[llvm-testresults] Grawp-NO-LTO-PIC nightly tester results

[Evan Cheng]

On Apr 2, 2010, at 10:39 AM, Evan Cheng wrote:

> 
> On Apr 2, 2010, at 10:25 AM, Evan Cheng wrote:
> 
>> It's i386. 176.gcc fails this way:
>> 
>> /Developer/usr/bin/llvm-gcc -DSPEC_CPU2000 -I/Volumes/Joker/echeng/llvm/projects/llvm-test/External/SPEC/CINT2000/176.gcc -I/Users/echeng/LLVM/llvm/p\
>> rojects/llvm-test/External/SPEC/CINT2000/176.gcc -I/Users/echeng/LLVM/llvm/include -I/Users/echeng/LLVM/llvm/projects/llvm-test/include -I../../../..\
>> /include -I/Users/echeng/LLVM/llvm/include -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -DNDEBUG -I /Volumes/Joker/echeng/external/speccpu2000/benchspec/CINT2\
>> 000/176.gcc/src/  -O3 -m32 -c /Volumes/Joker/echeng/external/speccpu2000/benchspec/CINT2000/176.gcc/src/real.c -o Output/real.bc -emit-llvm
>> /Volumes/Joker/echeng/external/speccpu2000/benchspec/CINT2000/176.gcc/src/real.c: In function 'asctoeg':
>> /Volumes/Joker/echeng/external/speccpu2000/benchspec/CINT2000/176.gcc/src/real.c:4762: warning: incompatible implicit declaration of built-in functio\
>> n 'strlen'
>> Assertion failed: (L->isLCSSAForm(*DT)), function runOnLoop, file /Volumes/Penguin/echeng/BUILDIT/llvmCore.roots/llvmCore~obj/src/lib/Transforms/Util\
>> s/LCSSA.cpp, line 167.
>> /Volumes/Joker/echeng/external/speccpu2000/benchspec/CINT2000/176.gcc/src/real.c: At top level:
>> /Volumes/Joker/echeng/external/speccpu2000/benchspec/CINT2000/176.gcc/src/real.c:6122: internal compiler error: Abort trap: 6
>> Please submit a full bug report,
>> with preprocessed source if appropriate.
>> See <URL:http://developer.apple.com/bugreporter> for instructions.
>> make[1]: [Output/real.bc] Error 1 (ignored)
> 
> This is a tough one. So far I have not found a way to reproduce this with opt.

It's crashing in lcssa. However, the incoming ir is already broken. It has to do with phi not dominating all uses. Bob, I suspect this might be your changes.

Evan

> 
> Evan
> 
>> 
>> I'll look at other issues.
>> 
>> Evan
>> 
>> On Apr 2, 2010, at 10:20 AM, Bob Wilson wrote:
>> 
>>> 
>>> On Apr 2, 2010, at 9:57 AM, Daniel Dunbar wrote:
>>> 
>>>> Bullet probably has never passed, I seem to recall it didn't work for all testers when enabled. Dunno about the others.
>>> 
>>> Well, it _appeared_ to pass on 3-30.  Are you saying that it just got lucky somehow?
>>> 
>>> Jakob just asked me if this tester is i386, but I don't know how to tell.  Who owns it?  It might be helpful to take a look at the logs to see how these tests are failing.
>>> 
>>>> 
>>>> - Daniel
>>>> 
>>>> On Apr 2, 2010, at 9:56 AM, Bob Wilson wrote:
>>>> 
>>>>> Nurbs is back but there are still some unresolved failures on this tester:
>>>>> 
>>>>> 176.gcc
>>>>> 403.gcc
>>>>> 483.xalancbmk
>>>>> Bullet
>>>>> 
>>>>> xalancbmk has been failing for a while, but the other 3 all passed until recently.  See the 3-30 build:
>>>>> http://smooshlab.apple.com/perf/db_nt_internal/nightlytest/6854/?full=1
>>>>> 
>>>>> Jakob had guessed earlier that the Bullet failure might have been due to PR6750 but it didn't come back when he fixed that PR, so I guess it is something different.
>>>>> 
>>>>> On Apr 2, 2010, at 9:19 AM, daniel_dunbar at apple.com wrote:
>>>>> 
>>>>>> http://smooshlab.apple.com/perf/db_nt_internal/nightlytest/6954/
>>>>>> Nickname: Grawp-NO-LTO-PIC:5
>>>>>> Name:  grawp.apple.com
>>>>>> 
>>>>>> Run: 6954, Start Time: 2010-04-02 06:21:17, End Time: 2010-04-02 09:16:15
>>>>>> Comparing To: 6919, Start Time: 2010-04-01 06:30:01, End Time: 2010-04-01 09:24:09
>>>>>> *** WARNING ***: comparison is against a different machine (Grawp-NO-LTO-PIC:4)
>>>>>> 
>>>>>> --- Changes Summary ---
>>>>>> New Test Passes: 4
>>>>>> New Test Failures: 0
>>>>>> Added Tests: 0
>>>>>> Removed Tests: 0
>>>>>> Significant Changes: 4
>>>>>> 
>>>>>> --- Tests Summary ---
>>>>>> Total Tests: 3897
>>>>>> Total Test Failures: 950
>>>>>> 
>>>>>> Total Test Failures By Type:
>>>>>> Bitcode: 2
>>>>>> CBE: 64
>>>>>> GCCAS: 2
>>>>>> JIT: 433
>>>>>> JIT codegen: 433
>>>>>> LLC: 4
>>>>>> LLC compile: 4
>>>>>> LLC-BETA: 4
>>>>>> LLC_BETA compile: 4
>>>>>> 
>>>>>> --- Changes Detail ---
>>>>>> New Test Passes:
>>>>>> Externals/Nurbs/nurbs [LLC, LLC-BETA, LLC compile, LLC_BETA compile]
>>>>>> 
>>>>>> New Test Failures:
>>>>>> 
>>>>>> Added Tests:
>>>>>> 
>>>>>> Removed Tests:
>>>>>> 
>>>>>> Significant Changes in Test Results:
>>>>>> LLC:
>>>>>> Externals/SPEC/CINT2006/471_omnetpp/471_omnetpp: 7.84% (2.0400 => 2.2000)
>>>>>> SingleSource/Benchmarks/Shootout-C++/strcat: -6.90% (0.2900 => 0.2700)
>>>>>> LLC-BETA:
>>>>>> SingleSource/Benchmarks/Shootout-C++/strcat: -6.90% (0.2900 => 0.2700)
>>>>>> CBE:
>>>>>> MultiSource/Applications/spiff/spiff: 6.07% (4.9400 => 5.2400)
>>>>>> _______________________________________________
>>>>>> llvm-testresults mailing list
>>>>>> llvm-testresults at cs.uiuc.edu
>>>>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-testresults
>>>>> 
>>>> 
>>> 
>>> 
>>> _______________________________________________
>>> llvm-testresults mailing list
>>> llvm-testresults at cs.uiuc.edu
>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-testresults
>> 
>