[llvm-bugs] [Bug 20358] operating on unknown FP operands is a bad idea

[via llvm-bugs]

https://llvm.org/bugs/show_bug.cgi?id=20358

Sanjay Patel <spatel+llvm at rotateright.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
           See Also|                            |https://llvm.org/bugs/show_
                   |                            |bug.cgi?id=8100,
                   |                            |https://llvm.org/bugs/show_
                   |                            |bug.cgi?id=6050
         Resolution|---                         |WONTFIX

--- Comment #3 from Sanjay Patel <spatel+llvm at rotateright.com> ---
As noted in the patch review thread, this is a specific example of a general
problem. We shouldn't deoptimize normal FP code for the presumably rare
denormal occurrence. It's also not clear that all targets are as awful as x86
at handling denorms.

We really need an optimization flag/mode that tells us we're not allowed to
speculate FP ops. 

This probably also relates to honoring the FENV (bug 8100) and knowing which
exceptions are enabled (bug 6050).

Resolving as 'WONTFIX'.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20151124/67301aec/attachment.html>