https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Intel-Fortran-11-MPICH2-32bit-Windows-based-system-Bad/m-p/878540#M3262 Hello,<BR /><BR />I have written 2 codes. The first one is a <STRONG>non-parallel </STRONG>code and the second one uses <STRONG>MPICH2</STRONG>. They are compiled with <STRONG>Intel Fortran 11</STRONG> (Trial Version) on <STRONG>32 bits Windows</STRONG> system. <BR /><BR />The options for the compilations are <STRONG>/fast</STRONG> . So, the first code is compiled by the following command:<BR /><STRONG><BR />ifort /fast -o a1.exe mycode.f90</STRONG> <BR /><BR />The second code, implementing MPICH2, is compiled by:<BR /><BR /><STRONG>ifort /fast -o a2.exe mycode.f90 /link fmpich2.lib</STRONG><BR /> <BR /><BR />The compilations were successfully finished and gave 2 executables.<BR /><BR />When I launch <STRONG>a2.exe</STRONG> using only 1 CPU, I find it runs almost <STRONG>2 times slower</STRONG> than the <STRONG>a1.exe</STRONG>. Note that two executables treat the <STRONG>same problem</STRONG> and give the <STRONG>same results</STRONG> (Sure!).<BR /><BR />I seem the Intel Fortran optimization (here /fast) fails when I compile a mpich code.<BR /><BR /><BR />Anyone has experience on this problem? Please give me some advice!<BR /><BR />Thank you,<BR /><BR />tnvquynh.<BR /><BR /><BR /><BR /> Wed, 17 Jun 2009 04:58:37 GMT tnvquynh 2009-06-17T04:58:37Z https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Intel-Fortran-11-MPICH2-32bit-Windows-based-system-Bad/m-p/878540#M3262 Hello,<BR /><BR />I have written 2 codes. The first one is a <STRONG>non-parallel </STRONG>code and the second one uses <STRONG>MPICH2</STRONG>. They are compiled with <STRONG>Intel Fortran 11</STRONG> (Trial Version) on <STRONG>32 bits Windows</STRONG> system. <BR /><BR />The options for the compilations are <STRONG>/fast</STRONG> . So, the first code is compiled by the following command:<BR /><STRONG><BR />ifort /fast -o a1.exe mycode.f90</STRONG> <BR /><BR />The second code, implementing MPICH2, is compiled by:<BR /><BR /><STRONG>ifort /fast -o a2.exe mycode.f90 /link fmpich2.lib</STRONG><BR /> <BR /><BR />The compilations were successfully finished and gave 2 executables.<BR /><BR />When I launch <STRONG>a2.exe</STRONG> using only 1 CPU, I find it runs almost <STRONG>2 times slower</STRONG> than the <STRONG>a1.exe</STRONG>. Note that two executables treat the <STRONG>same problem</STRONG> and give the <STRONG>same results</STRONG> (Sure!).<BR /><BR />I seem the Intel Fortran optimization (here /fast) fails when I compile a mpich code.<BR /><BR /><BR />Anyone has experience on this problem? Please give me some advice!<BR /><BR />Thank you,<BR /><BR />tnvquynh.<BR /><BR /><BR /><BR /> Wed, 17 Jun 2009 04:58:37 GMT https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Intel-Fortran-11-MPICH2-32bit-Windows-based-system-Bad/m-p/878540#M3262 tnvquynh 2009-06-17T04:58:37Z https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Intel-Fortran-11-MPICH2-32bit-Windows-based-system-Bad/m-p/878541#M3263 <DIV style="margin:0px;"></DIV> Run with a profiler such as PTU, preferably with static link, and high threshold for yielding CPU during spin waits. Or, use an MPI profiler, such as Intel Trace Analyzer/Collector, which should be compatible with MPICH2. You are getting more into the territory of the HPC forum.<BR /> Wed, 17 Jun 2009 05:17:02 GMT https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Intel-Fortran-11-MPICH2-32bit-Windows-based-system-Bad/m-p/878541#M3263 TimP 2009-06-17T05:17:02Z