On Thu, 6 Jun 2024, YunQiang Su wrote:
> Richard Biener <[email protected]> 于2024年5月28日周二 17:47写道：
> >
> > The following avoids accounting single-lane SLP to the discovery
> > limit. As the two testcases show this makes discovery fail,
> > unfortunately even not the same across targets. The following
> > should fix two FAILs for GCN as a side-effect.
> >
> > Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
> >
> > PR tree-optimization/115254
> > * tree-vect-slp.cc (vect_build_slp_tree): Only account
> > multi-lane SLP to limit.
> >
> > * gcc.dg/vect/slp-cond-2-big-array.c: Expect 4 times SLP.
> > * gcc.dg/vect/slp-cond-2.c: Likewise.
>
> With this patch, MIPS/MSA still has only 3 times SLP.
> I am digging the problem
I bet it's an issue with missed permutes. f3() requires interleaving
of two VnQImode vectors.
>
> > ---
> > .../gcc.dg/vect/slp-cond-2-big-array.c | 2 +-
> > gcc/testsuite/gcc.dg/vect/slp-cond-2.c | 2 +-
> > gcc/tree-vect-slp.cc | 31 +++++++++++--------
> > 3 files changed, 20 insertions(+), 15 deletions(-)
> >
> > diff --git a/gcc/testsuite/gcc.dg/vect/slp-cond-2-big-array.c
> > b/gcc/testsuite/gcc.dg/vect/slp-cond-2-big-array.c
> > index cb7eb94b3a3..9a9f63c0b8d 100644
> > --- a/gcc/testsuite/gcc.dg/vect/slp-cond-2-big-array.c
> > +++ b/gcc/testsuite/gcc.dg/vect/slp-cond-2-big-array.c
> > @@ -128,4 +128,4 @@ main ()
> > return 0;
> > }
> >
> > -/* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 3
> > "vect" } } */
> > +/* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 4
> > "vect" } } */
> > diff --git a/gcc/testsuite/gcc.dg/vect/slp-cond-2.c
> > b/gcc/testsuite/gcc.dg/vect/slp-cond-2.c
> > index 1dcee46cd95..08bbb3dbec6 100644
> > --- a/gcc/testsuite/gcc.dg/vect/slp-cond-2.c
> > +++ b/gcc/testsuite/gcc.dg/vect/slp-cond-2.c
> > @@ -128,4 +128,4 @@ main ()
> > return 0;
> > }
> >
> > -/* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 3
> > "vect" } } */
> > +/* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 4
> > "vect" } } */
> > diff --git a/gcc/tree-vect-slp.cc b/gcc/tree-vect-slp.cc
> > index 0dd9a4daf6a..bbfde8849c1 100644
> > --- a/gcc/tree-vect-slp.cc
> > +++ b/gcc/tree-vect-slp.cc
> > @@ -1725,21 +1725,26 @@ vect_build_slp_tree (vec_info *vinfo,
> > SLP_TREE_SCALAR_STMTS (res) = stmts;
> > bst_map->put (stmts.copy (), res);
> >
> > - if (*limit == 0)
> > + /* Single-lane SLP doesn't have the chance of run-away, do not account
> > + it to the limit. */
> > + if (stmts.length () > 1)
> > {
> > - if (dump_enabled_p ())
> > - dump_printf_loc (MSG_NOTE, vect_location,
> > - "SLP discovery limit exceeded\n");
> > - /* Mark the node invalid so we can detect those when still in use
> > - as backedge destinations. */
> > - SLP_TREE_SCALAR_STMTS (res) = vNULL;
> > - SLP_TREE_DEF_TYPE (res) = vect_uninitialized_def;
> > - res->failed = XNEWVEC (bool, group_size);
> > - memset (res->failed, 0, sizeof (bool) * group_size);
> > - memset (matches, 0, sizeof (bool) * group_size);
> > - return NULL;
> > + if (*limit == 0)
> > + {
> > + if (dump_enabled_p ())
> > + dump_printf_loc (MSG_NOTE, vect_location,
> > + "SLP discovery limit exceeded\n");
> > + /* Mark the node invalid so we can detect those when still in use
> > + as backedge destinations. */
> > + SLP_TREE_SCALAR_STMTS (res) = vNULL;
> > + SLP_TREE_DEF_TYPE (res) = vect_uninitialized_def;
> > + res->failed = XNEWVEC (bool, group_size);
> > + memset (res->failed, 0, sizeof (bool) * group_size);
> > + memset (matches, 0, sizeof (bool) * group_size);
> > + return NULL;
> > + }
> > + --*limit;
> > }
> > - --*limit;
> >
> > if (dump_enabled_p ())
> > dump_printf_loc (MSG_NOTE, vect_location,
> > --
> > 2.35.3
>
>
>
>
--
Richard Biener <[email protected]>
SUSE Software Solutions Germany GmbH,
Frankenstrasse 146, 90461 Nuernberg, Germany;
GF: Ivo Totev, Andrew McDonald, Werner Knoblich; (HRB 36809, AG Nuernberg)
