]> granicus.if.org Git - llvm/commitdiff
Update docs for fast-math flags.
authorJay Foad <jay.foad@gmail.com>
Fri, 18 Oct 2019 16:07:09 +0000 (16:07 +0000)
committerJay Foad <jay.foad@gmail.com>
Fri, 18 Oct 2019 16:07:09 +0000 (16:07 +0000)
This adds fneg, phi and select to the list of operations that may use
fast-math flags.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@375250 91177308-0d34-0410-b5e6-96231b3b80d8

docs/LangRef.rst

index d2fe2acc7d593d6e815862d163e33eeab6e42b2c..c3f401126e163a8d5d852797a2d52dde584634b3 100644 (file)
@@ -2459,9 +2459,10 @@ Code that requires different behavior than this should use the
 Fast-Math Flags
 ---------------
 
-LLVM IR floating-point operations (:ref:`fadd <i_fadd>`,
+LLVM IR floating-point operations (:ref:`fneg <i_fneg>`, :ref:`fadd <i_fadd>`,
 :ref:`fsub <i_fsub>`, :ref:`fmul <i_fmul>`, :ref:`fdiv <i_fdiv>`,
-:ref:`frem <i_frem>`, :ref:`fcmp <i_fcmp>`) and :ref:`call <i_call>`
+:ref:`frem <i_frem>`, :ref:`fcmp <i_fcmp>`), :ref:`phi <i_phi>`,
+:ref:`select <i_select>` and :ref:`call <i_call>`
 may use the following flags to enable otherwise unsafe
 floating-point transformations.