https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Pardiso-solver-no-threaded-speedup/m-p/1580862#M35928 <P>We find that linking the threaded MKL Pardiso solver provides no performance boost - or even slows down if many threads are specified via OPNOMP_NUM_THREADS.</P><P>The Pardiso diagnostic messages report multiple threads are being used.&nbsp; But the execution time slows.</P><P>Is this to be expected?</P><P>&nbsp;</P> Fri, 15 Mar 2024 16:24:59 GMT Dave17 2024-03-15T16:24:59Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Pardiso-solver-no-threaded-speedup/m-p/1580862#M35928 <P>We find that linking the threaded MKL Pardiso solver provides no performance boost - or even slows down if many threads are specified via OPNOMP_NUM_THREADS.</P><P>The Pardiso diagnostic messages report multiple threads are being used.&nbsp; But the execution time slows.</P><P>Is this to be expected?</P><P>&nbsp;</P> Fri, 15 Mar 2024 16:24:59 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Pardiso-solver-no-threaded-speedup/m-p/1580862#M35928 Dave17 2024-03-15T16:24:59Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Pardiso-solver-no-threaded-speedup/m-p/1581740#M35930 <P>First of all try to play with MKL_NUM_THREADS environment variables. </P><P>Then enable verbose mode ( setting msglvl == 1 ) and see the perf results for each of phases ( reordering, factorization, solving) and see the scalability results if any or any performance bottleneck </P><P>--Gennady</P><P><BR /></P><BR /> Tue, 19 Mar 2024 10:17:27 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Pardiso-solver-no-threaded-speedup/m-p/1581740#M35930 Gennady_F_Intel 2024-03-19T10:17:27Z