[LLVMdev] Use clang to generate LLVM IR with -O3, how to visit these load operations after -reg2mem

[Eric Lu]

Hi, David
Thanks for your reply.

On Wed, Apr 22, 2015 at 12:37 PM, David Blaikie <dblaikie at gmail.com> wrote:

> On Tue, Apr 21, 2015 at 9:29 PM, Eric Lu <eirc.lew at gmail.com> wrote:
> >
> > Best Regards!
> > Eric Lew
> >
> > On 周三, 4月 22, 2015 at 12:11 下午, David Blaikie <dblaikie at gmail.com>
> wrote:
> >
> > On Tue, Apr 21, 2015 at 8:34 PM, Eric Lu <eirc.lew at gmail.com> wrote:
> >> Hi, all
> >>
> >> I want to instrument load/store operations in LLVM IR. And I find the
> LLVM
> >> IR generated with flag -o3 is much more efficient than -o0, so I try to
> >> instrument on these optimized LLVM IR, but, some load operations are
> >> changed
> >> to Phi loads. Then, I try the -reg2mem flag to remove these phi nodes.
> >> The  problems are:
> >> 1) I use visitLoadInst to collect these load operations, and it seems I
> >> cannot reach these loads that are generated by reg2mem? What is the
> >> reason?
> >
> > Clang doesn't generate optimized IR - so if your instrumentation is in
> > Clang's IRGen (clang/lib/CodeGen) it won't see these. Clang generates
> > the unoptimized IR (you can see it with -Xclang -disable-llvm-optzns)
> >
> > I do instrumentation in the opt of llvm. Does clang itself do some
> > optimizations that speed up applications with -O3 compared to -O0?
>
> A little bit, but not much.
>
> >
> >
> >> 2) Is there any way to compile with -O3 but not generate phi nodes ?
> >
> > Not likely - without that it'd disable a lot of LLVM's optimization
> power.
> >
> > - David
> >
> >>
> >>
> >> --
> >> Best Regards!
> >> Eric Lew
> >>
> >> _______________________________________________
> >> LLVM Developers mailing list
> >> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> >> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
> >>
>



-- 
Best Regards!
Eric Lew
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20150422/01355d13/attachment.html>