http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58673
--- Comment #1 from Pat Haugen <pthaugen at gcc dot gnu.org> --- Created attachment 30972 --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=30972&action=edit testcase Another similar example, but this one is on a store to memory instead of a load. Reduced from CPU2006 benchmark 435.gromacs. [pthaugen@igoo p8_spec_errors]$ /home/pthaugen/install/gcc/trunk/bin/gcc -c -m64 -O3 -funroll-loops -mcpu=power8 do_gct.c ... do_gct.c:208:1: error: could not split insn } ^ (insn 90 93 94 (set (mem/c:TI (plus:DI (reg:DI 6 6) (const_int 368 [0x170])) [2 leg+0 S16 A128]) (reg:TI 10 10 [423])) do_gct.c:93 508 {*movti_ppc64} (expr_list:REG_DEAD (reg:DI 6 6) (expr_list:REG_DEAD (reg:TI 10 10 [423]) (nil)))) do_gct.c:208:1: internal compiler error: in final_scan_insn, at final.c:2949 0x105b23c3 _fatal_insn(char const*, rtx_def const*, char const*, int, char const*) /home/pthaugen/src/gcc/trunk/gcc/gcc/rtl-error.c:109 0x1034ecfb final_scan_insn(rtx_def*, _IO_FILE*, int, int, int*) /home/pthaugen/src/gcc/trunk/gcc/gcc/final.c:2949 0x1034f1e7 final(rtx_def*, _IO_FILE*, int) /home/pthaugen/src/gcc/trunk/gcc/gcc/final.c:2019 0x1034f4df rest_of_handle_final /home/pthaugen/src/gcc/trunk/gcc/gcc/final.c:4424 0x1034f4df execute /home/pthaugen/src/gcc/trunk/gcc/gcc/final.c:4499 Please submit a full bug report,