IOSS_EXPORT ParallelUtils
  {
  public:
    ParallelUtils() = default;
    explicit ParallelUtils(Ioss_MPI_Comm the_communicator);
 
    enum MinMax { DO_MAX, DO_MIN, DO_SUM };
 
#if defined(SEACAS_HAVE_MPI)
    IOSS_NODISCARD static Ioss_MPI_Comm comm_world()
    {
      return (Ioss_MPI_Comm)MPI_COMM_WORLD; // CHECK: ALLOW MPI_COMM_WORLD
    }
    IOSS_NODISCARD static Ioss_MPI_Comm comm_self() { return (Ioss_MPI_Comm)MPI_COMM_SELF; }
    IOSS_NODISCARD static Ioss_MPI_Comm comm_null() { return (Ioss_MPI_Comm)MPI_COMM_NULL; }
#else
    // NOTE: These values match those used in siMPI package.
    IOSS_NODISCARD static constexpr Ioss_MPI_Comm comm_world() { return -100; }
    IOSS_NODISCARD static constexpr Ioss_MPI_Comm comm_self() { return -100; }
    IOSS_NODISCARD static constexpr Ioss_MPI_Comm comm_null() { return 0; }
#endif
 
    /*!
     * See if any external properties specified via the
     * IOSS_PROPERTIES environment variable.  If any found, add to
     * `properties`.
     */
    void add_environment_properties(Ioss::PropertyManager &properties);
 
    /*!
     * Returns 'true' if 'name' is defined in the environment.
     * The value of the environment variable is returned in 'value'.
     * getenv system call is only done on processor 0.
     * If '!sync_parallel', then don't push to other processors.
     */
    bool get_environment(const std::string &name, std::string &value,
                         IOSS_MAYBE_UNUSED bool sync_parallel) const;
 
    /*!
     * Returns 'true' if 'name' is defined in the environment.  The
     * value of the environment variable is converted to an integer
     * and returned in 'value'.  No checking is done to ensure that
     * the environment variable points to a valid integer.  getenv
     * system call is only done on processor 0.  If '!sync_parallel',
     * then don't push to other processors.
     */
    bool get_environment(const std::string &name, int &value,
                         IOSS_MAYBE_UNUSED bool sync_parallel) const;
 
    /*!
     * Returns 'true' if 'name' is defined in the environment no
     * matter what the value. Returns false otherwise.
     * getenv system call is only done on processor 0.
     * If '!sync_parallel', then don't push to other processors.
     */
    IOSS_NODISCARD bool get_environment(const std::string     &name,
                                        IOSS_MAYBE_UNUSED bool sync_parallel) const;
 
    IOSS_NODISCARD std::string decode_filename(const std::string &filename, bool is_parallel) const;
 
    IOSS_NODISCARD Ioss_MPI_Comm communicator() const { return communicator_; }
    IOSS_NODISCARD int           parallel_size() const;
    IOSS_NODISCARD int           parallel_rank() const;
 
    void barrier() const;
 
    /*!
     * Global OR of attribute strings, the processors which have no
     * knowledge of the value should initialize to '0' and the
     * processors with knowledge set the appropriate values.
     */
    void attribute_reduction(IOSS_MAYBE_UNUSED int length, IOSS_MAYBE_UNUSED char buffer[]) const;
 
    /*!
     * Generate a "globally unique id" which is unique over all entities
     * of a specific type over all processors.
     * Used by some applications for uniquely identifying an entity.
     * If `rank` == -1, then use parallel_rank; otherwise use rank
     */
    IOSS_NODISCARD int64_t generate_guid(IOSS_MAYBE_UNUSED size_t id, int rank = -1) const;
 
    /*! Return min, max, average memory used by any process */
    void memory_stats(int64_t &min, int64_t &max, int64_t &avg) const;
 
    /*! Return high-water-mark min, max, average memory used by any process */
    /* May be inaccurate unless system maintains this information */
    void hwm_memory_stats(int64_t &min, int64_t &max, int64_t &avg) const;
 
    /*! Vector 'local_counts' contains the number of objects
     * local to this processor.  On exit, global_counts
     * contains the total number of objects on all processors.
     * Assumes that ordering is the same on all processors
     */
    void global_count(const IntVector &local_counts, IntVector &global_counts) const;
    void global_count(const Int64Vector &local_counts, Int64Vector &global_counts) const;
 
    template <typename T>
    IOSS_NODISCARD T global_minmax(IOSS_MAYBE_UNUSED T      local_minmax,
                                   IOSS_MAYBE_UNUSED MinMax which) const;
 
    template <typename T>
    void global_array_minmax(IOSS_MAYBE_UNUSED std::vector<T> &local_minmax,
                             IOSS_MAYBE_UNUSED MinMax          which) const;
 
    template <typename T> void gather(T my_value, std::vector<T> &result) const;
    template <typename T> void all_gather(T my_value, std::vector<T> &result) const;
    template <typename T> void gather(std::vector<T> &my_values, std::vector<T> &result) const;
    template <typename T> void all_gather(std::vector<T> &my_values, std::vector<T> &result) const;
    template <typename T>
    int gather(int vals_count, int size_per_val, std::vector<T> &my_values,
               std::vector<T> &result) const;
 
    template <typename T> void broadcast(T &my_value, int root = 0) const;
    template <typename T> void broadcast(std::vector<T> &my_value, int root = 0) const;
 
    void progress(const std::string &output) const;
 
  private:
    Ioss_MPI_Comm communicator_{comm_world()};
    mutable int   parallelSize_{-1};
    mutable int   parallelRank_{-1};
  };
 
#ifdef SEACAS_HAVE_MPI
  IOSS_NODISCARD inline MPI_Datatype mpi_type(double /*dummy*/) { return MPI_DOUBLE; }
  IOSS_NODISCARD inline MPI_Datatype mpi_type(float /*dummy*/) { return MPI_FLOAT; }
  IOSS_NODISCARD inline MPI_Datatype mpi_type(int /*dummy*/) { return MPI_INT; }
  IOSS_NODISCARD inline MPI_Datatype mpi_type(long int /*dummy*/) { return MPI_LONG_LONG_INT; }
  IOSS_NODISCARD inline MPI_Datatype mpi_type(long long int /*dummy*/) { return MPI_LONG_LONG_INT; }
  IOSS_NODISCARD inline MPI_Datatype mpi_type(unsigned int /*dummy*/) { return MPI_UNSIGNED; }
  IOSS_NODISCARD inline MPI_Datatype mpi_type(unsigned long int /*dummy*/)
  {
    return MPI_UNSIGNED_LONG;
  }
  IOSS_NODISCARD inline MPI_Datatype mpi_type(unsigned long long int /*dummy*/)
  {
    return MPI_UNSIGNED_LONG_LONG;
  }
  IOSS_NODISCARD inline MPI_Datatype mpi_type(char /*dummy*/) { return MPI_CHAR; }
 
  template <typename T>
  int MY_Alltoallv64(const std::vector<T> &sendbuf, const std::vector<int64_t> &sendcounts,
                     const std::vector<int64_t> &senddisp, std::vector<T> &recvbuf,
                     const std::vector<int64_t> &recvcounts, const std::vector<int64_t> &recvdisp,
                     Ioss_MPI_Comm comm)
  {
    int processor_count = 0;
    int my_processor    = 0;
    MPI_Comm_size(comm, &processor_count);
    MPI_Comm_rank(comm, &my_processor);
 
    // Verify that all 'counts' can fit in an integer. Symmetric
    // communication, so recvcounts are sendcounts on another processor.
    for (int i = 0; i < processor_count; i++) {
      int snd_cnt = static_cast<int>(sendcounts[i]);
      if (static_cast<int64_t>(snd_cnt) != sendcounts[i]) {
        std::ostringstream errmsg;
        errmsg << "ERROR: The number of items that must be communicated via MPI calls from\n"
               << "       processor " << my_processor << " to processor " << i << " is "
               << sendcounts[i]
               << "\n       which exceeds the storage capacity of the integers "
                  "used by MPI functions.\n";
        IOSS_ERROR(errmsg);
      }
    }
 
    size_t pow_2 = Ioss::Utils::power_2(processor_count);
 
    for (size_t i = 1; i < pow_2; i++) {
      MPI_Status status{};
 
      size_t exchange_proc = i ^ my_processor;
      if (exchange_proc < static_cast<size_t>(processor_count)) {
        int snd_cnt = static_cast<int>(
            sendcounts[exchange_proc]); // Converts from int64_t to int as needed by mpi
        int rcv_cnt = static_cast<int>(recvcounts[exchange_proc]);
        int tag     = 24713;
        if (static_cast<size_t>(my_processor) < exchange_proc) {
          MPI_Send((void *)&sendbuf[senddisp[exchange_proc]], snd_cnt, mpi_type(T(0)),
                   exchange_proc, tag, comm);
          MPI_Recv(&recvbuf[recvdisp[exchange_proc]], rcv_cnt, mpi_type(T(0)), exchange_proc, tag,
                   comm, &status);
        }
        else {
          MPI_Recv(&recvbuf[recvdisp[exchange_proc]], rcv_cnt, mpi_type(T(0)), exchange_proc, tag,
                   comm, &status);
          MPI_Send((void *)&sendbuf[senddisp[exchange_proc]], snd_cnt, mpi_type(T(0)),
                   exchange_proc, tag, comm);
        }
      }
    }
 
    // Take care of this processor's data movement...
    std::copy(&sendbuf[senddisp[my_processor]],
              &sendbuf[senddisp[my_processor] + sendcounts[my_processor]],
              &recvbuf[recvdisp[my_processor]]);
    return 0;
  }
 
  template <typename T>
  int MY_Alltoallv(const std::vector<T> &sendbuf, const std::vector<int64_t> &sendcnts,
                   const std::vector<int64_t> &senddisp, std::vector<T> &recvbuf,
                   const std::vector<int64_t> &recvcnts, const std::vector<int64_t> &recvdisp,
                   Ioss_MPI_Comm comm)
  {
// Wrapper to handle case where send/recv counts and displacements are 64-bit integers.
// Two cases:
// 1) They are of type 64-bit integers, but only storing data in the 32-bit integer range.
//    -- if (sendcnts[#proc-1] + senddisp[#proc-1] < 2^31, then we are ok
// 2) They are of type 64-bit integers, and storing data in the 64-bit integer range.
//    -- call special alltoallv which does point-to-point sends
#if IOSS_DEBUG_OUTPUT
    {
      Ioss::ParallelUtils utils(comm);
      int                 processor_count = utils.parallel_size();
 
      int              max_comm = sendcnts[processor_count - 1] + senddisp[processor_count - 1];
      std::vector<int> comm_size;
 
      utils.gather(max_comm, comm_size);
      int my_rank = utils.parallel_rank();
      if (my_rank == 0) {
        fmt::print("Send Communication Size: {}\n", fmt::join(comm_size, ", "));
      }
    }
#endif
#if 1
    int processor_count = 0;
    MPI_Comm_size(comm, &processor_count);
    size_t max_comm = sendcnts[processor_count - 1] + senddisp[processor_count - 1];
    size_t one      = 1;
    if (max_comm < one << 31) {
      // count and displacement data in range, need to copy to integer vector.
      std::vector<int> send_cnt(sendcnts.begin(), sendcnts.end());
      std::vector<int> send_dis(senddisp.begin(), senddisp.end());
      std::vector<int> recv_cnt(recvcnts.begin(), recvcnts.end());
      std::vector<int> recv_dis(recvdisp.begin(), recvdisp.end());
      return MPI_Alltoallv((void *)Data(sendbuf), Data(send_cnt), Data(send_dis), mpi_type(T(0)),
                           (void *)Data(recvbuf), Data(recv_cnt), Data(recv_dis), mpi_type(T(0)),
                           comm);
    }
    else {
#endif
      // Same as if each processor sent a message to every other process with:
      //     MPI_Send(sendbuf+senddisp[i]*sizeof(sendtype),sendcnts[i], sendtype, i, tag, comm);
      // And received a message from each processor with a call to:
      //     MPI_Recv(recvbuf+recvdisp[i]*sizeof(recvtype),recvcnts[i], recvtype, i, tag, comm);
      return MY_Alltoallv64(sendbuf, sendcnts, senddisp, recvbuf, recvcnts, recvdisp, comm);
#if 1
    }
#endif
  }
 
  template <typename T>
  int MY_Alltoallv(const std::vector<T> &sendbuf, const std::vector<int> &sendcnts,
                   const std::vector<int> &senddisp, std::vector<T> &recvbuf,
                   const std::vector<int> &recvcnts, const std::vector<int> &recvdisp,
                   Ioss_MPI_Comm comm)
  {
#if IOSS_DEBUG_OUTPUT
    {
      Ioss::ParallelUtils utils(comm);
      int                 processor_count = utils.parallel_size();
 
      int              max_comm = sendcnts[processor_count - 1] + senddisp[processor_count - 1];
      std::vector<int> comm_size;
 
      utils.gather(max_comm, comm_size);
      int my_rank = utils.parallel_rank();
      if (my_rank == 0) {
        fmt::print("Send Communication Size: {}\n", fmt::join(comm_size, ", "));
      }
    }
#endif
    return MPI_Alltoallv((void *)Data(sendbuf), const_cast<int *>(Data(sendcnts)),
                         const_cast<int *>(Data(senddisp)), mpi_type(T(0)), Data(recvbuf),
                         const_cast<int *>(Data(recvcnts)), const_cast<int *>(Data(recvdisp)),
                         mpi_type(T(0)), comm);
  }
#endif
 
  template <typename T>
  void ParallelUtils::global_array_minmax(IOSS_MAYBE_UNUSED std::vector<T> &local_minmax,
                                          IOSS_MAYBE_UNUSED MinMax          which) const
  {
    IOSS_PAR_UNUSED(local_minmax);
    IOSS_PAR_UNUSED(which);
#ifdef SEACAS_HAVE_MPI
    if (parallel_size() > 1 && !local_minmax.empty()) {
      if (Ioss::SerializeIO::isEnabled() && Ioss::SerializeIO::inBarrier()) {
        std::ostringstream errmsg;
        errmsg << "Attempting mpi while in barrier owned by " << Ioss::SerializeIO::getOwner();
        IOSS_ERROR(errmsg);
      }
 
      std::vector<T> maxout(local_minmax.size());
      MPI_Op         oper = MPI_MAX;
      if (which == Ioss::ParallelUtils::DO_MAX) {
        oper = MPI_MAX;
      }
      else if (which == Ioss::ParallelUtils::DO_MIN) {
        oper = MPI_MIN;
      }
      else if (which == Ioss::ParallelUtils::DO_SUM) {
        oper = MPI_SUM;
      }
 
      const int success =
          MPI_Allreduce((void *)(Data(local_minmax)), Data(maxout),
                        static_cast<int>(local_minmax.size()), mpi_type(T()), oper, communicator_);
      if (success != MPI_SUCCESS) {
        std::ostringstream errmsg;
        errmsg << "Ioss::ParallelUtils::global_array_minmax - MPI_Allreduce failed";
        IOSS_ERROR(errmsg);
      }
      // Now copy back into passed in array...
      for (size_t i = 0; i < local_minmax.size(); i++) {
        local_minmax[i] = maxout[i];
      }
    }
#endif
  }
 
} // namespace Ioss