.NET compact framework floating point support on x86 and ARM?
- From: "Wil S" <wstark_NOSPAM_04@xxxxxxxxx>
- Date: Fri, 15 Dec 2006 10:29:39 -0800
I'm trying to benchmark floating point operations invoked from C# running
WinCE 5.0 on both an AMD Geode (x86 compatible) platform and an i.Mx31
(ARM). Both chips have a floating point unit. It is my understanding that
neither an MS native C/C++ compiler nor the CF JITter will generate floating
point instructions for the ARM chip, resulting in dismal floating point
performance. I don't know about the JITting on Geode, but it seems like
being x86 compatible would be a help in leveraging work from the
full-framework JITter.
How can I tell if either JITter (ARM or x86) is producing native floating
point instructions? Is this published somewhere (that I've not been able to
find)? Or can someone describe a technique for determining this?
Thanks,
Wil
.
- Prev by Date: Re: .NET and multithread networks server application
- Next by Date: Re: # of current physical Threads > 4 000 000 0000
- Previous by thread: .NET and multithread networks server application
- Next by thread: Re: Generic collections vs. arrays on structs vs. class
- Index(es):
Relevant Pages
|