[LLVMdev] Register pressure mechanism in PRE or Smarter rematerialization/split/spiller/coalescing ?

[Lawrence]

Hi, Daniel:

Thanks, I tried that patch you provided, it is better than just disabling your previous patch, it has more improvements than degradations.

Do you want to post that patch or you want me to do that?

Regards

Lawrence Hu

-----Original Message-----
From: Daniel Berlin [mailto:dberlin at dberlin.org] 
Sent: Wednesday, July 15, 2015 1:36 PM
To: Lawrence
Cc: LLVM Developers Mailing List
Subject: Re: Register pressure mechanism in PRE or Smarter rematerialization/split/spiller/coalescing ?

On Wed, Jul 15, 2015 at 1:10 PM, Daniel Berlin <dberlin at dberlin.org> wrote:
> IMHO, This doesn't make a lot of sense to turn off this part on it's own.
> I would just use the enable-pre flag to turn off scalar PRE, as it 
> will cause the same issue in other cases as well.
> Is there some reason you aren't just doing that?
> I suspect if this is a performance win, that would be as well.
>

Ugh, actually, it should be a win with the following change:


diff --git a/lib/Transforms/Scalar/GVN.cpp b/lib/Transforms/Scalar/GVN.cpp index 2c47a8a..a3387e3 100644
--- a/lib/Transforms/Scalar/GVN.cpp
+++ b/lib/Transforms/Scalar/GVN.cpp
@@ -1767,7 +1767,7 @@ bool GVN::processNonLocalLoad(LoadInst *LI) {
   }

   // Step 4: Eliminate partial redundancy.
-  if (!EnablePRE || !EnableLoadPRE)
+  if (!EnableLoadPRE)
     return false;

   return PerformLoadPRE(LI, ValuesPerBlock, UnavailableBlocks);




This will disable Scalar PRE without disabling load PRE.


(note, again, however, that load PRE can create exactly the same GEP situation you are referring to)