[LLVMdev] Convert fdiv - X/Y -> X*1/Y

[Chad Rosier]

I believe the advice that Owen gave me was that if it's simple enough, we
should do it in both places. (Feel free to chime in if I'm misquoting you,
Owen.)

Performing the transformation early as an InstCombine can expose additional
opportunities to optimize code.  The purpose of the redundant DAGCombine
would be to catch the case where the optimization is exposed by another DAG
combine.  And as Jim also pointed out, the DAG combine has the benefit of
target-specific knowledge.

I agree, this will require quite a bit of testing.

 Chad

On Thu, Aug 8, 2013 at 1:13 PM, Shuxin Yang <shuxin.llvm at gmail.com> wrote:

>  I did few transformation in Instruction *InstCombiner::visitFDiv() in an
> attempt to remove some divs.
> I may miss this case.  If you need to implement this rule, it is better
> done in Instcombine than in DAG combine.
> Doing such xform early expose the redundancy of 1/y, which have positive
> impact to neighboring code,
> while DAG combine is bit blind.
>
> You should be very careful, reciprocal is very very very imprecise
> transformation. Sometimes you will see big different
> with and without this xform.
>
> On 8/8/13 9:25 AM, Chad Rosier wrote:
>
> I would like to transform X/Y -> X*1/Y.  Specifically, I would like to
> convert:
>
> define void @t1a(double %a, double %b, double %d) {
> entry:
>   %div = fdiv fast double %a, %d
>   %div1 = fdiv fast double %b, %d
>   %call = tail call i32 @foo(double %div, double %div1)
>   ret void
> }
>
> to:
>
> define void @t1b(double %a, double %b, double %d) {
> entry:
>   %div = fdiv fast double 1.000000e+00, %d
>   %mul = fmul fast double %div, %a
>   %mul1 = fmul fast double %div, %b
>   %call = tail call i32 @foo(double %mul, double %mul1)
>   ret void
> }
>
> Is such a transformation best done as a (target-specific) DAG combine?
>
> A similar instcombine already exists for the X/C->X*1/C case (see the
> CvtFDivConstToReciprocal function in InstCombineMlDivRem.cpp), but I don't
> believe the above can be done as an instcombine as it creates a new
> instruction (in addition to replacing the original).  Also, I only want to
> perform the transformation if there are multiple uses of 1/Y (like in my
> test case).  Otherwise, the transformation replaces a fdiv with a fdiv+fmul
> pair, which I doubt would be profitable.
>
> FWIW, I'm also pretty sure this combine requires -fast-math.
>
> Can someone point me in the right direction?
>
>  Thanks,
>   Chad
>
>
> _______________________________________________
> LLVM Developers mailing listLLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.eduhttp://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>
>
>
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20130808/3a1254c2/attachment.html>