[LLVMdev] Question about Value Range Propagation

[Chris Lattner]

On Feb 22, 2011, at 7:25 AM, Reid Kleckner wrote:

> On Tue, Feb 22, 2011 at 6:19 AM, Duncan Sands <baldrick at free.fr> wrote:
>> the big problem with Patterson's VRP is that it is expensive in terms of
>> compile time.  LLVM used to have some passes (ABCD, predsimplify) that did
>> this kind of thing, but they were removed essentially because their compile
>> time was too great for the goodness they brought.
> 
> Any reason not to just leave them on at O3?  Based on the discussion
> around your simple condition propagation pass, it seemed predsimplify
> did delete dead code, but it didn't really improve generated code
> performance.  O3 seems the appropriate place to put expensive
> optimizations with diminishing returns.

Hi Reid,

-O3 compile time matters.  We don't turn things on in -O3 or -O4 that just burn compiler cycles but don't add value.

-Chris