This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: ICE on stage 2
- From: Kelley Cook <KelleyCook at comcast dot net>
- To: GCC Mailing List <gcc at gcc dot gnu dot org>, Jan Hubicka <jh at suse dot cz>
- Cc: Bobby McNulty Junior <bmj2001 at bellsouth dot net>,Christian =?UNKNOWN?Q?J=F6nsson?= <c dot christian dot joensson at telia dot com>,Jan Hubicka <jh at suse dot cz>
- Date: Thu, 16 Jan 2003 20:26:40 -0500
- Subject: Re: ICE on stage 2
On Thu, 16 Jan 2003 23:30:21 +0100, Jan Hubicka wrote:
>> [This followup was posted to gmane.comp.gcc.devel and a copy
>> was sent to the cited author.]
>>
>> In article <MPG.188e50ce649718d5989682@news.gmane.org>,
>> KelleyCook=Wuw85uim5zDR7s880joybQ@public.gmane.org says...
>> > In article <20030114111732.GA7036=qV6VTPpQhdfYtjvyW6yDsg@public.gmane.org>,
>> > c.christian.joensson=zq6IREYz3ykAvxtiuMwx3w@public.gmane.org says...
>> > > On Fri, Jan 10, 2003 at 08:11:09PM -0600, Bobby McNulty Junior wrote:
>> > > > It keeps happening since yesterday. GCC 3.4 (main trunk) keeps ICEing on me.
>> > > > I've compiled it using GCC-3.2 and -g -O2
>> > > > I686-pc-cygwin, newlib.
>> > >
>> > > <snip>
>> > >
>> > > This happens to me too...
>> >
>> > If you check message
>> >
>> > http://gcc.gnu.org/ml/gcc-patches/2003-01/msg00870.html
>> >
>> > I determined that Cygwin blew up after the Itanium 2
>> > scheduler merge:
>> >
>> > http://gcc.gnu.org/ml/gcc-patches/2003-01/msg00608.html
>> >
>>
>> Sorry apparently I screwed up, it was the patch previous to
>> that one which is causing the bootstrap failures under
>> cygwin. It is by Jan Hubicka:
>>
>> http://gcc.gnu.org/ml/gcc-patches/2003-01/msg00614.html
>>
>> I have verified that reversing it on the current mainline
>> allows a complete bootstrap on i686-pc-cygwin.
>
>I've commited fix for that patch about day or two later.
>Are you still having same problem?
Yes, just checked
>Can you just throw it into debuger
>on what insn it segfaults?
Unfortunately, I think it may be a garbage collection problem. It's
segfaulting in mirror_conflicts with as far as my limited gcc knowledge
is telling me rtx x pointing to a non-insn.
Normally we get
icc gen_sse_prologue_save insn_current_length insn_variable_length_p
insn_defaul
t_length bypass_p insn_default_latency {GC 11671k -> 4207k} {GC 13363k
-> 3606k}
{GC 6783k -> 3706k} {GC 8396k -> 3778k} {GC 5381k -> 3930k} {GC 5809k
-> 3852k}
and it crashes.
However if I run cc1 with --params ggc-min-heapsize=256000 it does
compile with a peak memory usage of ~150MB.
Execution times (seconds)
cfg construction : 0.63 ( 1%) usr 0.09 ( 1%) sys 0.75 ( 1%)
wall
cfg cleanup : 3.72 ( 6%) usr 0.09 ( 1%) sys 3.86 ( 5%)
wall
trivially dead code : 1.16 ( 2%) usr 0.00 ( 0%) sys 1.23 ( 2%)
wall
life analysis : 1.98 ( 3%) usr 0.00 ( 0%) sys 2.03 ( 2%)
wall
life info update : 1.28 ( 2%) usr 0.00 ( 0%) sys 1.35 ( 2%)
wall
preprocessing : 0.27 (0%) usr 0.53 ( 3%) sys 0.94 ( 1%)
wall
lexical analysis : 0.61 ( 1%) usr 0.25 ( 1%) sys 0.86 ( 1%)
wall
parser : 0.72 ( 1%) usr 0.17 ( 1%) sys 0.95 ( 1%)
wall
expand : 0.44 ( 1%) usr 0.03 (0%) sys 0.56 ( 1%)
wall
varconst : 0.03 (0%) usr 0.02 (0%) sys 0.03 (0%)
wall
integration : 0.08 (0%) usr 0.00 ( 0%) sys 0.08 (0%)
wall
jump : 1.18 ( 2%) usr 0.09 ( 1%) sys 1.50 ( 2%)
wall
CSE : 2.30 ( 4%) usr 0.00 ( 0%) sys 2.39 ( 3%)
wall
global CSE : 31.59 (52%) usr 15.52 (91%) sys 50.06 (61%)
wall
loop analysis : 0.08 (0%) usr 0.00 ( 0%) sys 0.08 (0%)
wall
bypass jumps : 1.55 ( 3%) usr 0.03 (0%) sys 1.58 ( 2%)
wall
CSE 2 : 1.22 ( 2%) usr 0.00 ( 0%) sys 1.25 ( 2%)
wall
branch prediction : 1.94 ( 3%) usr 0.00 ( 0%) sys 2.06 ( 3%)
wall
flow analysis : 0.17 (0%) usr 0.00 ( 0%) sys 0.17 (0%)
wall
combiner : 1.28 ( 2%) usr 0.00 ( 0%) sys 1.28 ( 2%)
wall
if-conversion : 0.27 (0%) usr 0.02 (0%) sys 0.28 (0%)
wall
regmove : 0.28 (0%) usr 0.00 ( 0%) sys 0.28 (0%)
wall
mode switching : 0.31 ( 1%) usr 0.00 ( 0%) sys 0.33 (0%)
wall
local alloc : 0.70 ( 1%) usr 0.00 ( 0%) sys 0.72 ( 1%)
wall
global alloc : 1.13 ( 2%) usr 0.01 (0%) sys 1.16 ( 1%)
wall
reload CSE regs : 2.53 ( 4%) usr 0.02 (0%) sys 2.70 ( 3%)
wall
flow 2 : 0.23 (0%) usr 0.02 (0%) sys 0.25 (0%)
wall
if-conversion 2 : 0.11 (0%) usr 0.00 ( 0%) sys 0.12 (0%)
wall
peephole 2 : 0.22 (0%) usr 0.02 (0%) sys 0.23 (0%)
wall
rename registers : 0.47 ( 1%) usr 0.00 ( 0%) sys 0.48 ( 1%)
wall
scheduling 2 : 0.76 ( 1%) usr 0.02 (0%) sys 0.78 ( 1%)
wall
reorder blocks : 0.09 (0%) usr 0.00 ( 0%) sys 0.11 (0%)
wall
shorten branches : 0.22 (0%) usr 0.00 ( 0%) sys 0.22 (0%)
wall
final : 0.44 ( 1%) usr 0.06 (0%) sys 0.56 ( 1%)
wall
symout : 0.00 ( 0%) usr 0.02 (0%) sys 0.05 (0%)
wall
rest of compilation : 0.58 ( 1%) usr 0.00 ( 0%) sys 0.69 ( 1%)
wall
TOTAL : 60.59 17.02 82.05
Anything else you can suggest for me to examine for you?