Re: Avoid multiple entry SCC regions

2014-07-28 Thread Jan Hubicka
Hi, here is updated version of the patch to avoid streaming trees that are local. I made it mostly to see what effect it have on collisions (as I think having unmergeable trees in the SCC hash artificially increase collision rate for trees derrived from this). To my surprise the number are quite

Avoid multiple entry SCC regions

2014-07-25 Thread Jan Hubicka
Hi, this patch make SCC hashes stronger by using DFS to get entry point independent order. The DFS refactoring work is Richards, only hash_tree changes are mine and I am responsible for all bugs :) Details are hopefully well documented in hash_scc. I tested the patch on Firefox and we manage to

Re: Avoid multiple entry SCC regions

2014-07-25 Thread Andi Kleen
Jan Hubicka hubi...@ucw.cz writes: I am lto bootstrapping/regtesting x86_64-linux and intend to comming once it passes. You'll have to redo it with hstates, sorry, as it conflicts with my patchkit which I checked in earlier. -Andi -- a...@linux.intel.com -- Speaking for myself only

Re: Avoid multiple entry SCC regions

2014-07-25 Thread Jan Hubicka
Jan Hubicka hubi...@ucw.cz writes: I am lto bootstrapping/regtesting x86_64-linux and intend to comming once it passes. You'll have to redo it with hstates, sorry, as it conflicts with my patchkit which I checked in earlier. Yep I noticed that earlier and I am testing updated patch. I

Re: Avoid multiple entry SCC regions

2014-07-25 Thread Jan Hubicka
Jan Hubicka hubi...@ucw.cz writes: I am lto bootstrapping/regtesting x86_64-linux and intend to comming once it passes. You'll have to redo it with hstates, sorry, as it conflicts with my patchkit which I checked in earlier. Hi, this is vairant of patch I comitted. Richard