[MonoDevelop] MonoDevelop OSX / Windows performance differences

[Mike Christensen]

I'm curious if anyone has done any perf benchmarking between the .NET
GC and the new Mono SGen GC..  If I understand, they're a similar
algorithm..  However, I've met the guy who wrote the .NET one and he's
friggen smart :)

Mike

On Sun, Mar 13, 2011 at 12:04 PM, Stifu <stifu at free.fr> wrote:
> That doesn't match what I've read or seen myself.
> The only thing that I found was significantly faster in Mono than in .NET
> was exception handling (which might be partly due to the fact Mono doesn't
> do as much as .NET, but I digress). Which doesn't matter too much, as
> exceptions are supposed to be exceptional.
>
> In most cases, I get similar performances and memory usages. Except for
> certain intensive operations, Windows technologies (WinForms, GDI+), or
> certain classes (DataSet, DataTable, etc...) where .NET is often clearly
> faster. Add to that Mono's GC (although it is being replaced), which
> sometimes performs much worse than the one of .NET.
>
> Anyway, back on topic, it's good to see you found the solution to your
> problem, Philippe.
> It would be interesting to see how these numbers compare to what you get on
> Windows.
>
>
> quandary wrote:
>>
>> According to some benchmarks I recently read (don't remember where
>> anymore),
>> mono was actually faster than .NET in most tested cases.
>>
>> And that's understandable.
>> First, the core mono runtime/VM is written mostly in C, while the core
>> .NET framework runtime/VM is written in C++, and second, mono on Linux
>> profits from the faster speed of Linux filesystems.
>>
>> (See Google's evaluation on EXT4 vs. XFS)
>>
>> On 03/13/2011 04:10 PM, Stifu wrote:
>>> I'm not saying Mono is always slower than .NET. But it is in many cases.
>>> The slowness you're seeing could be just a certain thing in Mono that is
>>> slow. It might also be a Mac-specific Mono issue.
>>>
>>>
>>> philippe.monteil wrote:
>>>> I am surprised by this explanation because the performance difference is
>>>> very important,
>>>> and because the first comparisons we have made between the .Net and Mono
>>>> versions of our
>>>> GIS engine show that their execution times are very close, if not better
>>>> for the Mono version...
>>>> We use Monodevelop on W7-64/BootCamp on one hand and on MaOSX on the
>>>> other, on the same i7/iMac...
>>>>
>>>
>>> --
>>> View this message in context:
>>> http://mono.1490590.n4.nabble.com/MonoDevelop-OSX-Windows-performance-differences-tp3351859p3352020.html
>>> Sent from the Mono - MonoDevelop IDE mailing list archive at Nabble.com.
>>> _______________________________________________
>>> Monodevelop-list mailing list
>>> Monodevelop-list at lists.ximian.com
>>> http://lists.ximian.com/mailman/listinfo/monodevelop-list
>>
>> _______________________________________________
>> Monodevelop-list mailing list
>> Monodevelop-list at lists.ximian.com
>> http://lists.ximian.com/mailman/listinfo/monodevelop-list
>>
>
>
> --
> View this message in context: http://mono.1490590.n4.nabble.com/MonoDevelop-OSX-Windows-performance-differences-tp3351859p3352396.html
> Sent from the Mono - MonoDevelop IDE mailing list archive at Nabble.com.
> _______________________________________________
> Monodevelop-list mailing list
> Monodevelop-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/monodevelop-list
>