[llvm-commits] [llvm] r120932 - in /llvm/trunk: lib/Target/X86/X86ISelLowering.cpp test/CodeGen/X86/select.ll

[Bob Wilson]

On Dec 5, 2010, at 8:39 PM, Chris Lattner wrote:

> 
> On Dec 5, 2010, at 8:25 PM, Evan Cheng wrote:
>>>> Right, they don't handle the integer overflow case.  In this particular example, what we really want is a way to say that the select is heavily biased, so that we get a jump instead of cmov (or cmov equivalent).  We really want:
>>>> 
>>>> mulq <reg>
>>>> jo Lfoo
>>>> jmp __Znam
>>>> Lfoo:
>>>> mov rdi, -1
>>>> jmp __Znam
>>> 
>>> Along these lines, I wonder if it would be enough to just mark IR select/condbr instructions with an instruction MDNode like !highlybiased (optionally with a direction).  This could be preserved down to codegen level and used for expansion there.  This would also be a straight-forward way to model __builtin_expect.
>> 
>> Do we want to make it a more generic MDNode so we can use it to encode branch probability?
> 
> Yes, making it more general than just "biased" makes sense.

I've been thinking about and talking with a few people about doing exactly that for recording profile information as well.  I don't yet have enough details to make a real proposal, but if you're planning to do anything like that soon, I'd be interested.