[PATCH] Add support for fast-math flags to the FCmp instruction.

[hfinkel at anl.gov]

LGTM.


REPOSITORY
  rL LLVM

================
Comment at: docs/LangRef.rst:6984
@@ +6983,3 @@
+only flags that have any effect on its semantics are those that allow
+assumptions to be made about the domain of input arguments; namely
+``nnan``, ``ninf``, and ``nsz``. See :ref:`fastmath` for more information.
----------------
I'd rather word this as, "that allow assumptions to be made about the values of the input arguments", because implying that there is a mathematical 'domain' that includes NaN might be confusing.


================
Comment at: docs/LangRef.rst:6988
@@ -6987,3 +6996,1 @@
 
-Note that the code generator does not yet support vector types with the
-``fcmp`` instruction.
----------------
Why are you deleting this statement in this patch? Is this a separate cleanup?

http://reviews.llvm.org/D9793

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