https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822262#M4893 <DIV><SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;">Another question is that alought DGEMM is repeatable, if the result will shift with real result of matrix</SPAN></DIV><DIV><SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;">multiplication. And how much it will shift?</SPAN></DIV><DIV><SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;"><BR /></SPAN></DIV><DIV><SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;"><BR /></SPAN><DIV></DIV></DIV> Mon, 16 May 2011 18:42:40 GMT Emily_Zhu 2011-05-16T18:42:40Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822258#M4889 Is anyone can solve the problem that mkl functions didn't get consistant result?<DIV>we use 10.3.3, It looks any function like DGEMM, DSYEV get different results</DIV><DIV>The different is big since for 1000*1000 matrix, the max difference for single element is</DIV><DIV>the matrix is 1.0e-14.</DIV><DIV></DIV><DIV>We also try 16 boundary alignment, but it doesn't work.</DIV> Fri, 13 May 2011 17:14:43 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822258#M4889 Emily_Zhu 2011-05-13T17:14:43Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822259#M4890 We've generally seen that if you are running your program on the same computer, with the same number of threads and your memory is aligned on 16-byte boundaries then you'll get consistent results from DGEMM from one run to the next. <BR /><BR />We'd need to know a little more about your program and the system on which you are running. Fri, 13 May 2011 23:46:10 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822259#M4890 Todd_R_Intel 2011-05-13T23:46:10Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822260#M4891 If I set each computer as mkl_sequential and use 16 boundary alignment.<DIV>but send whole computation to cluster, if the final result will be same?</DIV> Mon, 16 May 2011 06:47:33 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822260#M4891 Emily_Zhu 2011-05-16T06:47:33Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822261#M4892 <DIV id="tiny_quote"> <DIV style="margin-left: 2px; margin-right: 2px;">Quoting <A rel="/en-us/services/profile/quick_profile.php?is_paid=&amp;user_id=520446" class="basic" href="https://community.intel.com/en-us/profile/520446/">Emily Zhu</A></DIV> <DIV style="background-color: #e5e5e5; padding: 5px; border: 1px; border-style: inset; margin-left: 2px; margin-right: 2px;"><I>If I set each computer as mkl_sequential and use 16 boundary alignment.<DIV>but send whole computation to cluster, if the final result will be same?</DIV></I></DIV></DIV><P></P>If each node has the same software installation and same CPU brand, it's reasonable to hope for identical results. Mon, 16 May 2011 14:16:34 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822261#M4892 TimP 2011-05-16T14:16:34Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822262#M4893 <DIV><SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;">Another question is that alought DGEMM is repeatable, if the result will shift with real result of matrix</SPAN></DIV><DIV><SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;">multiplication. And how much it will shift?</SPAN></DIV><DIV><SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;"><BR /></SPAN></DIV><DIV><SPAN style="font-family: Verdana, Arial, Helvetica, sans-serif;"><BR /></SPAN><DIV></DIV></DIV> Mon, 16 May 2011 18:42:40 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/no-consistent-result-with-mkl-functions/m-p/822262#M4893 Emily_Zhu 2011-05-16T18:42:40Z