https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Difference-of-calculcation-result-between-i7-and-xeon-cpu/m-p/1015638#M19471 <P>Hi.</P> <P>I make the numerical calculation program and run two computers. (One is used i7 cpu, others Xeon cpu.)</P> <P>The numerical result is slightly different. (input data is same).</P> <P>My program is very sensitive to numerical difference, so i confused which one is correct.</P> <P>Why the differece is occured?</P> <P>&nbsp;</P> <P>the program is complied same computer under environment</P> <P>Complier : MS VS2010 C++ Complier</P> <P><SPAN style="font-size: 13.008px; line-height: 19.512px;">MKL library Version : 11.1.4.237</SPAN></P> <P><SPAN style="line-height: 19.512px;">Linked library : mkl_core.lib, mkl_intel_lp64.lib, mkl_intel_thread.lib, libiomp5md.lib</SPAN></P> <P>&nbsp;</P> <P><SPAN style="line-height: 19.512px;">the program mainly used "DGEMM",&nbsp;</SPAN><SPAN style="font-size: 13.008px; line-height: 19.512px;">"PARDISO" functions for solving linear equation directly(not iterative method).</SPAN></P> <P><SPAN style="line-height: 19.512px;">And the number of thread set 1.</SPAN></P> Tue, 13 Oct 2015 06:21:04 GMT DongHee_Bae 2015-10-13T06:21:04Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Difference-of-calculcation-result-between-i7-and-xeon-cpu/m-p/1015638#M19471 <P>Hi.</P> <P>I make the numerical calculation program and run two computers. (One is used i7 cpu, others Xeon cpu.)</P> <P>The numerical result is slightly different. (input data is same).</P> <P>My program is very sensitive to numerical difference, so i confused which one is correct.</P> <P>Why the differece is occured?</P> <P>&nbsp;</P> <P>the program is complied same computer under environment</P> <P>Complier : MS VS2010 C++ Complier</P> <P><SPAN style="font-size: 13.008px; line-height: 19.512px;">MKL library Version : 11.1.4.237</SPAN></P> <P><SPAN style="line-height: 19.512px;">Linked library : mkl_core.lib, mkl_intel_lp64.lib, mkl_intel_thread.lib, libiomp5md.lib</SPAN></P> <P>&nbsp;</P> <P><SPAN style="line-height: 19.512px;">the program mainly used "DGEMM",&nbsp;</SPAN><SPAN style="font-size: 13.008px; line-height: 19.512px;">"PARDISO" functions for solving linear equation directly(not iterative method).</SPAN></P> <P><SPAN style="line-height: 19.512px;">And the number of thread set 1.</SPAN></P> Tue, 13 Oct 2015 06:21:04 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Difference-of-calculcation-result-between-i7-and-xeon-cpu/m-p/1015638#M19471 DongHee_Bae 2015-10-13T06:21:04Z https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Difference-of-calculcation-result-between-i7-and-xeon-cpu/m-p/1015639#M19472 <P>yes, this is expected behavior with floating point computations. since &nbsp;version 11.1, MKL addresses such issues. Please refer to the documentation about CNR -&nbsp;https://software.intel.com/en-us/node/528577&nbsp;</P> <P>&nbsp;</P> Tue, 13 Oct 2015 09:08:06 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/Difference-of-calculcation-result-between-i7-and-xeon-cpu/m-p/1015639#M19472 Gennady_F_Intel 2015-10-13T09:08:06Z