[Bug c/16852] New: cc1 fails attempting to allocate too much memory

[Bug c/16852] New: cc1 fails attempting to allocate too much memory

[lucier at math dot purdue dot edu]

gcc -I../include -no-cpp-precomp -Wall -W -Wno-unused -O1 -fno-math-errno
-fschedule-insns2 -fno-trapping-math -fno-strict-aliasing -fomit-frame-pointer
-fPIC -fno-common -g -save-temps -save-temps -DHAVE_CONFIG_H -c _t-c-1.c
gcc: unrecognized option `-no-cpp-precomp'
*** malloc: vm_allocate(size=1990995968) failed (error code=3)
*** malloc[5586]: error: Can't allocate region

cc1: out of memory allocating 1990994304 bytes after a total of 0 bytes
make[1]: *** [_t-c-1.o] Error 1
make: *** [all] Error 1
[descartes:~/programs/gambc40b4-devel] lucier% gcc -v
Reading specs from /pkgs/gcc-mainline/lib/gcc/powerpc-apple-darwin7.4.0/3.5.0/specs
Configured with: ../configure --prefix=/pkgs/gcc-mainline
Thread model: posix
gcc version 3.5.0 20040730 (experimental)


The input file is at

http://www.math.purdue.edu/~lucier/GNATS/GNATS-11/_t-c-1.i.gz

-- 
           Summary: cc1 fails attempting to allocate too much memory
           Product: gcc
           Version: 3.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: lucier at math dot purdue dot edu
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: powerpc-apple-darwin7.4.0
  GCC host triplet: powerpc-apple-darwin7.4.0
GCC target triplet: powerpc-apple-darwin7.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] cc1 fails attempting to allocate too much memory

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c                           |rtl-optimization
           Keywords|                            |ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] cc1 fails attempting to allocate too much memory

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-03 06:51 -------
#0  0x9002b9bc in exit ()
#1  0x0040cbd4 in xexit (code=1) at /Users/apinski/src/devel/clean/gcc/libiberty/xexit.c:52
#2  0x004021c4 in xmalloc_failed (size=1990994304) at /Users/apinski/src/devel/clean/gcc/
libiberty/xmalloc.c:132
#3  0x00402254 in xcalloc (nelem=248874288, elsize=8) at /Users/apinski/src/devel/clean/gcc/
libiberty/xmalloc.c:161
#4  0x00343904 in global_alloc (file=0x0) at /Users/apinski/src/devel/clean/gcc/gcc/global.c:527
#5  0x001f64e8 in rest_of_compilation () at /Users/apinski/src/devel/clean/gcc/gcc/passes.c:693
#6  0x001f74a0 in execute_pass_list (pass=0x57234c) at /Users/apinski/src/devel/clean/gcc/gcc/
tree-optimize.c:453
#7  0x001f76f8 in tree_rest_of_compilation (fndecl=0x7e0380, nested_p=0 '\0') at /Users/apinski/src/
devel/clean/gcc/gcc/tree-optimize.c:556
#8  0x0001c2d8 in c_expand_body (fndecl=0x7e0380) at /Users/apinski/src/devel/clean/gcc/gcc/c-
decl.c:6377
#9  0x001e499c in cgraph_expand_function (node=0xd92080) at /Users/apinski/src/devel/clean/gcc/
gcc/cgraphunit.c:801
#10 0x001e4b64 in cgraph_assemble_pending_functions () at /Users/apinski/src/devel/clean/gcc/gcc/
cgraphunit.c:296
#11 0x001e5368 in cgraph_finalize_function (decl=0x7e0380, nested=0 '\0') at /Users/apinski/src/
devel/clean/gcc/gcc/cgraphunit.c:377
#12 0x0001c500 in c_finalize (fndecl=0x7e0380) at /Users/apinski/src/devel/clean/gcc/gcc/c-decl.c:
6247
#13 0x0001cb08 in finish_function () at /Users/apinski/src/devel/clean/gcc/gcc/c-decl.c:6349
#14 0x000047d0 in yyparse () at c-parse.y:368
#15 0x0000896c in c_parse_file () at c-parse.y:2966
#16 0x0004fd80 in c_common_parse_file (set_yydebug=1) at /Users/apinski/src/devel/clean/gcc/gcc/
c-opts.c:1090
#17 0x0008ea1c in toplev_main (argc=5693780, argv=0x41502540) at /Users/apinski/src/devel/
clean/gcc/gcc/toplev.c:981
#18 0x000020c8 in _start (argc=14, argv=0xbffff9d5, envp=0xbffffa11) at /SourceCache/Csu/
Csu-49/crt.c:268
#19 0x00001f3c in start ()
(gdb) up
#1  0x0040cbd4 in xexit (code=1) at /Users/apinski/src/devel/clean/gcc/libiberty/xexit.c:52
52        exit (code);
(gdb) 
#2  0x004021c4 in xmalloc_failed (size=1990994304) at /Users/apinski/src/devel/clean/gcc/
libiberty/xmalloc.c:132
132       xexit (1);
(gdb) 
#3  0x00402254 in xcalloc (nelem=248874288, elsize=8) at /Users/apinski/src/devel/clean/gcc/
libiberty/xmalloc.c:161
161         xmalloc_failed (nelem * elsize);
(gdb) 
#4  0x00343904 in global_alloc (file=0x0) at /Users/apinski/src/devel/clean/gcc/gcc/global.c:527
527       conflicts = xcalloc (max_allocno * allocno_row_words, sizeof (INT_TYPE));
(gdb) l
522       allocno_row_words = (max_allocno + INT_BITS - 1) / INT_BITS;
523
524       /* We used to use alloca here, but the size of what it would try to
525          allocate would occasionally cause it to exceed the stack limit and
526          cause unpredictable core dumps.  Some examples were > 2Mb in size.  */
527       conflicts = xcalloc (max_allocno * allocno_row_words, sizeof (INT_TYPE));
528
529       allocnos_live = xmalloc (allocno_row_words * sizeof (INT_TYPE));
530
531       /* If there is work to be done (at least one reg to allocate),
(gdb) p max_allocno
$1 = <unknown type>
(gdb) p allocno_row_words
$2 = <unknown type>
(gdb) p (int)allocno_row_words
$3 = 1972
(gdb) p (int)max_allocno
$4 = 126204

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] cc1 fails attempting to allocate too much memory

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-03 06:56 -------
This is failing when allocating the registers.
Seems like there are a large number of code in this function.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] cc1 fails attempting to allocate too much memory

[lucier at math dot purdue dot edu]

------- Additional Comments From lucier at math dot purdue dot edu  2004-08-03 23:34 -------
Subject: Re:  cc1 fails attempting to allocate too much memory


On Tuesday, August 3, 2004, at 01:56  AM, pinskia at gcc dot gnu dot 
org wrote:
> Seems like there are a large number of code in this function.

Well, yes, but it worked on June 19 when I filed this unrelated bug 
report:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16088

Brad



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] [3.5 Regression] cc1 fails attempting to allocate too much memory

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|cc1 fails attempting to     |[3.5 Regression] cc1 fails
                   |allocate too much memory    |attempting to allocate too
                   |                            |much memory
   Target Milestone|---                         |3.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] [3.5 Regression] cc1 fails attempting to allocate too much memory

[lucier at math dot purdue dot edu]

------- Additional Comments From lucier at math dot purdue dot edu  2004-09-01 04:13 -------
With commands like this:

 cvs update -D "July 22, 2004"

I found that the example program compiled on July 20 and didn't on July 24.

Bootstrap failed on July 22.

I'll work on this more later.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] [3.5 Regression] cc1 fails attempting to allocate too much memory

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-01 04:34 -------
Hmm, the only thing major that went into between those days were Deigo's aliasing changes (and 
some cleanups which should not have any effect on code generation at all).
Which was:
2004-07-22  Diego Novillo  <dnovillo@redhat.com>

        * tree-into-ssa.c (set_livein_block): Fix typo in comment.
        (rewrite_ssa_into_ssa): Start iterating over SSA names at 1.
        Release SSA names that have been re-renamed.
        ....



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] [3.5 Regression] cc1 fails attempting to allocate too much memory

[lucier at math dot purdue dot edu]

------- Additional Comments From lucier at math dot purdue dot edu  2004-09-01 18:12 -------
Subject: Re:  [3.5 Regression] cc1 fails attempting to allocate too much memory

It worked on July 23, and didn't work on July 24.

Brad



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852

[Bug rtl-optimization/16852] [3.5 Regression] cc1 fails attempting to allocate too much memory

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-08 19:45 -------
Fixed.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16852