[PATCH] Experiment with keeping GEPs near calls

[Quentin Colombet]

Hi Daniel,

I do not know if you still expect feedback from this post, as we already discuss the matter on IRC.
I add a couple of comment here for the record.

Note: Please move the commit list to llvm instead of CFE :).

Thanks,
Q.


================
Comment at: lib/CodeGen/RegisterCoalescer.cpp:847
@@ +846,3 @@
+               << TII->isTriviallyReMaterializable(DefMI, AA) << "\n";
+  //if (!TII->isAsCheapAsAMove(DefMI))
+  //  return false;
----------------
This test is used to avoid rematerializing expensive operation.

================
Comment at: lib/CodeGen/RegisterCoalescer.cpp:849
@@ +848,3 @@
+  //  return false;
+  //if (!TII->isTriviallyReMaterializable(DefMI, AA))
+  //  return false;
----------------
This check is used to ensure that we won’t introduce correctness issues.
The following code only supports trivial rematerialization. For more general rematerialization, you have to do more interfere checks.

http://reviews.llvm.org/D7259

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/