[llvm] r184105 - Switch spill weights from a basic loop depth estimation to BlockFrequencyInfo.

[Benjamin Kramer]

On 17.06.2013, at 21:22, Chandler Carruth <chandlerc at google.com> wrote:

> This is really awesome.
> 
> On Mon, Jun 17, 2013 at 12:00 PM, Benjamin Kramer <benny.kra at googlemail.com> wrote:
> This is good for a 5% speedup on zlib's deflate (x86_64), there were some very
> unfortunate spilling decisions in its hottest loop in longest_match().
> 
> Could you add some reduced form of this loop to the regression tests just so we don't lose track of this loop structure going forward?

I've been pondering this, the loop is somewhat large though. I'll see if I can trim it down to an acceptable size but I'm afraid it will become too fragile for a regression test.

- Ben