[llvm-commits] [PATCH] Add support for lowering FNEG with AVX

[Chad Rosier]

Before committing I'll take care of the missing newline at EOF.

On Dec 14, 2011, at 2:13 PM, Chad Rosier wrote:

> With AVX enabled the included test case clang would assert:
> Assertion failed: (VT.getSizeInBits() == Operand.getValueType().getSizeInBits() && "Cannot BITCAST between types of different sizes!")
> 
> This patch add support for lowering FNEG when AVX is enabled.
> 
> Hoping for feedback from someone with more AVX experience (e.g., Bruno, Craig, Evan, Eli).
> 
> Chad
> 
> <LowerFNEG_AVX.patch>