This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [graphite] Loop tiling
- From: Albert Cohen <Albert dot Cohen at inria dot fr>
- To: Sebastian Pop <sebpop at gmail dot com>
- Cc: Tobias Grosser <grosser at fim dot uni-passau dot de>, Cedric Bastoul <cedric dot bastoul at inria dot fr>, GCC Patches <GCC at gcc dot gnu dot org>, Konrad Trifunovic <konrad dot trifunovic at gmail dot com>, louis-noel dot pouchet at inria dot fr, Karthik A <chill dot shanky at gmail dot com>, Harle Christophe <christophe dot harle at amd dot com>, Adrien ELICHE <aeliche at isty dot uvsq dot fr>
- Date: Wed, 11 Jun 2008 07:35:39 +0200
- Subject: Re: [graphite] Loop tiling
- References: <1213030724.1133.73.camel@tobilaptop> <cb9d34b20806091025j1bc05588qc0c7d0092d7fbfae@mail.gmail.com>
Hi all,
Yes, Sebastian is Right for now, and thanks for pointing out the references.
Yet in the longer term, we are working on CLooG extensions to facilitate
tiling with parametric sizes, multidomensional tiling, increased
scalability, no risks of integer overflows, and better quality of
generated code... :-).
The magic is mostly in a couple of papers by Radjopadhye and his
students, at PLDI'07 and SC'07. We are working on extending it to
affine-by-statement scheduling and imperfectly nested loops, and also to
make it run within one single run of CLooG (see those papers for details).
More info at an upcoming phone meeting, or if Cedric wants to comment a
little more.
Albert
Sebastian Pop wrote:
> Hi Tobi,
>
> Thanks for working on this.
>
> Solution 2 is the right one.
>
> On Mon, Jun 9, 2008 at 11:58 AM, Tobias Grosser
> <grosser@fim.uni-passau.de> wrote:
>> 1. It is not very efficient. As we have many multiplications in the
>> code.
>> Will later gcc optimization passes convert these multiplications to
>> additions?
>>
>
> Yes, you should not worry about scalar optimizations at all.
>
>> 2. I could not find any documentation about scattering functions, that
>> describe this way of using the scattering functions.
>>
>
> I think that the best place to look at this is:
> page 4 of http://www.lri.fr/~girbal/publi/ics05_facilitating.ps
>
> also have a look at the other reports from:
> http://www.lri.fr/~girbal/site_wrapit/
>
> Albert probably has some other pointers for reports that describe how
> to do loop tiling.
>
>> Another graphite specific problem is, how we connect the real loop
>> variables to the cloog variables. Before loop tiling this was easy.
>> Loops of the same depth in the cloog output correspond to the loops of
>> the same depth in the original gimple loop nest. But know we have to add
>> some data structure, that connects the gimple loops, with the cloog
>> loops.
>>
>
> This was also the case before: we needed a map between the old
> induction variable and the new ones.
>
> Sebastian Pop
> --
> AMD - GNU Tools