https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Inconsistent-results-for-Pardiso-with-multiple-threads/m-p/860849#M7504 Hi,<BR /><BR />I am running a finite element analysis code which uses Pardiso as the linear solver. If I use one thread, the log files are identical except for the time of the run. If I use four threads, there are significant differences in the number of iterations and right-hand-side evaluations. Do you have any idea why this might be?<BR /><BR />Thanks,<BR /><BR />Dave Fri, 26 Mar 2010 00:16:55 GMT Rawlins__David 2010-03-26T00:16:55Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Inconsistent-results-for-Pardiso-with-multiple-threads/m-p/860849#M7504 Hi,<BR /><BR />I am running a finite element analysis code which uses Pardiso as the linear solver. If I use one thread, the log files are identical except for the time of the run. If I use four threads, there are significant differences in the number of iterations and right-hand-side evaluations. Do you have any idea why this might be?<BR /><BR />Thanks,<BR /><BR />Dave Fri, 26 Mar 2010 00:16:55 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Inconsistent-results-for-Pardiso-with-multiple-threads/m-p/860849#M7504 Rawlins__David 2010-03-26T00:16:55Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Inconsistent-results-for-Pardiso-with-multiple-threads/m-p/860850#M7505 Hi Dave, <DIV>the similar question has been discussed <A href="http://software.intel.com/en-us/forums/showthread.php?t=70105">here</A>.</DIV> <DIV>May be make sense to add that the problem will depends on conditional numbers of the inputmatrices.</DIV> <DIV>especially if the inputmatricesis ill-conditioned.</DIV> <DIV>--Gennady</DIV> <DIV></DIV> Fri, 26 Mar 2010 11:30:04 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Inconsistent-results-for-Pardiso-with-multiple-threads/m-p/860850#M7505 Gennady_F_Intel 2010-03-26T11:30:04Z