VCluster_meta_function.hpp

/*
 * VCluster_meta_function.hpp
 *
 *  Created on: Dec 8, 2016
 *      Author: i-bird
 */
 
#ifndef OPENFPM_VCLUSTER_SRC_VCLUSTER_VCLUSTER_META_FUNCTION_HPP_
#define OPENFPM_VCLUSTER_SRC_VCLUSTER_VCLUSTER_META_FUNCTION_HPP_
 
#include "memory/BHeapMemory.hpp"
#include "Packer_Unpacker/has_max_prop.hpp"
 
static inline bool is_mpi_rdma_cuda_active()
{
#if defined(MPIX_CUDA_AWARE_SUPPORT) && MPIX_CUDA_AWARE_SUPPORT
            return true;
#else
            return false;
#endif
}
 
template<bool result, typename T, typename S, template<typename> class layout_base, typename Memory>
struct unpack_selector_with_prp
{
    template<typename op,
             int ... prp>
    static void call_unpack(S & recv,
                            openfpm::vector_fr<BMemory<Memory>> & recv_buf,
                            openfpm::vector<size_t> * sz,
                            openfpm::vector<size_t> * sz_byte,
                            op & op_param,
                            size_t opt)
    {
        if (sz_byte != NULL)
            sz_byte->resize(recv_buf.size());
 
        for (size_t i = 0 ; i < recv_buf.size() ; i++)
        {
            T unp;
 
            ExtPreAlloc<Memory> & mem = *(new ExtPreAlloc<Memory>(recv_buf.get(i).size(),recv_buf.get(i)));
            mem.incRef();
 
            Unpack_stat ps;
 
            Unpacker<T,Memory>::template unpack<>(mem, unp, ps);
 
            size_t recv_size_old = recv.size();
            // Merge the information
 
            op_param.template execute<true,T,decltype(recv),decltype(unp),layout_base,prp...>(recv,unp,i,opt);
 
            size_t recv_size_new = recv.size();
 
            if (sz_byte != NULL)
                sz_byte->get(i) = recv_buf.get(i).size();
            if (sz != NULL)
                sz->get(i) = recv_size_new - recv_size_old;
 
            mem.decRef();
            delete &mem;
        }
    }
};
 
template<typename op, typename Vt, typename S, template<typename> class layout_base, typename v_mpl>
struct unpack_each_prop_buffer
{
    S & recv;
 
    openfpm::vector_fr<BMemory<HeapMemory>> & recv_buf;
 
    size_t i;
 
    op & op_param;
 
    openfpm::vector<size_t> * sz;
 
    openfpm::vector<size_t> * sz_byte;
 
    inline unpack_each_prop_buffer(S & recv,
                                   openfpm::vector_fr<BMemory<HeapMemory>> & recv_buf,
                                   op & op_param,
                                   size_t i,
                                   openfpm::vector<size_t> * sz,
                                   openfpm::vector<size_t> * sz_byte)
    :recv(recv),recv_buf(recv_buf),op_param(op_param),i(i),sz(sz),sz_byte(sz_byte)
    {};
 
    template<typename T>
    inline void operator()(T& t) const
    {
        // here we get the the type of the property at position T::value
        typedef typename boost::mpl::at<typename T::value_type::type,boost::mpl::int_<T::value> >::type prp_type;
 
        // here we get the the type of the property at position T::value
        typedef typename boost::mpl::at<v_mpl,boost::mpl::int_<T::value>>::type prp_num;
 
        // calculate the number of received elements
        size_t n_ele = recv_buf.get(i).size() / sizeof(prp_type);
 
        // add the received particles to the vector
        PtrMemory * ptr1 = new PtrMemory(recv_buf.get(i).getPointer(),recv_buf.get(i).size());
 
        // create vector representation to a piece of memory already allocated
        openfpm::vector<typename Vt::value_type,PtrMemory,layout_base,openfpm::grow_policy_identity> v2;
 
        v2.template setMemory<prp_num::value>(*ptr1);
 
        // resize with the number of elements
        v2.resize(n_ele);
 
        // Merge the information
 
        size_t recv_size_old = recv.size();
 
        op_param.template execute<false,T,decltype(recv),decltype(v2),layout_base,prp_num::value>(recv,v2,i);
 
        size_t recv_size_new = recv.size();
 
        if (sz_byte != NULL)
            sz_byte->get(i) = recv_buf.get(i).size();
        if (sz != NULL)
            sz->get(i) = recv_size_new - recv_size_old;
    }
};
 
template<typename sT, template<typename> class layout_base,typename Memory>
struct process_receive_mem_traits_inte
{
    size_t i;
 
    openfpm::vector_fr<BMemory<Memory>> & recv_buf;
 
    openfpm::vector<typename sT::value_type,PtrMemory,layout_base,openfpm::grow_policy_identity> & v2;
 
    size_t n_ele = 0;
 
    // options
    size_t opt;
 
    inline process_receive_mem_traits_inte(openfpm::vector<typename sT::value_type,PtrMemory,layout_base,openfpm::grow_policy_identity> & v2,
                                           openfpm::vector_fr<BMemory<Memory>> & recv_buf,
                                           size_t i,
                                           size_t opt)
    :i(i),recv_buf(recv_buf),v2(v2),opt(opt)
    {};
 
    template<typename T>
    inline void operator()(T& t)
    {
        typedef typename boost::mpl::at<typename sT::value_type::type,T>::type type_prp;
 
        // calculate the number of received elements
        this->n_ele = recv_buf.get(i).size() / sizeof(type_prp);
 
        PtrMemory * ptr1;
 
        if (opt & MPI_GPU_DIRECT)
        {
#if defined(MPIX_CUDA_AWARE_SUPPORT) && MPIX_CUDA_AWARE_SUPPORT
            // add the received particles to the vector
            ptr1 = new PtrMemory(recv_buf.get(i).getDevicePointer(),recv_buf.get(i).size());
#else
            // add the received particles to the vector
            ptr1 = new PtrMemory(recv_buf.get(i).getPointer(),recv_buf.get(i).size());
#endif
        }
        else
        {
            // add the received particles to the vector
            ptr1 = new PtrMemory(recv_buf.get(i).getPointer(),recv_buf.get(i).size());
        }
 
        v2.template setMemory<T::value>(*ptr1);
 
        ++i;
    }
};
 
template<bool inte_or_lin,typename T, typename S, template<typename> class layout_base,typename Memory>
struct unpack_selector_with_prp_lin
{
    template<typename op, unsigned int ... prp> static int call_unpack_impl(S & recv,
                                                                             openfpm::vector_fr<BMemory<Memory>> & recv_buf,
                                                                             openfpm::vector<size_t> * sz,
                                                                             openfpm::vector<size_t> * sz_byte,
                                                                             op & op_param,
                                                                             size_t i,
                                                                             size_t opt)
    {
        // create vector representation to a piece of memory already allocated
        openfpm::vector<typename T::value_type,PtrMemory,layout_base,openfpm::grow_policy_identity> v2;
 
        process_receive_mem_traits_inte<T,layout_base,Memory> prmti(v2,recv_buf,i,opt);
 
        boost::mpl::for_each_ref<boost::mpl::range_c<int,0,T::value_type::max_prop>>(prmti);
 
        v2.resize(prmti.n_ele);
 
        // Merge the information
 
        size_t recv_size_old = recv.size();
 
        op_param.template execute<false,T,decltype(recv),decltype(v2),layout_base,prp...>(recv,v2,i,opt);
 
        size_t recv_size_new = recv.size();
 
        if (sz_byte != NULL)
            sz_byte->get(i) = recv_buf.get(i).size();
        if (sz != NULL)
            sz->get(i) = recv_size_new - recv_size_old;
 
        return sizeof...(prp);
    }
};
 
template<typename T, typename S, template<typename> class layout_base, typename Memory>
struct unpack_selector_with_prp_lin<true,T,S,layout_base,Memory>
{
    template<typename op, unsigned int ... prp> static int call_unpack_impl(S & recv,
                                                                             openfpm::vector_fr<BMemory<Memory>> & recv_buf,
                                                                             openfpm::vector<size_t> * sz,
                                                                             openfpm::vector<size_t> * sz_byte,
                                                                             op & op_param,
                                                                             size_t i,
                                                                             size_t opt)
    {
        // calculate the number of received elements
        size_t n_ele = recv_buf.get(i).size() / sizeof(typename T::value_type);
 
        // add the received particles to the vector
        PtrMemory * ptr1 = new PtrMemory(recv_buf.get(i).getPointer(),recv_buf.get(i).size());
        ptr1->incRef();
 
        {
        // create vector representation to a piece of memory already allocated
        openfpm::vector<typename T::value_type,PtrMemory,layout_base,openfpm::grow_policy_identity> v2;
 
        v2.setMemory(*ptr1);
 
        // resize with the number of elements
        v2.resize(n_ele);
 
        // Merge the information
 
        size_t recv_size_old = recv.size();
 
        op_param.template execute<false,T,decltype(recv),decltype(v2),layout_base,prp...>(recv,v2,i,opt);
 
        size_t recv_size_new = recv.size();
 
        if (sz_byte != NULL)
            sz_byte->get(i) = recv_buf.get(i).size();
        if (sz != NULL)
            sz->get(i) = recv_size_new - recv_size_old;
        }
 
        ptr1->decRef();
        delete ptr1;
        return 1;
    }
};
 
typedef aggregate<int,int> dummy_type;
 
//
template<typename T, typename S, template<typename> class layout_base, typename Memory>
struct unpack_selector_with_prp<true,T,S,layout_base,Memory>
{
    template<typename op, unsigned int ... prp> static void call_unpack(S & recv,
                                                                        openfpm::vector_fr<BMemory<Memory>> & recv_buf,
                                                                        openfpm::vector<size_t> * sz,
                                                                        openfpm::vector<size_t> * sz_byte,
                                                                        op & op_param,
                                                                        size_t opt)
    {
        if (sz_byte != NULL)
            sz_byte->resize(recv_buf.size());
 
        for (size_t i = 0 ; i < recv_buf.size() ; )
        {
            i += unpack_selector_with_prp_lin<is_layout_mlin<layout_base<dummy_type>>::value,T,S,layout_base,Memory>::template call_unpack_impl<op,prp...>(recv,recv_buf,sz,sz_byte,op_param,i,opt);
        }
    }
};
 
 
template<typename T>
struct call_serialize_variadic {};
 
template<int ... prp>
struct call_serialize_variadic<index_tuple<prp...>>
{
    template<typename T> inline static void call_pr(T & send, size_t & tot_size)
    {
        Packer<T,HeapMemory>::template packRequest<prp...>(send,tot_size);
    }
 
    template<typename T> inline static void call_pack(ExtPreAlloc<HeapMemory> & mem, T & send, Pack_stat & sts)
    {
        Packer<T,HeapMemory>::template pack<prp...>(mem,send,sts);
    }
 
    template<typename op, typename T, typename S, template<typename> class layout_base, typename Memory>
    inline static void call_unpack(S & recv,
                                   openfpm::vector_fr<BMemory<Memory>> & recv_buf,
                                   openfpm::vector<size_t> * sz,
                                   openfpm::vector<size_t> * sz_byte,
                                   op & op_param,
                                   size_t opt)
    {
        const bool result = has_pack_gen<typename T::value_type>::value == false && is_vector<T>::value == true;
 
        unpack_selector_with_prp<result, T, S,layout_base,Memory>::template call_unpack<op,prp...>(recv, recv_buf, sz, sz_byte, op_param,opt);
    }
};
 
template<typename sT>
struct set_buf_pointer_for_each_prop
{
    sT & v;
 
    openfpm::vector<const void *> & send_buf;
 
    size_t opt;
 
    inline set_buf_pointer_for_each_prop(sT & v, openfpm::vector<const void *> & send_buf, size_t opt)
    :v(v),send_buf(send_buf),opt(opt)
    {};
 
    template<typename T>
    inline void operator()(T& t) const
    {
        // If we have GPU direct activated use directly the cuda buffer
        if (opt & MPI_GPU_DIRECT)
        {
#if defined(MPIX_CUDA_AWARE_SUPPORT) && MPIX_CUDA_AWARE_SUPPORT
            send_buf.add(v.template getDeviceBuffer<T::value>());
#else
            v.template deviceToHost<T::value>();
            send_buf.add(v.template getPointer<T::value>());
#endif
        }
        else
        {
            send_buf.add(v.template getPointer<T::value>());
        }
    }
};
 
template<typename sT>
struct set_buf_size_for_each_prop
{
    sT & v;
 
    openfpm::vector<size_t> & sz;
 
    inline set_buf_size_for_each_prop(sT & v, openfpm::vector<size_t> & sz)
    :v(v),sz(sz)
    {};
 
    template<typename T>
    inline void operator()(T& t) const
    {
        typedef typename boost::mpl::at<typename sT::value_type::type,T>::type type_prp;
 
        sz.add(sizeof(type_prp)*v.size());
    }
};
 
template<typename T, bool impl = is_multiple_buffer_each_prp<T>::value >
struct pack_unpack_cond_with_prp_inte_lin
{
    static void set_buffers(T & send, openfpm::vector<const void *> & send_buf, size_t opt)
    {
        send_buf.add(send.getPointer());
    }
 
    static void set_size_buffers(T & send, openfpm::vector<size_t> & sz)
    {
        sz.add(send.size()*sizeof(typename T::value_type));
    }
 
    static void construct_prc(openfpm::vector<size_t> & prc_send, openfpm::vector<size_t> & prc_send_)
    {
        for (size_t i = 0 ; i < prc_send.size() ; i++)
        {
            prc_send_.add(prc_send.get(i));
        }
    }
};
 
// memory_traits_inte
template<typename T>
struct pack_unpack_cond_with_prp_inte_lin<T,true>
{
    static void set_buffers(T & send, openfpm::vector<const void *> & send_buf, size_t opt)
    {
        set_buf_pointer_for_each_prop<T> sbp(send,send_buf,opt);
 
        boost::mpl::for_each_ref<boost::mpl::range_c<int,0,T::value_type::max_prop>>(sbp);
    }
 
    static void set_size_buffers(T & send, openfpm::vector<size_t> & sz)
    {
        set_buf_size_for_each_prop<T> sbp(send,sz);
 
        boost::mpl::for_each_ref<boost::mpl::range_c<int,0,T::value_type::max_prop>>(sbp);
    }
 
    static void construct_prc(openfpm::vector<size_t> & prc_send, openfpm::vector<size_t> & prc_send_)
    {
        for (size_t i = 0 ; i < prc_send.size() ; i++)
        {
            for (size_t j = 0 ; j < T::value_type::max_prop ; j++)
            {prc_send_.add(prc_send.get(i));}
        }
    }
};
 
template<bool cond,
         typename op,
         typename T,
         typename S,
         template <typename> class layout_base,
         unsigned int ... prp>
struct pack_unpack_cond_with_prp
{
    static void packingRequest(T & send, size_t & tot_size, openfpm::vector<size_t> & sz)
    {
        typedef typename ::generate_indexes<int, has_max_prop<T, has_value_type_ofp<T>::value>::number, MetaFuncOrd>::result ind_prop_to_pack;
        if (has_pack_gen<typename T::value_type>::value == false && is_vector<T>::value == true)
        {
            pack_unpack_cond_with_prp_inte_lin<T>::set_size_buffers(send,sz);
        }
        else
        {
            call_serialize_variadic<ind_prop_to_pack>::call_pr(send,tot_size);
 
            sz.add(tot_size);
        }
    }
 
    static void packing(ExtPreAlloc<HeapMemory> & mem, T & send, Pack_stat & sts, openfpm::vector<const void *> & send_buf, size_t opt = 0)
    {
        typedef typename ::generate_indexes<int, has_max_prop<T, has_value_type_ofp<T>::value>::number, MetaFuncOrd>::result ind_prop_to_pack;
        if (has_pack_gen<typename T::value_type>::value == false && is_vector<T>::value == true)
        {
            pack_unpack_cond_with_prp_inte_lin<T>::set_buffers(send,send_buf,opt);
        }
        else
        {
            send_buf.add(mem.getPointerEnd());
            call_serialize_variadic<ind_prop_to_pack>::call_pack(mem,send,sts);
        }
    }
 
    template<typename Memory>
    static void unpacking(S & recv,
                          openfpm::vector_fr<BMemory<Memory>> & recv_buf,
                          openfpm::vector<size_t> * sz,
                          openfpm::vector<size_t> * sz_byte,
                          op & op_param,
                          size_t opt)
    {
        typedef index_tuple<prp...> ind_prop_to_pack;
        call_serialize_variadic<ind_prop_to_pack>::template call_unpack<op,T,S,layout_base>(recv, recv_buf, sz, sz_byte, op_param,opt);
    }
};
 
 
 
template<bool sr>
struct op_ssend_recv_add_sr
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp> static void execute(D & recv,S & v2, size_t i, size_t opt)
    {
        if (opt & MPI_GPU_DIRECT)
        {
#if defined(MPIX_CUDA_AWARE_SUPPORT) && MPIX_CUDA_AWARE_SUPPORT
 
            // Merge the information
            recv.template add_prp_device<typename T::value_type,
                              PtrMemory,
                              openfpm::grow_policy_identity,
                              openfpm::vect_isel<typename T::value_type>::value,
                              layout_base,
                              prp...>(v2);
#else
            size_t old_size = recv.size();
 
            // Merge the information
            recv.template add_prp<typename T::value_type,
                              PtrMemory,
                              openfpm::grow_policy_identity,
                              openfpm::vect_isel<typename T::value_type>::value,
                              layout_base,
                              prp...>(v2);
 
            recv.template hostToDevice<prp...>(old_size,old_size+v2.size()-1);
 
#endif
 
        }
        else
        {
            // Merge the information
            recv.template add_prp<typename T::value_type,
                              PtrMemory,
                              openfpm::grow_policy_identity,
                              openfpm::vect_isel<typename T::value_type>::value,
                              layout_base,
                              prp...>(v2);
        }
    }
};
 
template<>
struct op_ssend_recv_add_sr<true>
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    static void execute(D & recv,S & v2, size_t i,size_t opt)
    {
        // Merge the information
        recv.template add_prp<typename T::value_type,
                              HeapMemory,
                              typename T::grow_policy,
                              openfpm::vect_isel<typename T::value_type>::value,
                              layout_base,
                              prp...>(v2);
    }
};
 
template<typename op>
struct op_ssend_recv_add
{
    template<bool sr,
             typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    static void execute(D & recv,S & v2, size_t i, size_t opt)
    {
        // Merge the information
        op_ssend_recv_add_sr<sr>::template execute<T,D,S,layout_base,prp...>(recv,v2,i,opt);
    }
};
 
template<bool sr,template<typename,typename> class op, typename vector_type_opart>
struct op_ssend_recv_merge_impl
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    inline static void execute(D & recv,S & v2,size_t i,vector_type_opart & opart)
    {
        // Merge the information
        recv.template merge_prp_v<op,
                                  typename T::value_type,
                                  PtrMemory,
                                  openfpm::grow_policy_identity,
                                  layout_base,
                                  typename vector_type_opart::value_type,
                                  prp...>(v2,opart.get(i));
    }
};
 
template<template<typename,typename> class op, typename vector_type_opart>
struct op_ssend_recv_merge_impl<true,op,vector_type_opart>
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    inline static void execute(D & recv,S & v2,size_t i,vector_type_opart & opart)
    {
        // Merge the information
        recv.template merge_prp_v<op,
                                  typename T::value_type,
                                  HeapMemory,
                                  openfpm::grow_policy_double,
                                  layout_base,
                                  typename vector_type_opart::value_type,
                                  prp...>(v2,opart.get(i));
    }
};
 
template<template<typename,typename> class op, typename vector_type_opart>
struct op_ssend_recv_merge
{
    vector_type_opart & opart;
 
    op_ssend_recv_merge(vector_type_opart & opart)
    :opart(opart)
    {}
 
    template<bool sr,
             typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    void execute(D & recv,S & v2,size_t i,size_t opt)
    {
        op_ssend_recv_merge_impl<sr,op,vector_type_opart>::template execute<T,D,S,layout_base,prp...>(recv,v2,i,opart);
    }
};
 
template<bool sr,template<typename,typename> class op, typename vector_type_opart, typename vector_type_prc_offset>
struct op_ssend_recv_merge_gpu_impl
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    inline static void execute(D & recv,S & v2,size_t i,vector_type_opart & opart, vector_type_prc_offset & prc_off)
    {
        prc_off.template deviceToHost<0>();
 
        unsigned int start = 0;
        unsigned int stop = prc_off.template get<0>(i / sizeof...(prp));
 
        if (i != 0)
        {start = prc_off.template get<0>(i / sizeof...(prp)-1);}
 
        // Merge the information
        recv.template merge_prp_v_device<op,
                                  typename T::value_type,
                                  PtrMemory,
                                  openfpm::grow_policy_identity,
                                  layout_base,
                                  vector_type_opart,
                                  prp...>(v2,opart,start,stop);
    }
};
 
template<template<typename,typename> class op, typename vector_type_opart, typename vector_type_prc_offset>
struct op_ssend_recv_merge_gpu_impl<true,op,vector_type_opart,vector_type_prc_offset>
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    inline static void execute(D & recv,S & v2,size_t i,vector_type_opart & opart, vector_type_prc_offset & prc_off)
    {
        std::cout << __FILE__ << ":" << __LINE__ << " Error: not implemented" << std::endl;
    }
};
 
template<template<typename,typename> class op, typename vector_type_opart, typename vector_type_prc_offset>
struct op_ssend_recv_merge_gpu
{
    vector_type_opart & opart;
 
    vector_type_prc_offset & prc_offset;
 
    op_ssend_recv_merge_gpu(vector_type_opart & opart, vector_type_prc_offset & prc_offset)
    :opart(opart),prc_offset(prc_offset)
    {}
 
    template<bool sr,
             typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    void execute(D & recv,S & v2,size_t i,size_t opt)
    {
        op_ssend_recv_merge_gpu_impl<sr,op,vector_type_opart,vector_type_prc_offset>::template execute<T,D,S,layout_base,prp...>(recv,v2,i,opart,prc_offset);
    }
};
 
template<bool sr>
struct op_ssend_gg_recv_merge_impl
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    inline static void execute(D & recv,S & v2,size_t i,size_t & start)
    {
        // Merge the information
        recv.template merge_prp_v<replace_,
                                  typename T::value_type,
                                  PtrMemory,
                                  openfpm::grow_policy_identity,
                                  layout_base,
                                  prp...>(v2,start);
 
        start += v2.size();
    }
};
 
template<bool sr>
struct op_ssend_gg_recv_merge_impl_run_device
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    inline static void execute(D & recv,S & v2,size_t i,size_t & start)
    {
        // Merge the information
        recv.template merge_prp_v<replace_,
                                  typename T::value_type,
                                  typename S::Memory_type,
                                  openfpm::grow_policy_identity,
                                  layout_base,
                                  prp...>(v2,start);
 
        recv.template hostToDevice<prp ...>(start,start+v2.size()-1);
 
        start += v2.size();
    }
};
 
template<bool sr>
struct op_ssend_gg_recv_merge_impl_run_device_direct
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp>
    inline static void execute(D & recv,S & v2,size_t i,size_t & start)
    {
        // Merge the information
        recv.template merge_prp_device<replace_,
                                  typename T::value_type,
                                  typename S::Memory_type,
                                  openfpm::grow_policy_identity,
                                  prp...>(v2,start);
 
        start += v2.size();
    }
};
 
template<>
struct op_ssend_gg_recv_merge_impl<true>
{
    template<typename T,
             typename D,
             typename S,
             template <typename> class layout_base,
             int ... prp> inline static void execute(D & recv,S & v2,size_t i,size_t & start)
    {
        // Merge the information
        recv.template merge_prp_v<replace_,
                                  typename T::value_type,
                                  HeapMemory,
                                  typename S::grow_policy,
                                  layout_base,
                                  prp...>(v2,start);
 
        // from
        start += v2.size();
    }
};
 
struct op_ssend_gg_recv_merge
{
    size_t start;
 
    op_ssend_gg_recv_merge(size_t start)
    :start(start)
    {}
 
    template<bool sr, typename T, typename D, typename S, template<typename> class layout_base, int ... prp> void execute(D & recv,S & v2,size_t i,size_t opt)
    {
        op_ssend_gg_recv_merge_impl<sr>::template execute<T,D,S,layout_base,prp...>(recv,v2,i,start);
    }
};
 
struct op_ssend_gg_recv_merge_run_device
{
    size_t start;
 
    op_ssend_gg_recv_merge_run_device(size_t start)
    :start(start)
    {}
 
    template<bool sr, typename T, typename D, typename S, template<typename> class layout_base, int ... prp> void execute(D & recv,S & v2,size_t i,size_t opt)
    {
        bool active = is_mpi_rdma_cuda_active();
        if (active == true)
        {op_ssend_gg_recv_merge_impl_run_device_direct<sr>::template execute<T,D,S,layout_base,prp...>(recv,v2,i,start);}
        else
        {op_ssend_gg_recv_merge_impl_run_device<sr>::template execute<T,D,S,layout_base,prp...>(recv,v2,i,start);}
    }
};
 
 
 
 
#endif /* OPENFPM_VCLUSTER_SRC_VCLUSTER_VCLUSTER_META_FUNCTION_HPP_ */