[Mono-list] Benchmark Results - Mono vs .Net
Jeffrey Stedfast
fejj at novell.com
Mon Apr 23 23:43:35 EDT 2007
I've been playing around with this benchmark and have discovered that
some of the reasons mono performs slower than MS is that (g)mcs doesn't
optimize things like:
r += i++;
very well. If you instead write the code as:
r += i; i++;
the performance of the arithmetic loops under mono matches that of the
MS .Net vm.
As far as the other tests... I dunno yet :)
Jeff
On Fri, 2007-04-13 at 09:46 +0200, Henk Tiggelaar wrote:
> Below are the results of benchmarks under Mono 1.2.3.1 and
> Microsoft.NET 2.0. Benchmark code used can be found at
> http://www.tommti-systems.de/main-Dateien/reviews/languages/benchmarks.html
>
> Mono 1.2.3.1 results
> ================
> Int arithmetic elapsed time: 8094 ms
> Double arithmetic elapsed time: 12141 ms
> long arithmetic elapsed time: 26406 ms
> Trig elapsed time: 2749 ms
> IO elapsed time: 3204 ms
> Array elapsed time: 406 ms
> Exception elapsed time: 3719 ms
> HashMap elapsed time: 234 ms
> HashMaps elapsed time: 5265 ms
> HeapSort elapsed time: 609 ms
> Vector elapsed time: 9953 ms
> Matrix Multiply elapsed time: 71562 ms
> Nested Loop elapsed time: 10359 ms
> String Concat. (fixed) elapsed time: 578 ms
> Total C# benchmark time: 155279 ms
>
> .NET 2.0 results
> ================
> Int arithmetic elapsed time: 5812 ms
> Double arithmetic elapsed time: 7249 ms
> long arithmetic elapsed time: 16265 ms
> Trig elapsed time: 2281 ms
> IO elapsed time: 2499 ms
> Array elapsed time: 203 ms
> Exception elapsed time: 26687 ms
> HashMap elapsed time: 124 ms
> HashMaps elapsed time: 3999 ms
> HeapSort elapsed time: 531 ms
> Vector elapsed time: 9890 ms
> Matrix Multiply elapsed time: 33687 ms
> Nested Loop elapsed time: 22265 ms
> String Concat. (fixed) elapsed time: 359 ms
> Total C# benchmark time: 131851 ms
> _______________________________________________
> Mono-list maillist - Mono-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/mono-list
More information about the Mono-list
mailing list