On Thu, 14 Apr 2016, Bill Schmidt wrote:
> On Thu, 2016-04-14 at 13:22 +0200, Richard Biener wrote:
> > The following fixes PR70130 - improved SLP capabilities now run into
> > the realignment code on ppc which doesn't properly verify that all
> > vector loads emitted by vectorizable_load share the same alignment.
> >
> > Bootstrap / regtest pending on x86_64-unknown-linux-gnu.
> >
> > Bootstrapped / tested on ppc64le by Alan.
> >
> > Richard.
> >
> > 2016-04-14 Richard Biener <rguent...@suse.de>
> > Alan Modra <amo...@gmail.com>
> >
> > PR tree-optimization/70130
> > * tree-vect-data-refs.c (vect_supportable_dr_alignment): Detect
> > when alignment stays not the same and no not use the realign
> > scheme then.
> >
> > * gcc.dg/vect/O3-pr70130.c: New testcase.
> >
> > Index: gcc/tree-vect-data-refs.c
> > ===================================================================
> > *** gcc/tree-vect-data-refs.c (revision 234970)
> > --- gcc/tree-vect-data-refs.c (working copy)
> > *************** vect_supportable_dr_alignment (struct da
> > *** 5983,5992 ****
> > || targetm.vectorize.builtin_mask_for_load ()))
> > {
> > tree vectype = STMT_VINFO_VECTYPE (stmt_info);
> > ! if ((nested_in_vect_loop
> > ! && (TREE_INT_CST_LOW (DR_STEP (dr))
> > ! != GET_MODE_SIZE (TYPE_MODE (vectype))))
> > ! || !loop_vinfo)
> > return dr_explicit_realign;
> > else
> > return dr_explicit_realign_optimized;
> > --- 5983,6001 ----
> > || targetm.vectorize.builtin_mask_for_load ()))
> > {
> > tree vectype = STMT_VINFO_VECTYPE (stmt_info);
> > !
> > ! /* If we are doing SLP then the accesses need not have the
> > ! same alignment, instead it depends on the SLP group size. */
> > ! if (loop_vinfo
> > ! && STMT_SLP_TYPE (stmt_info)
> > ! && (LOOP_VINFO_VECT_FACTOR (loop_vinfo)
> > ! * GROUP_SIZE (vinfo_for_stmt (GROUP_FIRST_ELEMENT
> > (stmt_info))))
> > ! % TYPE_VECTOR_SUBPARTS (vectype) != 0)
>
> Parentheses here look wrong. Should be one fewer ending paren on the "*
> GROUP_SIZE" line and one more on the following line, right?
Yeah, though it shouldn't matter in practice. I'll fix things up
before committing.
Richard.
> Bill
>
> > ! ;
> > ! else if (!loop_vinfo
> > ! || (nested_in_vect_loop
> > ! && (TREE_INT_CST_LOW (DR_STEP (dr))
> > ! != GET_MODE_SIZE (TYPE_MODE (vectype)))))
> > return dr_explicit_realign;
> > else
> > return dr_explicit_realign_optimized;
> > Index: gcc/testsuite/gcc.dg/vect/O3-pr70130.c
> > ===================================================================
> > *** gcc/testsuite/gcc.dg/vect/O3-pr70130.c (revision 0)
> > --- gcc/testsuite/gcc.dg/vect/O3-pr70130.c (working copy)
> > ***************
> > *** 0 ****
> > --- 1,94 ----
> > + /* { dg-do run } */
> > + /* { dg-require-effective-target vsx_hw { target powerpc*-*-* } } */
> > + /* { dg-additional-options "-mcpu=power7" { target powerpc*-*-* } } */
> > +
> > + struct foo
> > + {
> > + short a[3][16][16];
> > + short pad;
> > + } images[8];
> > +
> > + void __attribute__ ((noinline, noclone))
> > + Loop_err (struct foo *img, const int s[16][2], int s0)
> > + {
> > + int i, j;
> > +
> > + for (j = 0; j < 16; j++)
> > + {
> > + for (i=0; i < 16; i++)
> > + {
> > + img->a[0][j][i] = s[i][0];
> > + img->a[1][j][i] = s[j][1];
> > + img->a[2][j][i] = s0;
> > + }
> > + }
> > + }
> > +
> > + const int s[16][2] = { { 1, 16 }, { 2, 15 }, { 3, 14 }, { 4, 13 },
> > + { 5, 12 }, { 6, 11 }, { 7, 10 }, { 8, 9 },
> > + { 9, 8 }, { 10, 7 }, { 11, 6 }, { 12, 5 },
> > + { 13, 4 }, { 14, 3 }, { 15, 2 }, { 16, 1 } };
> > + const struct foo expected
> > + = { { { { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 },
> > + { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 } },
> > + { { 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16
> > },
> > + { 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15 },
> > + { 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14 },
> > + { 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13 },
> > + { 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12 },
> > + { 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11 },
> > + { 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10 },
> > + { 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9 },
> > + { 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8 },
> > + { 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7 },
> > + { 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6 },
> > + { 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5 },
> > + { 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4 },
> > + { 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3 },
> > + { 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2 },
> > + { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 } },
> > + { { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1
> > },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 },
> > + { -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 } } },
> > + 0 };
> > +
> > + int
> > + main (void)
> > + {
> > + int i;
> > +
> > + for (i = 0; i < 8; i++)
> > + Loop_err (images + i, s, -1);
> > +
> > + for (i = 0; i < 8; i++)
> > + if (__builtin_memcmp (&expected, images + i, sizeof (expected)))
> > + __builtin_abort ();
> > + return 0;
> > + }
> >
>
>
>
--
Richard Biener <rguent...@suse.de>
SUSE LINUX GmbH, GF: Felix Imendoerffer, Jane Smithard, Graham Norton, HRB
21284 (AG Nuernberg)
