[LLVMbugs] [Bug 7330] New: miscompile of SPEC CPU2006's 450.soplex

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Wed Jun 9 07:06:47 CDT 2010


http://llvm.org/bugs/show_bug.cgi?id=7330

           Summary: miscompile of SPEC CPU2006's 450.soplex
           Product: dragonegg
           Version: 2.7
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: New Bugs
        AssignedTo: baldrick at free.fr
        ReportedBy: kenneth.hoste at ugent.be
                CC: llvmbugs at cs.uiuc.edu


Compiling the 450.soplex SPEC CPU2006 benchmark on Linux/x86-64 using
dragonegg-2.7 (with gcc-4.5.0) makes the SPEC framework complain about a
miscompile:

  Running (#1) 450.soplex test base x86-64_llvm-2.7_dragonegg_O2 default
/local/SPEC_CPU2006_v1.1_src/bin/specinvoke -d
/local/SPEC_CPU2006_v1.1_src/benchspec/CPU2006/450.soplex/run/run_base_test_x86-64_llvm-2.7_dragonegg_O2.0000
-e speccmds.err -o speccmds.stdout -f speccmds.cmd -C
/local/SPEC_CPU2006_v1.1_src/bin/specinvoke -E -d
/local/SPEC_CPU2006_v1.1_src/benchspec/CPU2006/450.soplex/run/run_base_test_x86-64_llvm-2.7_dragonegg_O2.0000
-c 1 -e compare.err -o compare.stdout -f compare.cmd

*** Miscompare of test.mps.info; for details see
   
/local/SPEC_CPU2006_v1.1_src/benchspec/CPU2006/450.soplex/run/run_base_test_x86-64_llvm-2.7_dragonegg_O2.0000/test.mps.info.mis

*** Miscompare of test.out; for details see
   
/local/SPEC_CPU2006_v1.1_src/benchspec/CPU2006/450.soplex/run/run_base_test_x86-64_llvm-2.7_dragonegg_O2.0000/test.out.mis

The test.out output file shows this:

Finished solving (status=-6)
aborted due to iteration limit
solution value is: 6.0605730e+04

While, according to the model output provided by SPEC, it should be showing
something like: 

Finished solving (status=1, objValue=1.727911e+05)
solution value is: 1.7279107e+05

It seems like the iterative algorithm at the core of soplex is unable to
converge.

The problem arises both at -O0 and -O2.
The issue does however not occur when the same benchmark is built using
llvm-gcc-4.2 -O2.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.


More information about the LLVMbugs mailing list