This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
RE: [plugins-ici-cloning-instrumentation] new GCC plugin developements
- From: "Grigori Fursin" <grigori dot fursin at inria dot fr>
- To: "'Basile STARYNKEVITCH'" <basile at starynkevitch dot net>
- Cc: <ctuning-discussions at googlegroups dot com>, "'Joern Rennecke'" <amylaar at spamcop dot net>, "'Ian Lance Taylor'" <iant at google dot com>, "'Yuanjie Huang'" <huangyuanjie at ict dot ac dot cn>, "'Liang Peng'" <pengliang at ict dot ac dot cn>, "'Zbigniew Chamski'" <zbigniew dot chamski at gmail dot com>, "'GCC Mailing List'" <gcc at gcc dot gnu dot org>, "'Yuri Kashnikoff'" <yuri dot kashnikoff at gmail dot com>, "'Diego Novillo'" <dnovillo at google dot com>
- Date: Fri, 6 Nov 2009 18:08:58 +0100
- Subject: RE: [plugins-ici-cloning-instrumentation] new GCC plugin developements
- References: <4AE6E471.4020200@starynkevitch.net> <4AE7014A.1000903@starynkevitch.net> <mcr3a54lwb4.fsf@dhcp-172-17-9-151.mtv.corp.google.com> <4AE70C5E.4050005@starynkevitch.net> <84fc9c000910270839v2d9efe0dw829c8647f361c96f@mail.gmail.com> <4AE7164D.9010200@starynkevitch.net> <84fc9c000910270855w736df367qe511d8db280aaeb4@mail.gmail.com> <2dc303d60910271056h17038110ib63c53cfa374f5c7@mail.gmail.com> <002c01ca5746$9dd43da0$d97cb8e0$@fursin@inria.fr> <20091102074959.p8410ulv28sg0w44-nzlynne@webmail.spamcop.net> <008601ca5e09$93164fb0$b942ef10$@fursin@inria.fr> <20091105082557.75c2estyoog8ss0c-nzlynne@webmail.spamcop.net> <00b101ca5e1f$6ecf31b0$4c6d9510$@fursin@inria.fr> <mcrocng1r8s.fsf@dhcp-172-17-9-151.mtv.corp.google.com> <019b01ca5eea$c70e45c0$552ad140$@fursin@inria.fr> <20091106095419.fvs0otmbwowg0scw-nzlynne@webmail.spamcop.net> <4AF446BE.5050602@starynkevitch.net> <01c101ca5efb$f8b5d9b0$ea218d10$@fursin@inria.fr> <4AF455FE.9090603@starynkevitch.net>
Yes, that is true and that is one of the most usable features of ICI so far
since we use it to manipulate passes, add new analysis passes of generic function cloning
(during GSOC'09), etc. So we will need a hard-wired hook to substitute pass manager.
Other hooks such as fine-grain tuning for another GSOC'09 project
are still experimental and are not hardwired, but just referenced by some strings ...
Cheers,
Grigori
> -----Original Message-----
> From: Basile STARYNKEVITCH [mailto:basile@starynkevitch.net]
> Sent: Friday, November 06, 2009 6:00 PM
> To: Grigori Fursin
> Cc: ctuning-discussions@googlegroups.com; 'Joern Rennecke'; 'Ian Lance Taylor'; 'Yuanjie
> Huang'; 'Liang Peng'; 'Zbigniew Chamski'; 'GCC Mailing List'; 'Yuri Kashnikoff'; 'Diego
> Novillo'
> Subject: Re: [plugins-ici-cloning-instrumentation] new GCC plugin developements
>
> Grigori Fursin wrote:
> > Hi Basile,
> >
> > Well, the point of ICI is actually not to hardwire specific events at the beginning
> > but allow users/researchers easily add their own hooks referenced by any string
> > to simplify prototyping of new ideas. Then if they see that it is useful and they
> > can persuade the community about that, those hooks go to the mainline and are hardwired
> > in the PLUGIN_* event ...
>
>
> My intuition was that ICI need a single (or probably a few) hook to replace the entire pass
> manager....
> Then the ICI plugin would call the existing GCC passes in the order it wants....
>
> Regards.
> --
> Basile STARYNKEVITCH http://starynkevitch.net/Basile/
> email: basile<at>starynkevitch<dot>net mobile: +33 6 8501 2359
> 8, rue de la Faiencerie, 92340 Bourg La Reine, France
> *** opinions {are only mines, sont seulement les miennes} ***