Re: improving auto increment expressions detection across basic blocks.

2 msg	Doubt about scheduling
3 msg	omp workshare (PR35423) & beginner questions
1 msg	TREE_ADDRESSABLE flag
1 msg	gcc-4.2-20080409 is now available
4 msg	Where is scheduling going wrong? - GCC-4.1.2
3 msg	sizeof(array) with variable-length array parameter
12 msg	more m32c brokenness
20 msg	Problem with reloading in a new backend...
1 msg	KwGdOwMp 4329040559400834703
5 msg	Problem with Fix for PR 35163 ?
2 msg	gcc4.x handling of switch(enum)
1 msg	gcc-4.1-20080407 is now available
1 msg	An x86 gcc 4.3 branch
3 msg	[RFC] Implement __builtin_bswap* for ARMv6
16 msg	Copyright assignment wiki page

improving auto increment expressions detection ...
\ Ramana Radhakrishnan (7 Apr 2008)
. \ Andrew Pinski (7 Apr 2008)
. . \ Ramana Radhakrishnan (7 Apr 2008)
. . . \ Kenneth Zadeck (8 Apr 2008)
. . . . \ Mark Mitchell (8 Apr 2008)
. . . . . \ Kenneth Zadeck (8 Apr 2008)
. . . . . . \ Kenneth Zadeck (9 Apr 2008)
. . \ Ramana Radhakrishnan (9 Apr 2008)

118 msg	US-CERT Vulnerability Note VU#162289
2 msg	comments do not match code.
4 msg	RFC Test suite fix testing of no_trampolines
2 msg	Translating and proofreading your professional ...

Subject:	Re: improving auto increment expressions detection across basic blocks.
Group:	Gcc
From:	Ramana Radhakrishnan
Date:	9 Apr 2008

Hi Andrew,

I've been looking at doing this possibly in the store sinking pass and
have the following query as below.

On Mon, Apr 7, 2008 at 9:11 PM, Andrew Pinski <pinskia> wrote:
> On Mon, Apr 7, 2008 at 3:31 AM, Ramana Radhakrishnan <ramana.r> wrote:
> > The basic case is as explained below.
> >
> > for (i = 0; i < 100; i ++)
> > {
> > if (....)
> > {
> > a[i] = something;
> > }
> > else
> > a[i] = something else..
> >
> > }
>
> If it is this case, I think the store to a[i] should really sinked
> below the if statement instead of doing anything fancy with
> autoincrement.

Reg. the store sinking that you point out I've been looking at
tree-ssa-sink.c see if I can do something there .In sink_local it
appears as though the whole thing works by using dominator information
. The sinking process works by going block by block and attempting to
sink to a location which is dominated by this node.

Correct me if I am wrong but if you looked at the basic blocks in
which the stores are already contained, they would not be considered
for store sinking because they don't dominate anything really in the
CFG. So one wouldn't be able to do this in the current infrastructure
that we have in GCC. Is there any other place that one could look at ?

Cheers
Ramana

>
> Thanks,
> Andrew Pinski
>

--
Ramana Radhakrishnan