llvm dev - Mar 2013 - [LLVMdev] Dependence Analysis on Machine code

Hi,

I am trying to do dependence analysis (loop dependences) on machine code.
But, scalar evolution doesn't work as expected (like at IR level - opt
pass).
I am not getting enough clue as to how to proceed..

One naive way of getting the dependence information can be
- Do dependence analysis at IR level and attach the information as meta
data.
  And then use this information at machine code level.
Can someone suggest better approach ??


-- 
Regards,
Rahul Patil.




--
View this message in context:
http://llvm.1065342.n5.nabble.com/Dependence-Analysis-on-Machine-code-tp55989.html
Sent from the LLVM - Dev mailing list archive at Nabble.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20130315/c39d9498/attachment.html>

On Mar 15, 2013, at 3:18 AM, rahul <rahul3527 at gmail.com> wrote:
> 
> Hi,
> 
> I am trying to do dependence analysis (loop dependences) on machine code.
> But, scalar evolution doesn't work as expected (like at IR level - opt
pass).
> I am not getting enough clue as to how to proceed..
> 
> One naive way of getting the dependence information can be
> - Do dependence analysis at IR level and attach the information as meta
data.
>   And then use this information at machine code level.
> Can someone suggest better approach ??
It sounds like you have a good clue how to proceed. I can't think of a
better way.
-Andy

Thanks Andrew for the reply :)
I will try this out and see if it gives some satisfactory results.

--Rahul


On Tue, Mar 19, 2013 at 1:38 AM, Andrew Trick <atrick at apple.com> wrote:
>
> On Mar 15, 2013, at 3:18 AM, rahul <rahul3527 at gmail.com> wrote:
>
> >
> > Hi,
> >
> > I am trying to do dependence analysis (loop dependences) on machine
code.
> > But, scalar evolution doesn't work as expected (like at IR level -
opt
> pass).
> > I am not getting enough clue as to how to proceed..
> >
> > One naive way of getting the dependence information can be
> > - Do dependence analysis at IR level and attach the information as
meta
> data.
> >   And then use this information at machine code level.
> > Can someone suggest better approach ??
>
> It sounds like you have a good clue how to proceed. I can't think of a
> better way.
> -Andy
>
>

-- 
Regards,
Rahul Patil.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20130319/79777454/attachment.html>