[Bug optimization/9895] [SH] GCC unable to retain array values in registers

[Bug optimization/9895] [SH] GCC unable to retain array values in registers

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895


pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |SUSPENDED


------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-22 00:52 -------
The problem is that GCC likes to throw things to the stack too soon, other bugs like this have been 
suspended so suspending this one too.

[Bug optimization/9895] [SH] GCC unable to retain array values in registers

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2003-12-02 18:27 -------
Moving target to tree-ssa as SRA could be expanded to work with arrays.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|SUSPENDED                   |NEW
   Last reconfirmed|0000-00-00 00:00:00         |2003-12-02 18:27:23
               date|                            |
   Target Milestone|---                         |tree-ssa


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895

[Bug optimization/9895] [SH] GCC unable to retain array values in registers

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895

[Bug tree-optimization/9895] [SH] GCC unable to retain array values in registers

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.5.0                       |---


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895

[Bug tree-optimization/9895] [SH] GCC unable to retain array values in registers

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-01-13 19:06 -------
Now the problem is two fold, we don't do SRA after loop unrolling and we have a problem in that the 
invariant not being updated for an ADDR_EXPR (both of these have been filed).

See PR 18754 and PR 18755.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |18754, 18755
   Last reconfirmed|2004-08-22 07:41:43         |2005-01-13 19:06:18
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895

[Bug tree-optimization/9895] [SH] GCC unable to retain array values in registers

[pinskia at gcc dot gnu dot org]

-- 
Bug 9895 depends on bug 18755, which changed state.

Bug 18755 Summary: invariant/constant not recomputed when ADDR_EXPR changed
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18755

           What    |Old Value                   |New Value
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895

[Bug tree-optimization/9895] GCC unable to retain array values in registers

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-07 21:22 -------
For this one, GCC is over estimating the size of the loop.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[SH] GCC unable to retain   |GCC unable to retain array
                   |array values in registers   |values in registers


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895

[Bug tree-optimization/9895] GCC unable to retain array values in registers

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
RTL opts now manage to clean this up enough (-O2 -funroll-loops):

integer_calc_one:
.LFB0:
        .cfi_startproc
        pushl   %edi
        .cfi_def_cfa_offset 8
        .cfi_offset 7, -8
        pushl   %esi
        .cfi_def_cfa_offset 12
        .cfi_offset 6, -12
        pushl   %ebx
        .cfi_def_cfa_offset 16
        .cfi_offset 3, -16
        subl    $32, %esp
        .cfi_def_cfa_offset 48
        movl    48(%esp), %ebx
        call    rand
        movl    %eax, %esi
        movl    %ebx, %eax
        sarl    $31, %eax
        shrl    $24, %eax
        leal    (%ebx,%eax), %ecx
        movzbl  %cl, %ecx
        subl    %eax, %ecx
        addl    $2, %ecx
        leal    (%esi,%ecx), %ebx
        addl    28(%esp), %esi
        movl    %ebx, %eax
        cltd
        idivl   %ecx
        movl    %ecx, %edx
        movl    %eax, %edi
        subl    %ecx, %edi
        addl    %esi, %ecx
        imull   %edi, %edx
        addl    %ebx, %ecx
        addl    %ecx, %eax
        addl    %edi, %eax
        addl    %edx, %eax
        addl    24(%esp), %eax
        addl    $32, %esp
        .cfi_def_cfa_offset 16
        .cfi_def_cfa_offset 16
        popl    %ebx
        .cfi_restore 3
        .cfi_def_cfa_offset 12
        popl    %esi
        .cfi_restore 6
        .cfi_def_cfa_offset 8
        popl    %edi
        .cfi_restore 7
        .cfi_def_cfa_offset 4
        ret

[Bug tree-optimization/9895] GCC unable to retain array values in registers

[rguenth at gcc dot gnu dot org]

------- Comment #9 from rguenth at gcc dot gnu dot org  2008-04-27 17:21 -------
Which is also the reason for early unrolling not helping.  But at -O3 the
tree loop optimizer unrolls the loop after LIM/PRE "fixed" the size estimate.

But it's too late to clean up at this point (no strong memory optimizers left).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9895