--Boundary_(ID_2q/zAFgm2PvG3TYFwZddEg)
Content-type: TEXT/PLAIN; charset=US-ASCII
Content-transfer-encoding: 7BIT
Hello,
After the release of mandrake 6.1 which is supposed to include
pgcc-1.1.2 as the default compiler I was curious about performance of
current compilers, I know benchmarks are stupid things that doesn't
mean anything, but anyway I decide to compile gcc-2.96 (CVS 23/9/99)
pgcc-2.95.1 and gcc-2.95.1 and test them with nbench-byte-2.1 I use
-O6 -fomit-frame-pointer -fno-exceptions -fexpensive-optimizations
-fforce-mem -fforce-addr -funroll-loops -march=pentiumpro, although some
of the compilers doesn't support -O6 or some of the options are redundant.
The resulting binaries size were:
nbench-2.95 58196
nbench-2.96 59112
nbench-pgcc 59352
I test them on a PII 450 with 512M as I already said I know benchmarks are
not reliable ways to test compilers, but I was surprise by the performance of
pgcc-2.95.1, it is supposed to be the faster of them, although 2.96 includes
the new ia32 backend it is still in a beta stage, but pgcc-2.95.1 is based on
2.95.1 but it seems slower than it??
- german
<>-------------------------------------+-----------------------------------<>
One O.S. to rule them all, | German Gomez Garcia
One O.S. to find them. | german@pinon.ccu.uniovi.es
One O.S. to bring them all |
and in the darkness bind them. | "Wur Qanar Wur Stilor Wur Kas"
<>-------------------------------------+-----------------------------------<>
--Boundary_(ID_2q/zAFgm2PvG3TYFwZddEg)
Content-id: <Pine.HPP.3.91.990924103428.7660B@pinon.ccu.uniovi.es>
Content-type: TEXT/PLAIN; name=offtopic; charset=US-ASCII
Content-description:
Content-disposition: attachment; filename=offtopic
Content-transfer-encoding: 7BIT
TEST : Iterations/sec. : Old Index : New Index
: : Pentium 90* : AMD K6/233*
--------------------:------------------:-------------:------------
NUMERIC SORT : 196.58 : 5.04 : 1.66
STRING SORT : 18.053 : 8.07 : 1.25
BITFIELD : 5.8898e+07 : 10.10 : 2.11
FP EMULATION : 11.878 : 5.70 : 1.32
FOURIER : 3446 : 3.92 : 2.20
ASSIGNMENT : 2.5594 : 9.74 : 2.53
IDEA : 414.01 : 6.33 : 1.88
HUFFMAN : 198.17 : 5.50 : 1.75
NEURAL NET : 3.8831 : 6.24 : 2.62
LU DECOMPOSITION : 113.03 : 5.86 : 4.23
C compiler : gcc version 2.7.2.3
MEMORY INDEX : 1.881
INTEGER INDEX : 1.637
FLOATING-POINT INDEX: 2.901
--------------------:------------------:-------------:------------
NUMERIC SORT : 266.36 : 6.83 : 2.24
STRING SORT : 19.617 : 8.77 : 1.36
BITFIELD : 6.3688e+07 : 10.92 : 2.28
FP EMULATION : 13.2 : 6.33 : 1.46
FOURIER : 4921.4 : 5.60 : 3.14
ASSIGNMENT : 3.1238 : 11.89 : 3.08
IDEA : 581.4 : 8.89 : 2.64
HUFFMAN : 224.64 : 6.23 : 1.99
NEURAL NET : 9149.4 : 14697.77 : 6182.42
LU DECOMPOSITION : 184.8 : 9.57 : 6.91
C compiler : gcc driver version 2.96 executing gcc version 2.95.1
MEMORY INDEX : 2.121
INTEGER INDEX : 2.037
FLOATING-POINT INDEX: 51.218
--------------------:------------------:-------------:------------
NUMERIC SORT : 265.96 : 6.82 : 2.24
STRING SORT : 19.592 : 8.75 : 1.36
BITFIELD : 6.2761e+07 : 10.77 : 2.25
FP EMULATION : 17.729 : 8.51 : 1.96
FOURIER : 4872 : 5.54 : 3.11
ASSIGNMENT : 3.4182 : 13.01 : 3.37
IDEA : 561.15 : 8.58 : 2.55
HUFFMAN : 210.49 : 5.84 : 1.86
NEURAL NET : 5.6267 : 9.04 : 3.80
LU DECOMPOSITION : 189.56 : 9.82 : 7.09
C compiler : gcc version 2.96
MEMORY INDEX : 2.174
INTEGER INDEX : 2.138
FLOATING-POINT INDEX: 4.378
--------------------:------------------:-------------:------------
NUMERIC SORT : 264.08 : 6.77 : 2.22
STRING SORT : 19.444 : 8.69 : 1.34
BITFIELD : 6.6431e+07 : 11.40 : 2.38
FP EMULATION : 13.126 : 6.30 : 1.45
FOURIER : 4528.1 : 5.15 : 2.89
ASSIGNMENT : 3.0558 : 11.63 : 3.02
IDEA : 579.77 : 8.87 : 2.63
HUFFMAN : 221.4 : 6.14 : 1.96
NEURAL NET : 3.9683 : 6.37 : 2.68
LU DECOMPOSITION : 184.64 : 9.57 : 6.91
C compiler : gcc driver version 2.96 executing gcc version pgcc-2.95.1
MEMORY INDEX : 2.129
INTEGER INDEX : 2.021
FLOATING-POINT INDEX: 3.770
--Boundary_(ID_2q/zAFgm2PvG3TYFwZddEg)--
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/