Some results from SPECint 2000 with Kurt's 2nd Patch for C++inlining

[Andreas Jaeger]

I've used Kurt's second inline patch on GCC 3.1 CVS and run the
SPECint2000 programm 252.eon with it.

Without Kurt's patch:

$ time make bootstrap

real    39m54.271s
user    36m28.650s
sys     1m46.330s

$ time runspec --action=build -l 252.eon
real    14m4.484s
user    13m19.120s
sys     0m20.800s

Runtime:
                                     Estimated
                   Base      Base      Base   
   Benchmarks    Ref Time  Run Time   Ratio   
   252.eon           1300       459       283 
   252.eon           1300       458       284 
   252.eon           1300       459       283*

With Kurt's patch:

$ time make bootstrap

real    40m3.185s
user    36m24.260s
sys     1m45.810s

$ time runspec --action=build -l 252.eon
real    14m2.593s
user    13m6.430s
sys     0m21.010s

Runtime:
   252.eon           1300       437       297 
   252.eon           1300       437       297*
   252.eon           1300       437       297 

The compile time shows no difference but the execution is 5% faster
(459s -> 437 s).

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj