/****************************************************************************
**  CUBE        http://www.scalasca.org/                                   **
*****************************************************************************
**  Copyright (c) 1998-2016                                                **
**  Forschungszentrum Juelich GmbH, Juelich Supercomputing Centre          **
**                                                                         **
**  Copyright (c) 2009-2015                                                **
**  German Research School for Simulation Sciences GmbH,                   **
**  Laboratory for Parallel Programming                                    **
**                                                                         **
**  This software may be modified and distributed under the terms of       **
**  a BSD-style license.  See the COPYING file in the package base         **
**  directory for details.                                                 **
****************************************************************************/


/**
 * \file cubew_example.c
   \brief A small example of using "libcubew4.a".
   One creates a cube file "example.cube", defined structure of metrics, call tree, machine, cartesian topology
   and fills teh severity matrix.
   At the end the file will be written on the disk.
 */

#include <stdio.h>
#include <stdint.h>
#include <stdlib.h>
#include <unistd.h>
#include <errno.h>
#include <string.h>
#include <math.h>

#include "cubew_cube.h"



#define CNODES_NUMBER 10
#ifndef NWRITERS
#define NWRITERS 1
#endif

#ifndef NTHREADS
#define NTHREADS 64
#endif


#ifndef CNODES
#define CNODES 300
#endif


#ifndef COMPRESSION
#define COMPRESSION CUBE_FALSE
#endif

#define NMETRICS 40

static int current_random = 0; // uswd to create a stable random loooking sequence of numbers (used for creating of a calltree)
static int MAXDEPTH       = 3;


#define NDIMS 2


void
create_cube_structure( cube_t*,
                       cube_metric** metrics,
                       cube_cnode**  cnodes,
                       unsigned      Nthreads );

void
fill_cube_type_double_row( cube_type_double* row,
                           unsigned          size,
                           double            value );
void
fill_cube_type_int64_row( cube_type_int64* row,
                          unsigned         size,
                          int64_t          value );

void
create_calltree( cube_t*       cube,
                 cube_region** regn,
                 cube_cnode**  codes,
                 int*          idx,
                 int*          reg_idx,
                 cube_cnode*   parent,
                 int           level );
int
get_next_random();
void
set_seed_random( int s );
void
calculate_max_depth();

// / Start point of this example.
int
main( int argc, char* argv[] )
{
    calculate_max_depth();

    // / actual number of writers might be different. For code it is transparent.
    unsigned ArraySize = CNODES;    // / In this pattern, number of cnodes

    unsigned Nthreads = NTHREADS;   // / In this pattern, number of threads


    // / Pattern has to know, which metrics are there to perform parallel writing.
    // / To destroy is only this aray, not metrics inside. Metrics are destroyed by "cube_free"
    cube_metric** metrics_vector = ( cube_metric** )calloc( NMETRICS, sizeof( cube_metric* ) );


    // / Pattern has to know, which cnodes are there to perform parallel writing.
    // / To destroy is only this aray, not cnodes inside. Cnodes are destroyed by "cube_free"
    cube_cnode** cnodes_vector = ( cube_cnode** )calloc( ArraySize, sizeof( cube_cnode* ) );




    char* cubefile = "example3";


    //    FILE* fp = fopen(cubefile, "w");
    cube_t* cube = cube_create( cubefile, CUBE_MASTER, CUBE_FALSE );
    if ( !cube )
    {
        fprintf( stderr, "cube_create failed!\n" );
        exit( 1 );
    }

    create_cube_structure( cube, metrics_vector, cnodes_vector, Nthreads );


    cube_type_double sev_double[ Nthreads ];
    cube_type_int64  sev_int64[ Nthreads ];



    /* generate header */
    cube_def_mirror( cube, "http://icl.cs.utk.edu/software/kojak/" );
    cube_def_mirror( cube, "http://www.fz-juelich.de/jsc/kojak/" );
    cube_def_attr( cube, "description", "a simple example" );
    cube_def_attr( cube, "experiment time", "November 1st, 2004" );

    cube_enable_flat_tree( cube, CUBE_FALSE );


    char* test1 = "Humpty Dumpty sat on a wall,"
                  "Humpty Dumpty had a great fall."
                  "Four-score Men and Four-score more,"
                  "Could not make Humpty Dumpty where he was before.";
    cube_write_misc_data( cube, "Humpty Dumpty", test1, strlen( test1 ) );

    unsigned mi = 0;
    for ( mi = 0; mi < NMETRICS; mi++ ) // / run over the metrics
    {
        carray* sequence = cube_get_cnodes_for_metric( cube, metrics_vector[ mi ] );

        unsigned ci = 0;
        for ( ci = 0; ci < sequence->size; ci++ )
        {
            if ( metrics_vector[ mi ]->dtype_params->type == CUBE_DATA_TYPE_DOUBLE )
            {
                fill_cube_type_double_row( sev_double, Nthreads, ci + mi );
                cube_write_sev_row_of_cube_type_double( cube, metrics_vector[ mi ], ( cube_cnode* )( sequence->data )[ ci ], sev_double );
            }
            if ( metrics_vector[ mi ]->dtype_params->type == CUBE_DATA_TYPE_INT64 )
            {
                fill_cube_type_int64_row( sev_int64, Nthreads, ci + mi );
                cube_write_sev_row_of_cube_type_int64( cube, metrics_vector[ mi ], ( cube_cnode* )( sequence->data )[ ci ], sev_int64 );
            }
        }
    }

    printf( "Test file %s complete.\n", cubefile );
    char* test2 = "Twas bryllyg, and ye slythy toves"
                  "Did gyre and gymble in ye wabe:"
                  "All mimsy were ye borogoves;"
                  "And ye mome raths outgrabe.";

    cube_write_misc_data( cube, "Jabberwocky", test2, strlen( test2 ) );

    free( metrics_vector );               /* content shouldn't be freed, it is done internally by cube */
    free( cnodes_vector );                /* content shouldn't be freed, it is done internally by cube */

    cube_free( cube );
    return 0;
}






void
fill_cube_type_double_row( cube_type_double* row, unsigned size, double value )
{
    unsigned i = 0;
    for ( i = 0; i < size; i++ )
    {
        row[ i ].value = value + i;
    }
}

void
fill_cube_type_int64_row( cube_type_int64* row, unsigned size, int64_t value )
{
    unsigned i = 0;
    for ( i = 0; i < size; i++ )
    {
        row[ i ].value = value + i;
    }
}



void
create_cube_structure( cube_t* cube, cube_metric** metrics, cube_cnode** cnodes, unsigned Nthreads )
{
    /* generate header */
    cube_def_mirror( cube, "http://icl.cs.utk.edu/software/kojak/" );
    cube_def_mirror( cube, "http://www.fz-juelich.de/jsc/kojak/" );
    cube_def_attr( cube, "description", "a simple example" );
    cube_def_attr( cube, "experiment time", "November 1st, 2004" );

    /* define metric tree */
    metrics[ 0 ] = cube_def_met( cube, "Time", "time", "FLOAT", "sec", "",
                                 "@mirror@patterns-2.1.html#execution",
                                 "Total CPU allocation time", NULL, CUBE_METRIC_EXCLUSIVE );
    metrics[ 1 ] = cube_def_met( cube, "Visits", "visits", "INTEGER", "occ", "",
                                 "http://www.cs.utk.edu/usr.html",
                                 "Number of visits", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 2 ] = cube_def_met( cube, "Wait at Barrier", "mpi_barrier_wait", "FLOAT", "sec", "",
                                 "http://www.cs.utk.edu/sys.html",
                                 "Waiting time in front of MPI barriers", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 3 ] = cube_def_met( cube, "Barrier Completion", "mpi_barrier_completion", "FLOAT", "sec", "",
                                 "http://www.cs.utk.edu/sys.html",
                                 "Time needed to finish an MPI barrier", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 4 ] = cube_def_met( cube, "Late Sender", "mpi_latesender", "FLOAT", "sec", "",
                                 "http://www.cs.utk.edu/sys.html",
                                 "Time a receiving process is waiting for a message", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 5 ] = cube_def_met( cube, "Messages in Wrong Order", "mpi_latesender_wo", "FLOAT", "sec", "",
                                 "http://www.cs.utk.edu/sys.html",
                                 "Late Sender situation due to messages received in the wrong order", metrics[ 4 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 6 ] = cube_def_met( cube, "Messages from different sources ", "mpi_lswo_defferent", "FLOAT", "sec", "",
                                 "http://www.cs.utk.edu/sys.html",
                                 "Wrong order situation due to messages received from different sources", metrics[ 5 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 7 ] = cube_def_met( cube, "Messages from same sources ", "mpi_lswo_same", "FLOAT", "sec", "",
                                 "http://www.cs.utk.edu/sys.html",
                                 "Wrong order situation due to messages received from same sources", metrics[ 5 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 8 ] = cube_def_met( cube, "Late Receiver", "mpi_latereceiver", "FLOAT", "sec", "",
                                 "http://www.cs.utk.edu/sys.html",
                                 "Time a sending process is waiting for the receiver to become ready", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 9 ] = cube_def_met( cube, "Early Reduce", "mpi_earlyreduce", "FLOAT", "sec", "",
                                 "http://www.cs.utk.edu/sys.html",
                                 "Waiting time due to an early receiver in MPI n-to-1 operation", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 10 ] = cube_def_met( cube, "Early Scan", "mpi_earlyscan", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Waiting time due to an early receiver in MPI scan operation", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 11 ] = cube_def_met( cube, "Late Broadcast", "mpi_latebroadcast", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Waiting time due to an early rreceiver in MPI 1-to-n operation", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 12 ] = cube_def_met( cube, "Wait at N x N", "mpi_wait_nxn", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Time due to inherent synchronization in MPI n-to-n operation", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 13 ] = cube_def_met( cube, "N x N completion", "mpi_nxn_completion", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Time needed to finish a n-to-n collective operation", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 14 ] = cube_def_met( cube, "P2P send synchronizations", "sync_send", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of point-to-point send synchronization operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 15 ] = cube_def_met( cube, "Late Receivers", "mpi_slr_count", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of Late Receivers instances in synchronizations", metrics[ 14 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 16 ] = cube_def_met( cube, "P2P recv synchronizations", "sync_recv", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of point-to-point receive synchronization operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 17 ] = cube_def_met( cube, "Late Senders", "mpi_sls_count", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of Late Sender instances in synchronizations", metrics[ 16 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 18 ] = cube_def_met( cube, "Messages in Wrong Order", "mpi_slswo_count", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of Late Sender instances in synchronizations were messages are receivesd in wrong order", metrics[ 17 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 19 ] = cube_def_met( cube, "Collective synchronizations", "sync_coll", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of collective synchronization oprations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 20 ] = cube_def_met( cube, "P2P send communications", "comms_send", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of point-to-point send communication operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 21 ] = cube_def_met( cube, "Late Receivers", "mpi_clr_count", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of Late Receiver instances in communications", metrics[ 20 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 22 ] = cube_def_met( cube, "P2P recv communications", "comms_recv", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of point-to-point receive communication operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 23 ] = cube_def_met( cube, "Late Sender", "mpi_cls_count", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of Late Sender instances in communications", metrics[ 22 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 24 ] = cube_def_met( cube, "Messages in Wrong Order", "mpi_clswo_count", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of Late Sender instances in communications were messages are received in wrong order", metrics[ 23 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 25 ] = cube_def_met( cube, "Collective exchange communications", "comms_cxch", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of collective communications as source and destination", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 26 ] = cube_def_met( cube, "Collective communications as source", "comms_csrc", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of collective communications as source", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 27 ] = cube_def_met( cube, "Collective communications as destination", "comms_cdst", "INTEGER", "occ", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of collective communications as destination", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 28 ] = cube_def_met( cube, "P2P bytes sent", "bytes_sent", "INTEGER", "bytes", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of bytes sent in point-to-point operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 29 ] = cube_def_met( cube, "P2P bytes received", "bytes_rcvd", "INTEGER", "bytes", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of bytes received in point-to-point operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 30 ] = cube_def_met( cube, "Collective bytes outgoing", "bytes_cout", "INTEGER", "bytes", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of bytes outgoing in collective operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 31 ] = cube_def_met( cube, "Collective bytes incoming", "bytes_cin", "INTEGER", "bytes", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of bytes incoming in collective operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 32 ] = cube_def_met( cube, "RMA bytes received", "bytes_get", "INTEGER", "bytes", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of bytes received in RMA operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 33 ] = cube_def_met( cube, "RMA bytes put", "bytes_put", "INTEGER", "bytes", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Number of bytes sent in RMA operations", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 34 ] = cube_def_met( cube, "Late Post", "mpi_rma_late_post", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Time spent in MPI RMA LAte Post inefficiency pattern", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 35 ] = cube_def_met( cube, "Early Wait", "mpi_rma_early_wait", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Time spent in MPI_Win_Wait waiting for last MPI_Win_complete.", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 36 ] = cube_def_met( cube, "Late Complet", "mpi_rma_late_complete", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Time spent in MPI_Win_Wait waiting between last rma access and last MPI_Win_complete.", metrics[ 35 ],  CUBE_METRIC_EXCLUSIVE );
    metrics[ 37 ] = cube_def_met( cube, "Early Transfer", "mpi_rma_early_transfer", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Time spent in MPI RMA Early Tansfer inefficiency pattern.", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 38 ] = cube_def_met( cube, "Wait at Fence", "mpi_rma_wait_at_fence", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Time spent in MPI_Win_fence, waiting for other processes", NULL,  CUBE_METRIC_EXCLUSIVE );
    metrics[ 39 ] = cube_def_met( cube, "Early Fence", "mpi_rma_early_fence", "FLOAT", "sec", "",
                                  "http://www.cs.utk.edu/sys.html",
                                  "Time spent in MPI_Win_fence while  waiting for pending PMA ops", metrics[ 38 ],  CUBE_METRIC_EXCLUSIVE );





    /* define call tree */
    char* mod = "/ICL/CUBE/example.c";
//      cube_region *regn0, *regn1, *regn2;

    cube_region** regn = ( cube_region** )calloc( 169, sizeof( cube_region* ) );


    char* names[ 169 ];

    names[ 0 ]   = "UNKNOWN";
    names[ 1 ]   = "TRACING";
    names[ 2 ]   = "MPI_Accumulate";
    names[ 3 ]   =  "MPI_Allgather";
    names[ 4 ]   =  "MPI_Allgatherv";
    names[ 5 ]   =  "MPI_Allreduce";
    names[ 6 ]   =  "MPI_Alltoall";
    names[ 7 ]   =   "MPI_Alltoallv";
    names[ 8 ]   =   "MPI_Alltoallw";
    names[ 9 ]   =   "MPI_Barrier";
    names[ 10 ]  =   "MPI_Bcast";
    names[ 11 ]  =   "MPI_Bsend";
    names[ 12 ]  =  "MPI_Bsend_init";
    names[ 13 ]  =   "MPI_Cancel";
    names[ 14 ]  =   "MPI_Cart_create";
    names[ 15 ]  =   "MPI_Comm_create";
    names[ 16 ]  =   "MPI_Comm_dup";
    names[ 17 ]  =   "MPI_Comm_free";
    names[ 18 ]  =   "MPI_Comm_group";
    names[ 19 ]  =    "MPI_Comm_remote_group";
    names[ 20 ]  =    "MPI_Comm_split";
    names[ 21 ]  =    "MPI_Exscan";
    names[ 22 ]  =    "MPI_File_close";
    names[ 23 ]  =    "MPI_File_delete";
    names[ 24 ]  =    "MPI_File_get_amode";
    names[ 25 ]  =    "MPI_File_get_atomicity";
    names[ 26 ]  =    "MPI_File_get_byte_offset";
    names[ 27 ]  =     "MPI_File_get_group";
    names[ 28 ]  =     "MPI_File_get_info";
    names[ 29 ]  =     "MPI_File_get_position";
    names[ 30 ]  =     "MPI_File_get_position_shared";
    names[ 31 ]  =    "MPI_File_get_size";
    names[ 32 ]  =    "MPI_File_get_type_extent";
    names[ 33 ]  =    "MPI_File_get_view";
    names[ 34 ]  =    "MPI_File_iread";
    names[ 35 ]  =   "MPI_File_iread_at";
    names[ 36 ]  =   "MPI_File_iread_shared";
    names[ 37 ]  =   "MPI_File_iwrite";
    names[ 38 ]  =    "MPI_File_iwrite_at";
    names[ 39 ]  =    "MPI_File_iwrite_shared";
    names[ 40 ]  =    "MPI_File_open";
    names[ 41 ]  =    "MPI_File_preallocated";
    names[ 42 ]  =    "MPI_File_read";
    names[ 43 ]  =    "MPI_File_read_all";
    names[ 44 ]  =    "MPI_File_read_all_begin";
    names[ 45 ]  =    "MPI_File_real_all_end";
    names[ 46 ]  =     "MPI_File_read_at";
    names[ 47 ]  =     "MPI_File_read_at_all";
    names[ 48 ]  =     "MPI_File_read_at_all_begin";
    names[ 49 ]  =     "MPI_File_read_at_all_end";
    names[ 50 ]  =    "MPI_File_read_ordered";
    names[ 51 ]  =    "MPI_File_read_ordered_begin";
    names[ 52 ]  =   "MPI_File_read_ordered_end";
    names[ 53 ]  =   "MPI_File_read_shared";
    names[ 54 ]  =   "MPI_File_seek";
    names[ 55 ]  =   "MPI_File_seek_shared";
    names[ 56 ]  =    "MPI_File_set_atomicity";
    names[ 57 ]  =    "MPI_File_set_info";
    names[ 58 ]  =    "MPI_File_set_size";
    names[ 59 ]  =    "MPI_File_set_view";
    names[ 60 ]  =    "MPI_File_sync";
    names[ 61 ]  =    "MPI_file_write";
    names[ 62 ]  =    "MPI_File_write_all";
    names[ 63 ]  =    "MPI_File_write_all_begin";
    names[ 64 ]  =     "MPI_File_write_all_end";
    names[ 65 ]  =     "MPI_File_write_at_all";
    names[ 66 ]  =     "MPI_File_write_at_all_begin";
    names[ 67 ]  =     "MPI_File_write_at_all_end";
    names[ 68 ]  =     "MPI_File_write_ordered";
    names[ 69 ]  =     "MPI_File_write_ordered_begin";
    names[ 70 ]  =    "MPI_File_write_ordered_end";
    names[ 71 ]  =    "MPI_File_write_shared";
    names[ 72 ]  =    "MPI_Finalize";
    names[ 73 ]  =    "MPI_Gather";
    names[ 74 ]  =   "mPI_Gatherv";
    names[ 75 ]  =   "MPI_Get";
    names[ 76 ]  =    "MPI_Graph_create";
    names[ 77 ]  =    "MPI_Group_difference";
    names[ 78 ]  =    "MPI_Group_excl";
    names[ 79 ]  =    "MPI_Group_ free";
    names[ 80 ]  =     "MPI_Group_incl";
    names[ 81 ]  =     "MPI_Group_intersection";
    names[ 82 ]  =     "MPI_Group_range_excl";
    names[ 83 ]  =     "MPI_Group_range_incl";
    names[ 84 ]  =     "MPI_Group_union";
    names[ 85 ]  =     "MPI_Ibsend";
    names[ 86 ]  =     "MPI_Init";
    names[ 87 ]  =     "MPI_Init_thread";
    names[ 88 ]  =     "MPI_Intercomm_create";
    names[ 89 ]  =     "MPI_Intercomm_merge";
    names[ 90 ]  =     "MPI_Irecv";
    names[ 91 ]  =     "MPI_Irsend";
    names[ 92 ]  =     "MPI_Isend";
    names[ 93 ]  =     "MPI_Issend";
    names[ 94 ]  =     "MPI_Probe";
    names[ 95 ]  =     "MPI_Put";
    names[ 96 ]  =     "MPI_Recv";
    names[ 97 ]  =     "MPI_Recv_init";
    names[ 98 ]  =     "MPI_Reduce";
    names[ 99 ]  =     "MPI_Reduce_scatter";
    names[ 100 ] =     "MPI_Request_free";
    names[ 101 ] =     "MPI_Rsend";
    names[ 102 ] =     "MPI_Rsend_init";
    names[ 103 ] =     "MPI_Scan";
    names[ 104 ] =     "MPI_Scatter";
    names[ 105 ] =     "MPI_Scatterv";
    names[ 106 ] =      "MPI_Send";
    names[ 107 ] =      "MPI_Send_init";
    names[ 108 ] =     "MPI_Sendrecv";
    names[ 109 ] =      "MPI_Sendrecv_replace";
    names[ 110 ] =      "MPI_Ssend";
    names[ 111 ] =      "MPI_Ssend_init";
    names[ 112 ] =      "MPI_Start";
    names[ 113 ] =      "MPI_Startall";
    names[ 114 ] =      "MPI_Test";
    names[ 115 ] =      "MPI_Testall";
    names[ 116 ] =      "MPI_Testany";
    names[ 117 ] =      "MPI_Testsome";
    names[ 118 ] =      "MPI_Wait";
    names[ 119 ] =      "MPI_Waitall";
    names[ 120 ] =      "MPI_Waitany";
    names[ 121 ] =      "MPI_Waitsome";
    names[ 122 ] =      "MPI_Win_complete";
    names[ 123 ] =      "MPI_Win_create";
    names[ 124 ] =      "MPI_Win_fence";
    names[ 125 ] =     "MPI_Win_free";
    names[ 126 ] =      "MPI_Win_lock";
    names[ 127 ] =      "MPI_Win_post";
    names[ 128 ] =     "MPI_Win_start";
    names[ 129 ] =     "MPI_Win_test";
    names[ 130 ] =     "MPI_Win_unlock";
    names[ 131 ] =     "MPI_Win_wait";
    names[ 132 ] =     "PARALLEL";
    names[ 133 ] =     "driver";
    names[ 134 ] =     "task_init";
    names[ 135 ] =     "bcast_init";
    names[ 136 ] =     "barriere_sync";
    names[ 137 ] =    "read_input";
    names[ 138 ] =     "bcast_real";
    names[ 139 ] =     "decomp";
    names[ 140 ] =     "inner_auto";
    names[ 141 ] =     "inner";
    names[ 142 ] =     "initialize";
    names[ 143 ] =     "initxs";
    names[ 144 ] =     "initsnc";
    names[ 145 ] =     "octant";
    names[ 146 ] =     "initgeom";
    names[ 147 ] =     "timers";
    names[ 148 ] =     "its=1                    ";
    names[ 149 ] =     "source";
    names[ 150 ] =     "sweep";
    names[ 151 ] =      "rcv_real";
    names[ 152 ] =     "snd_real";
    names[ 153 ] =     "global_init_sum";
    names[ 154 ] =     "flux_err";
    names[ 155 ] =      "global_real_max";
    names[ 156 ] =      "its=2                        ";
    names[ 157 ] =      "its=3                        ";
    names[ 158 ] =      "its=4                        ";
    names[ 159 ] =      "its=5                        ";
    names[ 160 ] =     "its=6                        ";
    names[ 161 ] =     "its=7                        ";
    names[ 162 ] =     "its=8                        ";
    names[ 163 ] =   "its=9                        ";
    names[ 164 ] =    "its=10                        ";
    names[ 165 ] =    "its=11                       ";
    names[ 166 ] =      "its=12                        ";
    names[ 167 ] =     "global_real_sum";
    names[ 168 ] =     "task_end";



    for ( unsigned i = 0; i < 169; i++ )
    {
        char* descr = "";
        if ( i > 1 && i < 132  )
        {
            descr = "MPI";
        }
        if ( i == 1 || i == 132 )
        {
            descr = "EPIK";
        }
        if ( i > 132 )
        {
            descr = "USR";
        }
        regn[ i ] = cube_def_region( cube, names[ i ], names[ i ], "MPI", "barrier",  -1, -1, "", descr, mod );
    }

    //     create calltree
    srand( 0 );
    int idx        = 0;
    int region_idx = 0;
    create_calltree( cube, regn, cnodes, &idx, &region_idx,  NULL, 0 );


    /* define location tree */
    cube_machine* mach = cube_def_mach( cube, "MSC", "" );
    cube_node*    node = cube_def_node( cube, "Athena", mach );

    cube_thread** thrds = ( cube_thread** )calloc( Nthreads, sizeof( cube_thread* ) );

    for ( unsigned i = 0; i < Nthreads; i++ )
    {
        cube_process* proc0 = cube_def_proc( cube, "Process 0", i, node );
        thrds[ i ] = cube_def_thrd( cube, "Thread 0", i, proc0 );
    }


    free( regn );




    /* define first topology */
    long            dimv0[ NDIMS ]    = { 5, 5 };
    int             periodv0[ NDIMS ] = { 1, 0 };
    cube_cartesian* cart0             = cube_def_cart( cube, NDIMS, dimv0, periodv0 );
    cube_cart_set_name( cart0, "My Topology 1" );

    long coordv[ NDIMS ] = { 0, 0 };
    cube_def_coords( cube, cart0, thrds[ 1 ], coordv );

    /* define second topology */
    long            dimv1[ NDIMS ]    = { 3, 3 };
    int             periodv1[ NDIMS ] = { 1, 0 };
    cube_cartesian* cart1             = cube_def_cart( cube, NDIMS, dimv1, periodv1 );
    cube_cart_set_name( cart1, "Application topology 3" );

    long coordv0[ NDIMS ] = { 0, 1 };
    long coordv1[ NDIMS ] = { 1, 0 };
    cube_def_coords( cube, cart1, thrds[ 0 ], coordv0 );
    cube_def_coords( cube, cart1, thrds[ 1 ], coordv1 );





    /* define second topology*/
    long            dimv2[ 4 ]    = { 3, 3, 3, 3 };
    int             periodv2[ 4 ] = { 1, 0, 0, 0 };
    cube_cartesian* cart2         = cube_def_cart( cube, 4, dimv2, periodv2 );

    long coordv20[ 4 ] = { 0, 1, 0, 0 };
    long coordv21[ 4 ] = { 1, 0, 0, 0 };
    cube_def_coords( cube, cart2, thrds[ 2 ], coordv20 );
    cube_def_coords( cube, cart2, thrds[ 3 ], coordv21 );
    cube_cart_set_name( cart2, "Second" );


    /* define third topology*/
    long            dimv3[ 14 ]    = { 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3 };
    int             periodv3[ 14 ] = { 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
    cube_cartesian* cart3          = cube_def_cart( cube, 14, dimv3, periodv3 );
    long            coordv32[ 14 ] = { 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2 };
    long            coordv33[ 14 ] = { 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
    cube_def_coords( cube, cart3, thrds[ 5 ], coordv32 );
    cube_def_coords( cube, cart3, thrds[ 10 ], coordv33 );
    cube_cart_set_name( cart3, "Third" );






    cube_set_statistic_name( cube, "mystat" );

    cube_set_metrics_title( cube, "Metric tree (QMCD App <ver.34>)" );
    cube_set_calltree_title( cube, "Calltree (serial run, #2345)" );
    cube_set_systemtree_title( cube, "System  ( Linux cluster < 43 & 23 >)" );


    free( thrds );
}




int
get_next_random()
{
    current_random = ( current_random + 5 ) % 13;
    return current_random;
}

void
set_seed_random( int s )
{
    current_random = s;
}



void
create_calltree( cube_t* cube, cube_region** regn, cube_cnode** cnodes, int* idx, int* region_idx,  cube_cnode* parent, int level )
{
    if ( ( *idx ) >= CNODES )
    {
        return;
    }
    if ( level >= MAXDEPTH )
    {
        return;
    }
    int _num_children = get_next_random();
    for ( unsigned i = 0; i < _num_children; i++ )
    {
        unsigned rand_reg = *region_idx;
        *region_idx = ( ( *region_idx + 1 ) % 169 );      // cyclic selection of regions
        cube_cnode* cnode = cube_def_cnode( cube, regn[ rand_reg ], parent );
        cnodes[ *idx ] = cnode;
        ( *idx )++;
        create_calltree( cube, regn, cnodes, idx, region_idx, cnode, level + 1 );
        if ( ( *idx ) >= CNODES )
        {
            return;                            // for the case, inside of those function was reached the end of array
        }
    }
}





void
calculate_max_depth()
{
    MAXDEPTH = ( int )( log( CNODES ) / log( 13 ) ) + 2;
}