topic SIGSEGV calling mkl_dcoomm() in Intel® oneAPI Math Kernel Library https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/SIGSEGV-calling-mkl-dcoomm/m-p/1179641#M29232 <P><SPAN style="font-size: 1em;">We have small matrices we're attempting to multiply, and sometimes it crashes.&nbsp; Not every time, and when it works the results appear to be correct.&nbsp;&nbsp;</SPAN></P> <PRE style="font-size: 13.008px; color: rgb(0, 0, 0); word-wrap: break-word; white-space: pre-wrap;"># SIGSEGV (0xb) at pc=0x00007f39486428ec, pid=15934, tid=0x00007f3934a26700 </PRE> <PRE style="font-size: 13.008px; color: rgb(0, 0, 0); word-wrap: break-word; white-space: pre-wrap;">siginfo: si_signo: 11 (SIGSEGV), si_code: 2 (SEGV_ACCERR), si_addr: 0x00007f3bd40bc000 </PRE> <PRE style="font-size: 13.008px; color: rgb(0, 0, 0); word-wrap: break-word; white-space: pre-wrap;"> Problematic frame: # C [libmkl_mc3.so+0xcc08ec] mkl_dft_mc3_gather_d_d+0xcc </PRE> <P>Arguments to the function:</P> <P>A is 18x24<BR /> B is 18x28<BR /> C is 24x28</P> <P>transa: T<BR /> matdescra: "G&nbsp; F&nbsp; "<BR /> m:18&nbsp;<BR /> n:28&nbsp;<BR /> k:24&nbsp;<BR /> nnz:119&nbsp;<BR /> alpha: 1<BR /> beta: 0<BR /> ldb: 18&nbsp;<BR /> ldc: 24&nbsp;</P> <PRE style="color: rgb(0, 0, 0); word-wrap: break-word; white-space: pre-wrap;">Are we calling the function incorrectly somehow? It appears like every time it crashes the process is attempting to read 32 bytes past the end of B's allocated space. Our theory is that it "works" when that memory is part of the process address space and crashes when it's not. </PRE> Wed, 03 Jan 2018 22:42:48 GMT Elliott__Tavis 2018-01-03T22:42:48Z SIGSEGV calling mkl_dcoomm() https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/SIGSEGV-calling-mkl-dcoomm/m-p/1179641#M29232 <P><SPAN style="font-size: 1em;">We have small matrices we're attempting to multiply, and sometimes it crashes.&nbsp; Not every time, and when it works the results appear to be correct.&nbsp;&nbsp;</SPAN></P> <PRE style="font-size: 13.008px; color: rgb(0, 0, 0); word-wrap: break-word; white-space: pre-wrap;"># SIGSEGV (0xb) at pc=0x00007f39486428ec, pid=15934, tid=0x00007f3934a26700 </PRE> <PRE style="font-size: 13.008px; color: rgb(0, 0, 0); word-wrap: break-word; white-space: pre-wrap;">siginfo: si_signo: 11 (SIGSEGV), si_code: 2 (SEGV_ACCERR), si_addr: 0x00007f3bd40bc000 </PRE> <PRE style="font-size: 13.008px; color: rgb(0, 0, 0); word-wrap: break-word; white-space: pre-wrap;"> Problematic frame: # C [libmkl_mc3.so+0xcc08ec] mkl_dft_mc3_gather_d_d+0xcc </PRE> <P>Arguments to the function:</P> <P>A is 18x24<BR /> B is 18x28<BR /> C is 24x28</P> <P>transa: T<BR /> matdescra: "G&nbsp; F&nbsp; "<BR /> m:18&nbsp;<BR /> n:28&nbsp;<BR /> k:24&nbsp;<BR /> nnz:119&nbsp;<BR /> alpha: 1<BR /> beta: 0<BR /> ldb: 18&nbsp;<BR /> ldc: 24&nbsp;</P> <PRE style="color: rgb(0, 0, 0); word-wrap: break-word; white-space: pre-wrap;">Are we calling the function incorrectly somehow? It appears like every time it crashes the process is attempting to read 32 bytes past the end of B's allocated space. Our theory is that it "works" when that memory is part of the process address space and crashes when it's not. </PRE> Wed, 03 Jan 2018 22:42:48 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/SIGSEGV-calling-mkl-dcoomm/m-p/1179641#M29232 Elliott__Tavis 2018-01-03T22:42:48Z At the first glance https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/SIGSEGV-calling-mkl-dcoomm/m-p/1179642#M29233 <P>At the first glance everything look correct. Could you give the example of this problem to check/debug problem on our side with the latest versions of MKL?&nbsp;</P> <P>&nbsp;</P> Thu, 04 Jan 2018 04:07:56 GMT https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/SIGSEGV-calling-mkl-dcoomm/m-p/1179642#M29233 Gennady_F_Intel 2018-01-04T04:07:56Z