Yes, thanks for pointing it out. We're aware of the situation and investigating.
r10651 was a correctness fix, so we'd prefer to not simply revert it. On Fri, Feb 10, 2012 at 15:57, M Capewell <[email protected]>wrote: > I think 10651 caused Kraken to tank. Here's the relative performance > of gaussian_blur we measure, with 10650 as baseline 1.0. As you can > see, reverting 10652 (at 10659) didn't fix the regression. > > 10674 3.2 > 10671 3.1 > 10669 3.1 > 10668 3.1 > 10659 3.0 > 10655 3.1 > 10652 3.1 > 10651 3.1 > 10650 1.0 > > - Martyn Capewell > > -- > v8-dev mailing list > [email protected] > http://groups.google.com/group/v8-dev > -- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
