[Mono-list] Disabling all bounds checking ...

Rodrigo Kumpera kumpera at gmail.com
Tue Jan 19 14:43:59 EST 2010


Look at MONO_ARCH_EMIT_BOUNDS_CHECK in ir-emit.h and mini-x86.h.

Making this macro generate a NOP should be enough.

Thou I would rather advice that you guys use unsafe pointers for this as it
allows
to do the same without hacking the runtime.


On Fri, Jan 15, 2010 at 4:46 PM, CarvalhoJunior <carvalho.heron at gmail.com>wrote:

>
> Hi,
>
>   As part of a research work on optimization of Mono for high performance
> computing applications, we need to perform some benchmarks with a version
> of
> Mono with all array bounds checking removed. For
> that, we have implemented a version of a subset of NAS Parallel Benchmarks
> (sequencial case) in C#, which is freely avaliable
> (http://code.google.com/p/npb-for-hpe/). Using it, we have compared
> performance of Mono and JVM. We are now working on parallelization of the
> code using MPI.NET and further "componentization" using our parallel
> components platform
> (http://code.google.com/p/hash-programming-environment/).
>
>   Can someone here to teach us how to remove all bounds checking in mono
> runtime ? I know this is only possible by modifying the source code and
> recompiling. I would like to know what modules are
> responsible to insert bounds checking code.
>
> Heron
> --
> View this message in context:
> http://old.nabble.com/Disabling-all-bounds-checking-...-tp27181840p27181840.html
> Sent from the Mono - General mailing list archive at Nabble.com.
>
> _______________________________________________
> Mono-list maillist  -  Mono-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/mono-list
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ximian.com/pipermail/mono-list/attachments/20100119/3d5aff8f/attachment.html 


More information about the Mono-list mailing list