https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95053
--- Comment #11 from Bill Seurer <seurer at linux dot vnet.ibm.com> --- /home/seurer/gcc/git/install/gcc-test/bin/gfortran -c -o module_ra_cam.fppized.o -I. -I./netcdf/include -I./inc -m64 -O0 -g3 -mcpu=power8 -Wno-deprecated-declarations -fconvert=big-endian -std=legacy module_ra_cam.fppized.f90 module_ra_cam.fppized.f90:6806:69: 6806 | asort(nxs) = 1.0_r8-(floor(cld(i,k)/cldeps)*cldeps) | 1 Error: Division by zero at (1) This error was also triggered by r11-205 but was not fixed by the patch. It compiles cleanly with r11-204 and earlier. I do not think (though cannot say for sure) that it really is a divide by zero. This is part of 521.wrf_r in the spec2017 test suite