topic Incomplete factorization... in Intel® oneAPI Math Kernel Library https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Incomplete-factorization/m-p/1000704#M18504 <P><SPAN style="font-size: 1em; line-height: 1.5;">Hello,</SPAN></P> <P>I am testing MKL ilu factorization. As it only works for ilu0, I provide a sparse matrix pattern that corresponds to ilu(k). One thing I observe is that the performance is getting slower when I increase the level of fills (I have compared with matlab for comparison). Is this expected ?</P> <P>For instance, I have the following results from mkl.</P> <P>level0, nnz = 9,027,150, time =&nbsp;0.598093, matlab =&nbsp;0.766352&nbsp;</P> <P>level1, nnz =&nbsp;12,816,050, time =&nbsp;1.3787, matlab =&nbsp;1.31876</P> <P>level2, nnz = 15,450,825, time =&nbsp;2.57172, matlab =&nbsp;1.9329</P> <P>This seems to be odd.&nbsp;<SPAN style="font-size: 1em; line-height: 1.5;">Does mkl perform something special ? any pivoting ? or more specialized for sparser patterns ? &nbsp;</SPAN></P> <P>&nbsp;</P> <P>Thanks,</P> <P>Kyungjoo</P> <P>&nbsp;</P> <P>&nbsp;</P> Tue, 07 Jul 2015 00:02:57 GMT Kim__Kyungjoo 2015-07-07T00:02:57Z Incomplete factorization... https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Incomplete-factorization/m-p/1000704#M18504 <P><SPAN style="font-size: 1em; line-height: 1.5;">Hello,</SPAN></P> <P>I am testing MKL ilu factorization. As it only works for ilu0, I provide a sparse matrix pattern that corresponds to ilu(k). One thing I observe is that the performance is getting slower when I increase the level of fills (I have compared with matlab for comparison). Is this expected ?</P> <P>For instance, I have the following results from mkl.</P> <P>level0, nnz = 9,027,150, time =&nbsp;0.598093, matlab =&nbsp;0.766352&nbsp;</P> <P>level1, nnz =&nbsp;12,816,050, time =&nbsp;1.3787, matlab =&nbsp;1.31876</P> <P>level2, nnz = 15,450,825, time =&nbsp;2.57172, matlab =&nbsp;1.9329</P> <P>This seems to be odd.&nbsp;<SPAN style="font-size: 1em; line-height: 1.5;">Does mkl perform something special ? any pivoting ? or more specialized for sparser patterns ? &nbsp;</SPAN></P> <P>&nbsp;</P> <P>Thanks,</P> <P>Kyungjoo</P> <P>&nbsp;</P> <P>&nbsp;</P> Tue, 07 Jul 2015 00:02:57 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Incomplete-factorization/m-p/1000704#M18504 Kim__Kyungjoo 2015-07-07T00:02:57Z