<div dir="ltr">Hi all,<div><br></div><div>I am doing a Mixed mode parallel simulation on cp2k 3.0. The package was installed by using the install_toolchain script. </div><div><br></div><div>Here is part of the performance when running:</div><div><br></div><div><br></div><div><div> -------------------------------------------------------------------------------</div><div> -                                                                             -</div><div> -                         MESSAGE PASSING PERFORMANCE                         -</div><div> -                                                                             -</div><div> -------------------------------------------------------------------------------</div><div><br></div><div> ROUTINE             CALLS                TOT TIME [s]          AVE VOLUME [Bytes]          PERFORMANCE [MB/s]</div><div> MP_Group                   5                       0.000</div><div> MP_Bcast             57138                    75.397                         10390.                                              7.87</div><div> MP_Allreduce     113420                2192.750                             296.                                              0.02</div><div> MP_Gather               102                      0.929                             640.                                             0.07</div><div> MP_Sync                    63                    83.271 </div><div> MP_Alltoall          113191               2289.461                       533721.                                           26.39</div><div> MP_SendRecv        6324                     6.835                             1283.                                           1.19</div><div> MP_ISendRecv    262260                   1.986                            51157.                                     6755.54</div><div> MP_Wait            2340964              3375.440</div><div> MP_comm_split            9                     0.438</div><div> MP_ISend           1357992                   9.288                           34472.                                     5039.93</div><div> MP_IRecv            1358200                  2.225                            34115.                                   20822.97</div><div> MP_Recv                   770                    2.647                         178840.                                          52.02</div><div> MP_Memory       1400158                   0.753</div><div> -------------------------------------------------------------------------------</div></div><div><br></div><div><font size="4"><br></font></div><div><font size="4">The total time for this calculation is 3 hours---10800s, so that means "MP_WAIT" takes nearly 1 hour. Anyone knows is it normal?</font></div><div><br></div><div><font size="4">Following is the libs I used when running:</font></div><div><br></div><div style="text-align: left;"><div><br></div><div><br></div><div><br></div><div><div> CP2K| version string:                                          CP2K version 3.0</div><div> CP2K| source code revision number:                                    svn:16458</div><div> CP2K| cp2kflags: libint fftw3 libxc pexsi elpa3 parallel mpi3 scalapack quip sm</div><div> CP2K|            m_dnn smm libderiv_max_am1=5 libint_max_am=6</div><div> CP2K| is freely available from                            https://www.cp2k.org/</div><div> CP2K| Program compiled at                          Wed Mar 16 01:20:41 GMT 2016</div><div> CP2K| Program compiled on                                                login1</div><div> CP2K| Program compiled for                                                local</div></div><div><br></div><div><br></div><div><br></div><div>Please give me some advise, the calculation is too slow, there must be something wrong. I've tried same input file and cores in ARCHER, it's 5 time faster.</div><div><br></div></div></div>