Re: [struct-reorg] strip struct frame from one-field structures.

["Richard Guenther" <richard dot guenther at gmail dot com>]

On Mon, May 12, 2008 at 3:00 PM, Razya Ladelsky <RAZYA@il.ibm.com> wrote:
> "Richard Guenther" <richard.guenther@gmail.com> wrote on 12/05/2008
>  15:35:22:
>
>
>
>  > On Mon, May 12, 2008 at 1:41 PM, Alon Dayan <ALOND@il.ibm.com> wrote:
>  > >
>  > >  This patch extends struct-reorg optimizations with an ability to
>  strip a
>  > >  structure frame from one-field structures.
>  > >
>  > >  For example, the variable 'a' of structure type:
>  > >
>  > >  typedef struct str
>  > >  {
>  > >         int* a;
>  > >  }
>  > >  str a;
>  > >
>  > >  will be replaced with
>  > >  int* a;
>  > >
>  > >  It works for both user-defined and compiler generated structures, as,
>  for
>  > >  example, those produced after structure peeling.
>  > >  The param struct-reorg-strip-struct was added to activate this
>  > >  optimization.
>  > >
>  > >  With this patch we measured additional 15% on 179.art over basic
>  > >  struct-reorg on p5.
>  >
>  > Did you analyze why?  Because ...
>  >
>  > >  But more than that, it opens further opportunities to subsequent
>  > >  optimizations by simplifying variable type
>  > >  from 'structure' type to field type, that might be easier to analyze
>  and
>  > >  transform. For example, with this
>  > >  patch a one-field structure with an array field becomes a subject for
>  > >  matrix flattening and transposing,
>  > >  while without this patch it is not.
>  >
>  > ... I don't buy this.  It shouldn't make a difference to analysis if the
>  > int * is wrapped inside a struct or not.  And if so, the affected passes
>  > should be fixed.
>  >
>  > Richard.
>
>  i think the intentino is that instead of having an array of structures,
>  we now have an array of int*, which makes the array a 2 dimensional array,
>  thereby a candidate for matrix transformations.

I see.  But this is not what the example shows ;)  And still enhancing the
analysis phase for matrix transformations should automagically handle this
case as well.

Richard.