On Wed, Sep 1, 2021 at 8:28 AM Hongtao Liu <crazy...@gmail.com> wrote:
>
> On Tue, Aug 31, 2021 at 7:56 PM Richard Biener
> <richard.guent...@gmail.com> wrote:
> >
> > On Tue, Aug 31, 2021 at 12:18 PM Hongtao Liu <crazy...@gmail.com> wrote:
> > >
> > > On Mon, Aug 30, 2021 at 8:25 PM Richard Biener via Gcc-patches
> > > <gcc-patches@gcc.gnu.org> wrote:
> > > >
> > > > On Fri, Aug 27, 2021 at 8:53 AM liuhongt <hongtao....@intel.com> wrote:
> > > > >
> > > > > When gimple simplifcation try to combine op and vec_cond_expr to
> > > > > cond_op,
> > > > > it doesn't check if mask type matches. It causes an ICE when expand
> > > > > cond_op
> > > > > with mismatched mode.
> > > > > This patch add a function named
> > > > > cond_vectorized_internal_fn_supported_p
> > > > > to additionally check mask type than
> > > > > vectorized_internal_fn_supported_p.
> > > > >
> > > > > Bootstrapped and regtested on x86_64-linux-gnu{-m32,}.
> > > > > Ok for trunk?
> > > > >
> > > > > gcc/ChangeLog:
> > > > >
> > > > > PR middle-end/102080
> > > > > * internal-fn.c (cond_vectorized_internal_fn_supported_p):
> > > > > New functions.
> > > > > * internal-fn.h (cond_vectorized_internal_fn_supported_p):
> > > > > New declaration.
> > > > > * match.pd: Check the type of mask while generating cond_op in
> > > > > gimple simplication.
> > > > >
> > > > > gcc/testsuite/ChangeLog:
> > > > >
> > > > > PR middle-end/102080
> > > > > * gcc.target/i386/pr102080.c: New test.
> > > > > ---
> > > > > gcc/internal-fn.c | 22 ++++++++++++++++++++++
> > > > > gcc/internal-fn.h | 1 +
> > > > > gcc/match.pd | 24
> > > > > ++++++++++++++++--------
> > > > > gcc/testsuite/gcc.target/i386/pr102080.c | 16 ++++++++++++++++
> > > > > 4 files changed, 55 insertions(+), 8 deletions(-)
> > > > > create mode 100644 gcc/testsuite/gcc.target/i386/pr102080.c
> > > > >
> > > > > diff --git a/gcc/internal-fn.c b/gcc/internal-fn.c
> > > > > index 1360a00f0b9..8b2b65db1a7 100644
> > > > > --- a/gcc/internal-fn.c
> > > > > +++ b/gcc/internal-fn.c
> > > > > @@ -4102,6 +4102,28 @@ expand_internal_call (gcall *stmt)
> > > > > expand_internal_call (gimple_call_internal_fn (stmt), stmt);
> > > > > }
> > > > >
> > > > > +/* Check cond_op for vector modes since
> > > > > vectorized_internal_fn_supported_p
> > > > > + doesn't check if mask type matches. */
> > > > > +bool
> > > > > +cond_vectorized_internal_fn_supported_p (internal_fn ifn, tree type,
> > > > > + tree mask_type)
> > > > > +{
> > > > > + if (!vectorized_internal_fn_supported_p (ifn, type))
> > > > > + return false;
> > > > > +
> > > > > + machine_mode mask_mode;
> > > > > + machine_mode vmode = TYPE_MODE (type);
> > > > > + int size1, size2;
> > > > > + if (VECTOR_MODE_P (vmode)
> > > > > + && targetm.vectorize.get_mask_mode (vmode).exists(&mask_mode)
> > > > > + && GET_MODE_SIZE (mask_mode).is_constant (&size1)
> > > > > + && GET_MODE_SIZE (TYPE_MODE (mask_type)).is_constant (&size2)
> > > > > + && size1 != size2)
> > > >
> > > > Why do we check for equal size rather than just mode equality which
> > > I originally thought TYPE_MODE of vector(8) <signed-boolean:1> was
> > > not QImode, Changed the patch to check mode equality.
> > > Update patch.
> >
> > Looking at all this it seems the match.pd patterns should have not
> > used vectorized_internal_fn_supported_p but direct_internal_fn_supported_p
> > which is equivalent here because we're always working with vector modes?
> >
> > And then shouldn't we look at the actual optab whether the mask mode matches
> > the expectation rather than going around via the target hook which may not
> > have
> > enough context to decide which mask mode to use?
> How about this?
>
> +/* Return true if target supports cond_op with data TYPE and
> + mask MASK_TYPE. */
> +bool
> +cond_internal_fn_supported_p (internal_fn ifn, tree type,
> + tree mask_type)
> +{
> + tree_pair types = tree_pair (type, type);
> + optab tmp = direct_internal_fn_optab (ifn, types);
> + machine_mode vmode = TYPE_MODE (type);
> + insn_code icode = direct_optab_handler (tmp, vmode);
> + if (icode == CODE_FOR_nothing)
> + return false;
> +
> + machine_mode mask_mode = TYPE_MODE (mask_type);
> + /* Can't create rtx and use insn_operand_matches here. */
> + return insn_data[icode].operand[0].mode == vmode
> + && insn_data[icode].operand[1].mode == mask_mode;
> +}
> +
Yeah, sth like that, though the operand[0].mode test should be
redudnant. I think we should assert or have a whiltelist
for the internal function we support to be queried this way.
Not sure if we can directly access the 'cond_binary/cond_ternary'
classification used in internal-fn.def, that would be best.
Richard, what are your thoughts about all this?
Thanks,
Richard.
> Update patch
>
> >
> > In any case if the approach of the patch is correct shouldn't it do
> >
> > if (VECTOR_MODE_P (vmode)
> > && (!targetm.vectorize.get_mask_mode (vmode).exists(&mask_mode)
> > || mask_mode != TYPE_MODE (mask_type)))
> > return false;
> >
> > that is, not return true if there's no mask mode for the data mode?
> >
> > Given the first observation should we call the function
> > direct_cond_internal_fn_supported_p () instead and as to the second
> > observation, look at the optab operands mode?
> >
> > Richard.
> >
> > > > I think would work for non-constant sized modes as well? And when
> > > > using sizes you'd instead use maybe_ne (GET_MODE_SIZE (mask_mode),
> > > > GET_MODE_SIZE (TYPE_MODE (mask_type)))
> > > >
> > > > Thanks,
> > > > Richard.
> > > >
> > > > > + return false;
> > > > > +
> > > > > + return true;
> > > > > +}
> > > > > +
> > > > > /* If TYPE is a vector type, return true if IFN is a direct internal
> > > > > function that is supported for that type. If TYPE is a scalar
> > > > > type,
> > > > > return true if IFN is a direct internal function that is
> > > > > supported for
> > > > > diff --git a/gcc/internal-fn.h b/gcc/internal-fn.h
> > > > > index 19d0f849a5a..f0aea00103c 100644
> > > > > --- a/gcc/internal-fn.h
> > > > > +++ b/gcc/internal-fn.h
> > > > > @@ -236,5 +236,6 @@ extern void expand_PHI (internal_fn, gcall *);
> > > > > extern void expand_SHUFFLEVECTOR (internal_fn, gcall *);
> > > > >
> > > > > extern bool vectorized_internal_fn_supported_p (internal_fn, tree);
> > > > > +extern bool cond_vectorized_internal_fn_supported_p (internal_fn,
> > > > > tree, tree);
> > > > >
> > > > > #endif
> > > > > diff --git a/gcc/match.pd b/gcc/match.pd
> > > > > index e5bbb123a6a..72b1bc674db 100644
> > > > > --- a/gcc/match.pd
> > > > > +++ b/gcc/match.pd
> > > > > @@ -6987,14 +6987,18 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
> > > > > cond_op (COND_BINARY)
> > > > > (simplify
> > > > > (vec_cond @0 (view_convert? (uncond_op@4 @1 @2)) @3)
> > > > > - (with { tree op_type = TREE_TYPE (@4); }
> > > > > - (if (vectorized_internal_fn_supported_p (as_internal_fn
> > > > > (cond_op), op_type)
> > > > > + (with { tree op_type = TREE_TYPE (@4);
> > > > > + tree mask_type = TREE_TYPE (@0); }
> > > > > + (if (cond_vectorized_internal_fn_supported_p (as_internal_fn
> > > > > (cond_op),
> > > > > + op_type, mask_type)
> > > > > && element_precision (type) == element_precision (op_type))
> > > > > (view_convert (cond_op @0 @1 @2 (view_convert:op_type @3))))))
> > > > > (simplify
> > > > > (vec_cond @0 @1 (view_convert? (uncond_op@4 @2 @3)))
> > > > > - (with { tree op_type = TREE_TYPE (@4); }
> > > > > - (if (vectorized_internal_fn_supported_p (as_internal_fn
> > > > > (cond_op), op_type)
> > > > > + (with { tree op_type = TREE_TYPE (@4);
> > > > > + tree mask_type = TREE_TYPE (@0); }
> > > > > + (if (cond_vectorized_internal_fn_supported_p (as_internal_fn
> > > > > (cond_op),
> > > > > + op_type, mask_type)
> > > > > && element_precision (type) == element_precision (op_type))
> > > > > (view_convert (cond_op (bit_not @0) @2 @3 (view_convert:op_type
> > > > > @1)))))))
> > > > >
> > > > > @@ -7003,14 +7007,18 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
> > > > > cond_op (COND_TERNARY)
> > > > > (simplify
> > > > > (vec_cond @0 (view_convert? (uncond_op@5 @1 @2 @3)) @4)
> > > > > - (with { tree op_type = TREE_TYPE (@5); }
> > > > > - (if (vectorized_internal_fn_supported_p (as_internal_fn
> > > > > (cond_op), op_type)
> > > > > + (with { tree op_type = TREE_TYPE (@5);
> > > > > + tree mask_type = TREE_TYPE (@0); }
> > > > > + (if (cond_vectorized_internal_fn_supported_p (as_internal_fn
> > > > > (cond_op),
> > > > > + op_type, mask_type)
> > > > > && element_precision (type) == element_precision (op_type))
> > > > > (view_convert (cond_op @0 @1 @2 @3 (view_convert:op_type @4))))))
> > > > > (simplify
> > > > > (vec_cond @0 @1 (view_convert? (uncond_op@5 @2 @3 @4)))
> > > > > - (with { tree op_type = TREE_TYPE (@5); }
> > > > > - (if (vectorized_internal_fn_supported_p (as_internal_fn
> > > > > (cond_op), op_type)
> > > > > + (with { tree op_type = TREE_TYPE (@5);
> > > > > + tree mask_type = TREE_TYPE (@0); }
> > > > > + (if (cond_vectorized_internal_fn_supported_p (as_internal_fn
> > > > > (cond_op),
> > > > > + op_type, mask_type)
> > > > > && element_precision (type) == element_precision (op_type))
> > > > > (view_convert (cond_op (bit_not @0) @2 @3 @4
> > > > > (view_convert:op_type @1)))))))
> > > > > diff --git a/gcc/testsuite/gcc.target/i386/pr102080.c
> > > > > b/gcc/testsuite/gcc.target/i386/pr102080.c
> > > > > new file mode 100644
> > > > > index 00000000000..6a40a75e1c5
> > > > > --- /dev/null
> > > > > +++ b/gcc/testsuite/gcc.target/i386/pr102080.c
> > > > > @@ -0,0 +1,16 @@
> > > > > +#include<immintrin.h>
> > > > > +typedef float __m256 __attribute__((__vector_size__(32)));
> > > > > +__m256 _mm256_blendv_ps___Y, _mm256_blendv_ps___M, _mm256_mul_ps___A,
> > > > > + _mm256_mul_ps___B, IfThenElse___trans_tmp_9;
> > > > > +
> > > > > +void
> > > > > +__attribute__ ((target("avx")))
> > > > > +IfThenElse (__m256 no) {
> > > > > + IfThenElse___trans_tmp_9 = _mm256_blendv_ps (no,
> > > > > _mm256_blendv_ps___Y, _mm256_blendv_ps___M);
> > > > > +}
> > > > > +void
> > > > > +__attribute__ ((target("avx512vl")))
> > > > > +EncodedFromDisplay() {
> > > > > + __m256 __trans_tmp_11 = _mm256_mul_ps___A * _mm256_mul_ps___B;
> > > > > + IfThenElse(__trans_tmp_11);
> > > > > +}
> > > > > --
> > > > > 2.18.1
> > > > >
> > >
> > >
> > >
> > > --
> > > BR,
> > > Hongtao
>
>
>
> --
> BR,
> Hongtao
