[llvm-testresults] evchen-ppc32 Power Macintosh nightly tester results

Duncan Sands baldrick at free.fr
Sat May 30 20:18:42 PDT 2009


I'm guessing that the increase in bytecode is due to outputting
bodies of "extern inline" functions rather than just declarations
for them, as we used to do.

> Bytecode:
>  singlesource/Benchmarks/CoyoteBench/fftbench: -41.78% (14704 => 20848)
>  singlesource/Benchmarks/Misc-C++/bigfib: -53.91% (28608 => 44032)
>  singlesource/Benchmarks/Misc-C++/ray: -76.32% (10336 => 18224)
>  singlesource/Benchmarks/Misc-C++/sphereflake: -141.23% (6752 => 16288)
>  singlesource/Benchmarks/Shootout-C++/ackermann: -300.00% (2720 => 10880)
>  singlesource/Benchmarks/Shootout-C++/ary: -190.58% (3568 => 10368)
>  singlesource/Benchmarks/Shootout-C++/ary2: -148.60% (4576 => 11376)
>  singlesource/Benchmarks/Shootout-C++/ary3: -209.28% (3792 => 11728)
>  singlesource/Benchmarks/Shootout-C++/except: -215.38% (3328 => 10496)
>  singlesource/Benchmarks/Shootout-C++/fibo: -252.07% (1936 => 6816)
>  singlesource/Benchmarks/Shootout-C++/hash: -73.58% (9024 => 15664)
>  singlesource/Benchmarks/Shootout-C++/hash2: -73.32% (10736 => 18608)
>  singlesource/Benchmarks/Shootout-C++/hello: -297.14% (1680 => 6672)
>  singlesource/Benchmarks/Shootout-C++/lists: -90.83% (5408 => 10320)
>  singlesource/Benchmarks/Shootout-C++/lists1: -125.76% (6336 => 14304)
>  singlesource/Benchmarks/Shootout-C++/matrix: -224.22% (3568 => 11568)
>  singlesource/Benchmarks/Shootout-C++/methcall: -144.05% (3632 => 8864)
>  singlesource/Benchmarks/Shootout-C++/nestedloop: -321.54% (2080 => 8768)
>  singlesource/Benchmarks/Shootout-C++/objinst: -135.54% (3872 => 9120)
>  singlesource/Benchmarks/Shootout-C++/random: -256.30% (1904 => 6784)
>  singlesource/Benchmarks/Shootout-C++/reversefile: -78.28% (8176 => 14576)
>  singlesource/Benchmarks/Shootout-C++/sieve: -127.21% (4880 => 11088)
>  singlesource/Benchmarks/Shootout-C++/spellcheck: -107.12% (12576 => 26048)
>  singlesource/Benchmarks/Shootout-C++/strcat: -274.19% (2480 => 9280)
>  singlesource/Benchmarks/Shootout-C++/sumcol: -331.30% (2096 => 9040)
>  singlesource/Benchmarks/Shootout-C++/wc: -313.58% (2592 => 10720)
>  singlesource/Regression/C++/ofstream_ctor: -480.46% (1392 => 8080)
>  multisource/Applications/hexxagon/hexxagon: -16.63% (40320 => 47024)
>  multisource/Benchmarks/Prolangs-C++/city/city: -51.44% (16112 => 24400)
>  multisource/Benchmarks/Prolangs-C++/deriv1/deriv1: -108.33% (5952 => 12400)
>  multisource/Benchmarks/Prolangs-C++/deriv2/deriv2: -87.61% (7360 => 13808)
>  multisource/Benchmarks/Prolangs-C++/employ/employ: -78.68% (11856 => 21184)
>  multisource/Benchmarks/Prolangs-C++/office/office: -164.86% (4144 => 10976)
>  multisource/Benchmarks/Prolangs-C++/shapes/shapes: -36.04% (10656 => 14496)
>  external/SPEC/CINT2000/252.eon/252.eon: -8.64% (544944 => 592048)

Ciao,

Duncan.



More information about the llvm-testresults mailing list