https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Problem-with-mkl-dcsrmv/m-p/778569#M1257 Hello <BR /><BR />In the attached example (creating dense matrix sv - &gt; transforming into sparse matrix v -&gt; multiplying with vector u) there is a problem with the mkl_dcsrmv routine. I.e. the value stored in the product vector y is wrong.<BR />I would be very happy if someone could tell me what I do wrong.<BR /><BR />Heres the code:<BR /><A onclick="ndownload('http://software.intel.com/file/40302')">test.cpp</A><BR /><BR />Kind regards<BR />Wolfram<BR /> Fri, 09 Dec 2011 17:43:37 GMT Wolfram_Ruehaak 2011-12-09T17:43:37Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Problem-with-mkl-dcsrmv/m-p/778569#M1257 Hello <BR /><BR />In the attached example (creating dense matrix sv - &gt; transforming into sparse matrix v -&gt; multiplying with vector u) there is a problem with the mkl_dcsrmv routine. I.e. the value stored in the product vector y is wrong.<BR />I would be very happy if someone could tell me what I do wrong.<BR /><BR />Heres the code:<BR /><A onclick="ndownload('http://software.intel.com/file/40302')">test.cpp</A><BR /><BR />Kind regards<BR />Wolfram<BR /> Fri, 09 Dec 2011 17:43:37 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Problem-with-mkl-dcsrmv/m-p/778569#M1257 Wolfram_Ruehaak 2011-12-09T17:43:37Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Problem-with-mkl-dcsrmv/m-p/778570#M1258 I found the reason: as the solution which is solved is y = alpha*A*x + beta*y (and not just y=A*x) it is necessary to initialize y. In the given case all y values hasve to be set to 0 before calling to mkl_dcsrmv.<BR /><BR />Kind regards<BR />Wolfram Fri, 09 Dec 2011 18:37:26 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Problem-with-mkl-dcsrmv/m-p/778570#M1258 Wolfram_Ruehaak 2011-12-09T18:37:26Z