Re: [PATCH] alternative hirate for builtin_expert

2013-10-08 Thread Ramana Radhakrishnan
Can someone comment / approve it quickly so that we get AArch32 and AArch64 linux cross-builds back up ? Ok. Applied for Dehao as r203269 . Tests on arm came back ok. Ramana Thanks, Richard. regards Ramana Honza Dehao Honza

Re: [PATCH] alternative hirate for builtin_expert

2013-10-08 Thread Dehao Chen
Thanks for applying the patch. Backported to google-4_8 I still have some concern when inlining .part function into its original function: basically, the gimple_block for that call may be NULL, but it does not make sense to clear all block info for all stmts in the .part function. Dehao On Tue,

Re: [PATCH] alternative hirate for builtin_expert

2013-10-07 Thread Ramana Radhakrishnan
On 10/04/13 22:23, Jan Hubicka wrote: On Fri, Oct 4, 2013 at 11:54 AM, Jan Hubicka hubi...@ucw.cz wrote: I looked at this problem. Bug updated http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58619 This is a bug when updating block during tree-inline. Basically, it is legal for *n to be NULL. E.g.

Re: [PATCH] alternative hirate for builtin_expert

2013-10-07 Thread Richard Biener
On Mon, Oct 7, 2013 at 12:02 PM, Ramana Radhakrishnan ramra...@arm.com wrote: On 10/04/13 22:23, Jan Hubicka wrote: On Fri, Oct 4, 2013 at 11:54 AM, Jan Hubicka hubi...@ucw.cz wrote: I looked at this problem. Bug updated http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58619 This is a bug when

Re: [PATCH] alternative hirate for builtin_expert

2013-10-04 Thread Ramana Radhakrishnan
On 10/02/13 23:49, Rong Xu wrote: Here is the new patch. Honaz: Could you take a look? Thanks, -Rong On Wed, Oct 2, 2013 at 2:31 PM, Jan Hubicka hubi...@ucw.cz wrote: Thanks for the suggestion. This is much cleaner than to use binary parameter. Just want to make sure I understand it

Re: [PATCH] alternative hirate for builtin_expert

2013-10-04 Thread Xinliang David Li
Dehao, can you take a look? David On Fri, Oct 4, 2013 at 6:05 AM, Ramana Radhakrishnan ramra...@arm.com wrote: On 10/02/13 23:49, Rong Xu wrote: Here is the new patch. Honaz: Could you take a look? Thanks, -Rong On Wed, Oct 2, 2013 at 2:31 PM, Jan Hubicka hubi...@ucw.cz wrote: Thanks

Re: [PATCH] alternative hirate for builtin_expert

2013-10-04 Thread Rong Xu
My change on the probability of builtin_expect does have an impact on the partial inline (more outlined functions will get inline back to the original function). I think this triggers an existing issue. Dehao will explain this in his coming email. -Rong On Fri, Oct 4, 2013 at 6:05 AM, Ramana

Re: [PATCH] alternative hirate for builtin_expert

2013-10-04 Thread Jan Hubicka
I looked at this problem. Bug updated http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58619 This is a bug when updating block during tree-inline. Basically, it is legal for *n to be NULL. E.g. When gimple_block(id-gimple_call) is NULL, remap_blocks_to_null will be called to set *n to NULL. The

Re: [PATCH] alternative hirate for builtin_expert

2013-10-04 Thread Dehao Chen
On Fri, Oct 4, 2013 at 11:54 AM, Jan Hubicka hubi...@ucw.cz wrote: I looked at this problem. Bug updated http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58619 This is a bug when updating block during tree-inline. Basically, it is legal for *n to be NULL. E.g. When gimple_block(id-gimple_call) is

Re: [PATCH] alternative hirate for builtin_expert

2013-10-04 Thread Jan Hubicka
On Fri, Oct 4, 2013 at 11:54 AM, Jan Hubicka hubi...@ucw.cz wrote: I looked at this problem. Bug updated http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58619 This is a bug when updating block during tree-inline. Basically, it is legal for *n to be NULL. E.g. When

Re: [PATCH] alternative hirate for builtin_expert

2013-10-03 Thread Jan Hubicka
Here is the new patch. Honaz: Could you take a look? OK, thanks! Honza

Re: [PATCH] alternative hirate for builtin_expert

2013-10-02 Thread Jan Hubicka
Hi, Current default probability for builtin_expect is 0.9996. This makes the freq of unlikely bb very low (4), which suppresses the inlining of any calls within those bb. We used FDO data to measure the branch probably for the branch annotated with builtin_expert. For

Re: [PATCH] alternative hirate for builtin_expert

2013-10-02 Thread Rong Xu
On Wed, Oct 2, 2013 at 9:08 AM, Jan Hubicka hubi...@ucw.cz wrote: Hi, Current default probability for builtin_expect is 0.9996. This makes the freq of unlikely bb very low (4), which suppresses the inlining of any calls within those bb. We used FDO data to measure the branch

Re: [PATCH] alternative hirate for builtin_expert

2013-10-02 Thread Jan Hubicka
Thanks for the suggestion. This is much cleaner than to use binary parameter. Just want to make sure I understand it correctly about the orginal hitrate: you want to retire the hitrate in PRED_BUILTIN_EXPECT and always use the one specified in the biniltin-expect-probability parameter. Yes.

Re: [PATCH] alternative hirate for builtin_expert

2013-10-02 Thread Rong Xu
Here is the new patch. Honaz: Could you take a look? Thanks, -Rong On Wed, Oct 2, 2013 at 2:31 PM, Jan Hubicka hubi...@ucw.cz wrote: Thanks for the suggestion. This is much cleaner than to use binary parameter. Just want to make sure I understand it correctly about the orginal hitrate: you

Re: [PATCH] alternative hirate for builtin_expert

2013-10-01 Thread Rong Xu
ping. On Fri, Sep 27, 2013 at 3:53 PM, Rong Xu x...@google.com wrote: Hi, Current default probability for builtin_expect is 0.9996. This makes the freq of unlikely bb very low (4), which suppresses the inlining of any calls within those bb. We used FDO data to measure the branch

[PATCH] alternative hirate for builtin_expert

2013-09-27 Thread Rong Xu
Hi, Current default probability for builtin_expect is 0.9996. This makes the freq of unlikely bb very low (4), which suppresses the inlining of any calls within those bb. We used FDO data to measure the branch probably for the branch annotated with builtin_expert. For google internal