[llvm-commits] [llvm] r172868 - in /llvm/trunk: lib/Target/X86/X86ISelLowering.cpp test/CodeGen/X86/sandybridge-loads.ll test/CodeGen/X86/v8i1-masks.ll

[Nadav Rotem]

X86 cost model already knows that YMM loads take 2 cycles.  :)

On Jan 18, 2013, at 4:46 PM, Renato Golin Linaro <renato.golin at linaro.org> wrote:

> On 18 January 2013 23:10, Nadav Rotem <nrotem at apple.com> wrote:
> Author: nadav
> Date: Fri Jan 18 17:10:30 2013
> New Revision: 172868
> 
> URL: http://llvm.org/viewvc/llvm-project?rev=172868&view=rev
> Log:
> On Sandybridge loading unaligned 256bits using two XMM loads (vmovups and vinsertf128) is faster than using a single vmovups instruction.
> 
> Would the cost model need to be updated?
> 
> I'm just asking to understand better where the cost model fits in all things... ;)
> 
> cheers,
> --renato

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130118/eff14732/attachment.html>