You were right with largeExplorers benchmark. Tested with 3.8.6665-full.
1) Intel Pentium 4 Mobile - 1.8GHz (WIn)
arithmaticLoopBenchmark 2179 compiled 158 ratio: 13.791
bytecodeBenchmark 2951 compiled 622 ratio: 4.744
sendBenchmark 3402 compiled 2281 ratio: 1.491
doLoopsBenchmark 1920 compiled 3660 ratio: 0.525
largeExplorers 1913 compiled 2013 ratio: 0.950
compilerBenchmark 1855 compiled 1448 ratio: 1.281
Cumulative Time 10830.891 compiled 4739.508 ratio 2.285
2) AMD Semprom 3100+ 1.8GHz (Win)
arithmaticLoopBenchmark 1647 compiled 117 ratio: 14.077
bytecodeBenchmark 2746 compiled 635 ratio: 4.324
sendBenchmark 2097 compiled 892 ratio: 2.351
doLoopsBenchmark 1265 compiled 895 ratio: 1.413
largeExplorers 996 compiled 721 ratio: 1.381
compilerBenchmark 991 compiled 1000 ratio: 0.991
Cumulative Time 6526.543 compiled 2119.418 ratio 3.079
Man, I was born for statistics ...:)
Jakub
On 11/16/06,
[hidden email] <
[hidden email]> wrote:
>
> > One more question. How safe is it to stick with the exupery vm for
> > my non-exupery-loaded images?
>
> Loading Exupery from SqueakMap is perfectly safe. The main reason
> I provide the pre-build VMs is to make it easier to get started.
> They are also a useful basis for macro benchmarks as the code behind
> the benchmark will be the same.
>
> I suspect that some of the differences with largeExplorer performance
> was due to image differences. I benchmarked using 3.8, Andy used
> 3.8.1, and Jakub used the pre-built 3.9.
>
> Bryce
> _______________________________________________
> Exupery mailing list
>
[hidden email]
>
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/exupery>
_______________________________________________
Exupery mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/exupery
Locked
