This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

CVS performance regression over 2.7.2 on SPECint95 on sparc


Hi!

SPECint95 benchmarks compiled with gcc from updated from CVS yesterday
evening (PST) performe worse than when compiled with gcc-2.7.2.3
on sun-sparc-solaris2.7 on a 333MHz Ultra5 machine with 256MB RAM. 

The only flags used are:
-O3 -mv8   for 2.7.2.3
and
-O3 -mcpu=v8 for the CVS version

here are 

the results:

                     SPEC Benchmark CINT95 Summary


                           for gcc-CVS           for gcc-CVS
                  Base     Base      Base        Base      Base  
Benchmarks    Ref Time Run Time     Ratio    Run Time     Ratio  
------------  -------- --------  --------    --------  --------  
099.go            4600   449        10.2 *        394     11.7 * 
124.m88ksim       1900   234         8.12*        244      7.80* 
126.gcc           1700    13.7         --X        164        --X 
129.compress      1800   352         5.11*        224      8.05* 
130.li            1900   254         7.48*        256      7.42* 
132.ijpeg         2400   252         9.53*        246      9.77* 
134.perl          1900   207         9.17*        202      9.40* 
147.vortex        2700     2.40        --X        323      8.36* 

for both for gcc I get comparison errors. 

vortex compiled with gcc-CVS segfaults. 

Does anybody else see this? 
Any suggestions? 


Thanks.

	--dan

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]