[llvm-testresults] llvmgcc-bwilson__llvm-gcc_PROD__i386 nightly tester results
Daniel Dunbar
daniel_dunbar at apple.com
Fri Oct 29 09:35:05 PDT 2010
On Oct 29, 2010, at 8:59 AM, Bob Wilson wrote:
>
> On Oct 29, 2010, at 8:10 AM, Daniel Dunbar wrote:
>>
>> Is this tester running in parallel?
>
> Not unless that is the default. There should not have been anything else running on the machine during the tests. I don't know what would explain the performance variations, though.
>
>>
>> If so, any chance we could get it to be a -j1 run to get less noise?
>
> Let me know if I need to do something to force -j1.
Nope, should be the default. Maybe something else was running during the first or second runs?
>
>>
>> Thanks again for setting this up!!!
>
> Sure. I didn't receive a copy of the test results. I don't see it in the llvm-testresults archive, and my log shows that "lnt submit" still sent mail directly to you.
That is correct, I just replied all to get it onto the list. Should be going to llvm-testresults in the future (you can resubmit run to verify).
- Daniel
>
>>
>> - Daniel
>>
>> On Oct 29, 2010, at 12:44 AM, llvm-testresults at cs.uiuc.edu wrote:
>>
>>> llvmgcc-bwilson__llvm-gcc_PROD__i386 nightly tester results
>>>
>>> URL http://llvm.org/perf/db_default/simple/nts/2/
>>> Nickname llvmgcc-bwilson__llvm-gcc_PROD__i386:1
>>> Name curlew.apple.com
>>> Run ID Order Start Time End Time
>>> Current 2 117642 2010-10-29 07:01:13 2010-10-29 07:46:01
>>> Previous 1 117603 2010-10-28 21:21:09 2010-10-28 22:06:06
>>>
>>> Tests Summary
>>>
>>> Status Group #
>>> Performance Regressions 9
>>> Performance Improvements 16
>>> Existing Failures 1
>>> Unchanged Tests 915
>>> Total Tests 941
>>>
>>> Changes Detail
>>>
>>> Performance Regressions Δ Previous Current σ
>>> MultiSource/Benchmarks/McCat/08-main/main.exec 100.00% 0.0600 0.1200 0.0000
>>> SingleSource/Benchmarks/Shootout-C++/methcall.exec 56.93% 5.4100 8.4900 0.0000
>>> SingleSource/Benchmarks/Shootout/methcall.exec 51.45% 5.1900 7.8600 0.0000
>>> MultiSource/Benchmarks/Trimaran/enc-pc1/enc-pc1.exec 25.27% 0.9100 1.1400 0.0000
>>> SingleSource/Benchmarks/BenchmarkGame/fannkuch.exec 16.77% 3.1600 3.6900 0.0000
>>> SingleSource/Benchmarks/Misc/fbench.exec 8.81% 1.9300 2.1000 0.0000
>>> SingleSource/Benchmarks/Shootout/lists.exec 3.99% 6.2600 6.5100 0.0000
>>> MultiSource/Applications/JM/lencod/lencod.exec 3.79% 12.4100 12.8800 0.0000
>>> External/SPEC/CINT2000/164_gzip/164_gzip.exec 1.87% 16.0200 16.3200 0.0000
>>>
>>> Performance Improvements Δ Previous Current σ
>>> SingleSource/Benchmarks/Dhrystone/fldry.exec -10.90% 4.2200 3.7600 0.0000
>>> SingleSource/Benchmarks/Dhrystone/dry.exec -10.10% 3.8600 3.4700 0.0000
>>> External/SPEC/CFP2000/177_mesa/177_mesa.exec -4.88% 3.6900 3.5100 0.0000
>>> SingleSource/Benchmarks/Misc/oourafft.exec -2.50% 7.5900 7.4000 0.0000
>>> SingleSource/Benchmarks/Shootout/objinst.exec -1.98% 5.0400 4.9400 0.0000
>>> External/SPEC/CINT95/132_ijpeg/132_ijpeg.compile -1.64% 4.8700 4.7900 0.0000
>>> SingleSource/Benchmarks/Misc/salsa20.exec -1.37% 10.2100 10.0700 0.0000
>>> MultiSource/Benchmarks/NPB-serial/is/is.exec -1.30% 10.0300 9.9000 0.0000
>>> External/SPEC/CINT95/134_perl/134_perl.compile -1.22% 6.5400 6.4600 0.0000
>>> MultiSource/Applications/lua/lua.exec -1.20% 30.8700 30.5000 0.0000
>>> MultiSource/Benchmarks/ASCI_Purple/SMG2000/smg2000.compile -1.16% 7.7400 7.6500 0.0000
>>> MultiSource/Benchmarks/Prolangs-C/TimberWolfMC/timberwolfmc.compile -1.12% 8.0700 7.9800 0.0000
>>> External/SPEC/CINT2006/456_hmmer/456_hmmer.compile -1.06% 7.5400 7.4600 0.0000
>>> SingleSource/Benchmarks/Shootout-C++/objinst.exec -1.04% 10.5700 10.4600 0.0000
>>> MultiSource/Benchmarks/MiBench/consumer-typeset/consumer-typeset.compile -1.02% 13.7400 13.6000 0.0000
>>> External/SPEC/CINT95/126_gcc/126_gcc.compile -1.00% 29.9500 29.6500 0.0000
>>>
>>> Existing Failures
>>> External/SPEC/CINT2006/483_xalancbmk/483_xalancbmk.exec
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-testresults/attachments/20101029/13f4ad52/attachment.html>
More information about the llvm-testresults
mailing list