[llvm-testresults] Grawp nightly tester results

[Evan Cheng]

The 179.art regression is very significant. I've tracked it down to  
84288 and 84292. Victor, please take a look asap.

Evan

On Oct 17, 2009, at 9:46 AM, Daniel Dunbar wrote:

> The LLC losses here look like they are for the worse, overall.
>
> - Daniel
>
> On Sat, Oct 17, 2009 at 4:02 AM,  <daniel_dunbar at apple.com> wrote:
>> http://smooshlab.apple.com/perf.cgi/db_nt_internal/nightlytest//2416
>> Name:  grawp.apple.com
>> Nickname: Grawp:4
>> Buildstatus: OK
>>
>>
>> LLC:
>>  Externals/SPEC/CFP2000/179_art/179_art: 27.47% (0.9100 => 1.1600)
>>  Externals/SPEC/CINT2006/445_gobmk/445_gobmk: 8.70% (0.2300 =>  
>> 0.2500)
>>  MultiSource/Applications/hexxagon/hexxagon: 5.61% (8.9200 => 9.4200)
>>  MultiSource/Benchmarks/Ptrdist/yacr2/yacr2: 15.91% (0.8800 =>  
>> 1.0200)
>>  SingleSource/Benchmarks/CoyoteBench/huffbench: -7.79% (19.1200 =>  
>> 17.6300)
>>  SingleSource/Benchmarks/McGill/chomp: 13.83% (0.9400 => 1.0700)
>>  SingleSource/Benchmarks/Misc-C++/stepanov_container: 6.19% (4.5200  
>> => 4.8000)
>>  SingleSource/Benchmarks/Misc/himenobmtxpa: -5.67% (3.3500 => 3.1600)
>>  SingleSource/Benchmarks/Shootout/sieve: 11.79% (5.0900 => 5.6900)
>> LLC-BETA compile:
>>  Externals/SPEC/CINT2000/300_twolf/300_twolf: -5.85% (2.5536 =>  
>> 2.4042)
>> _______________________________________________
>> llvm-testresults mailing list
>> llvm-testresults at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-testresults
>>