[Bug middle-end/52890] Revision 185336 causes 10% degradation on cpu2000 benchmark 252.eon

["pthaugen at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52890

--- Comment #1 from Pat Haugen <pthaugen at gcc dot gnu.org> 2012-04-06 21:13:04 UTC ---
My main question on this is why the MEM_REF has an alignment of 8 (i.e. byte)?

Breakpoint 1, expand_assignment (to=0xfffafed2790, from=<optimized out>, 
    nontemporal=<optimized out>) at
/home/pthaugen/src/gcc/temp3/gcc/gcc/expr.c:4653
4653            store_bit_field (mem, GET_MODE_BITSIZE (mode),
(gdb) prt to
 <mem_ref 0xfffafed2790
    type <real_type 0xfffb0057b60 double DF
        size <integer_cst 0xfffafe125e0 constant 64>
        unit size <integer_cst 0xfffafe12600 constant 8>
        align 8 symtab 0 alias set 2 canonical type 0xfffaff00f18 precision 64>

    arg 0 <ssa_name 0xfffb00e1d60
        type <reference_type 0xfffb0055f28 type <record_type 0xfffb0055d30
mrViewingHitRecord>
            public unsigned DI size <integer_cst 0xfffafe125e0 64> unit size
<integer_cst 0xfffafe12600 8>
            align 64 symtab 0 alias set 6 canonical type 0xfffb0055f28>
        visited var <parm_decl 0xfffafe41980 VHR>def_stmt GIMPLE_NOP

        version 1
        ptr-info 0xfffafe9af80>
    arg 1 <integer_cst 0xfffafe19b80 type <reference_type 0xfffb0055f28>
constant 0>
    test.cc:29:44>
(gdb) pr mem
(mem:DF (reg/v/f:DI 120 [ VHR ]) [5 MEM[(struct mrViewingHitRecord
&)VHR_1(D)]+0 S8 A8])