llvm dev - Dec 2013 - [LLVMdev] LLVM ARM VMLA instruction

Hi,

I was going through Code of LLVM instruction code generation for ARM. I
came across VMLA instruction hazards (Floating point multiply and
accumulate). I was comparing assembly code emitted by LLVM and GCC, where i
saw that GCC was happily using VMLA instruction for floating point while
LLVM never used it, instead it used a pair of VMUL and VADD instruction.

I wanted to know if there is any way in which these VMLA hazards can be
ignored and make LLVM to emit VMLA instructions? Is there any command line
option/compiler switch/FLAG for doing this? I tried '-ffast-math' but it
didn't work.


-- 
With regards,
Suyog Sarda
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20131218/05a4f2f5/attachment.html>

On 18 December 2013 07:55, suyog sarda <sardask01 at gmail.com> wrote:
> I wanted to know if there is any way in which these VMLA hazards can be
> ignored and make LLVM to emit VMLA instructions? Is there any command line
> option/compiler switch/FLAG for doing this? I tried '-ffast-math'
but it
> didn't work.
>
I believe the option you're looking for is: -mattr=-vmlx-forwarding

$ llc -mcpu=cortex-a9 -mattr=-vmlx-forwarding file.ll -0 file.s

cheers,
--renato
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20131218/6f4f160f/attachment.html>