Please review: Optimize vector multiply on X86

[Andrea_DiBiagio at sn.scee.net]

Hi Elena

Nadav Rotem <nrotem at apple.com> wrote on 26/06/2013 01:33:40:

> From: Nadav Rotem <nrotem at apple.com>
> To: "Demikhovsky, Elena" <elena.demikhovsky at intel.com>, 
> 
> Hi Elena, 
> 
> isSplatVector does not do what the name says. It checks it there is 
> a splat constant. It should have a different name. 

I agree with Nadav that the function name should be changed.
There is another function called 'isSplatVector' in X86ISelLowering.cpp: 
that function behaves differently from your new function since it checks 
for BUILD_VECTORs where all the operands are exactly the same. Also, 
operands are allowed to be of any type and not only ConstantSDNode or 
undef.
Calling your function the same way would be probably misleading. 

> 
> On 06/25/13, "Demikhovsky, Elena" <elena.demikhovsky at intel.com> wrote:
> Can anybody, please, take a look at this patch, I'd like to commit it.
> 
> Thank you.

Overall your patch looks good to me: I can see how vector multiply 
instructions are now correctly combined into shifts.

Cheers,
Andrea Di Biagio
SN Systems - Sony Computer Entertainment Group

**********************************************************************
This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed. 
If you have received this email in error please notify postmaster at scee.net
This footnote also confirms that this email message has been checked for 
all known viruses.
Sony Computer Entertainment Europe Limited
Registered Office: 10 Great Marlborough Street, London W1F 7LP, United 
Kingdom
Registered in England: 3277793
**********************************************************************

P Please consider the environment before printing this e-mail