https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Catasprophic-compiler-failure/m-p/1068863#M6957 Effect of pre-standard simd directive and openmp4 may not be always identical. Ideally compiler should warn about possible conflict. Tue, 17 Jan 2017 21:43:23 GMT TimP 2017-01-17T21:43:23Z https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Catasprophic-compiler-failure/m-p/1068861#M6955 <P>Hello. I've encountered an a compiler failure when trying to compile a file with -02 and above. This happens on both 2016 and 2017 versions on Linux. Seems to happen after an inclusion of a parallel openmp simd region. Is threre a way to make the compiler spit out some sort of a report or get any additional insight as to why?</P> <P>Cannot publish the source code.&nbsp;</P> Fri, 13 Jan 2017 03:34:43 GMT https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Catasprophic-compiler-failure/m-p/1068861#M6955 Paulius_V_ 2017-01-13T03:34:43Z https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Catasprophic-compiler-failure/m-p/1068862#M6956 <P>Found what was causing it:</P> <PRE class="brush:fortran;">#IFDEF A !$omp parallel do simd default(shared) schedule(static) #IFDEF B !$omp simd #ENDIF #ENDIF DO I=AA,BB ... ENDDO</PRE> <P>taking out the simd out of 'do simd' fixed it. Is there any difference in inlining the simd clause as opposed to having it on the next directive?</P> Fri, 13 Jan 2017 03:55:00 GMT https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Catasprophic-compiler-failure/m-p/1068862#M6956 Paulius_V_ 2017-01-13T03:55:00Z https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Catasprophic-compiler-failure/m-p/1068863#M6957 Effect of pre-standard simd directive and openmp4 may not be always identical. Ideally compiler should warn about possible conflict. Tue, 17 Jan 2017 21:43:23 GMT https://community.intel.com/t5/Intel-Moderncode-for-Parallel/Catasprophic-compiler-failure/m-p/1068863#M6957 TimP 2017-01-17T21:43:23Z