Re: [PATCH V3] VECT: Add decrement IV iteration loop control by variable amount support

[Richard Biener <rguenther@suse.de>]

On Wed, 26 Apr 2023, juzhe.zhong@rivai.ai wrote:

> From: Juzhe-Zhong <juzhe.zhong@rivai.ai>
> 
> This patch address all comments from Richard && Bernhard.
> Fix all codes following their comments:
> 
> 1. Remove ifn-specific handling of WHILE_LEN, so there is no change
>    in internal-fn.cc in this patch.
> 
> 2. Add comments for changing interface of "create_iv" function.
> 
> 3. Reorder the CODE of interface "create_iv", make CODE argument between
>    base and step.
> 
> 4. Fix the misleading and incorrect comments of "vect_set_loop_controls_by_while_len".
> 
> 5. remove redundant "enum" of "create_iv" interface and function.
> 
> Feel free to comment if I meet something.
> 
> Thanks.
> 
> gcc/ChangeLog:
> 
>         * cfgloopmanip.cc (create_empty_loop_on_edge): Adapt create_iv interface change.
>         * doc/md.texi: Add WHILE_LEN pattern description.
>         * gimple-loop-interchange.cc (tree_loop_interchange::map_inductions_to_loop): Adapt create_iv interface change.
>         * internal-fn.def (WHILE_LEN): New ifn.
>         * optabs.def (OPTAB_D): New optab pattern.
>         * tree-ssa-loop-ivcanon.cc (create_canonical_iv): Adapt create_iv interface change.
>         * tree-ssa-loop-ivopts.cc (create_new_iv): Ditto.
>         * tree-ssa-loop-manip.cc (create_iv): Ditto.
>         (tree_transform_and_unroll_loop): Ditto.
>         (canonicalize_loop_ivs): Ditto.
>         * tree-ssa-loop-manip.h (create_iv): Add code in the interface.
>         * tree-vect-data-refs.cc (vect_create_data_ref_ptr): New function.
>         * tree-vect-loop-manip.cc (vect_set_loop_controls_directly): Add decrementing IV loop control support.
>         (vect_set_loop_controls_by_while_len): New function.
>         (vect_set_loop_condition_partial_vectors): Add decrementing IV loop control support.
>         (vect_set_loop_condition_normal): Adapt create_iv interface change.
>         * tree-vect-loop.cc (vect_create_epilog_for_reduction): Ditto.
>         (vect_get_loop_len): Adapt SLP auto-vectorization for WHILE_LEN.
>         * tree-vect-stmts.cc (get_while_len_data_ref_ptr): New function.
>         (vectorizable_store): Add WHILE_LEN data reference pointer IV.
>         (vectorizable_load): Ditto.
>         * tree-vectorizer.h (vect_get_loop_len): Add new arguments to interface.
> 
> 
> ---
>  gcc/cfgloopmanip.cc            |   2 +-
>  gcc/doc/md.texi                |  34 ++++
>  gcc/gimple-loop-interchange.cc |   2 +-
>  gcc/internal-fn.def            |   1 +
>  gcc/optabs.def                 |   1 +
>  gcc/tree-ssa-loop-ivcanon.cc   |   2 +-
>  gcc/tree-ssa-loop-ivopts.cc    |   2 +-
>  gcc/tree-ssa-loop-manip.cc     |  18 ++-
>  gcc/tree-ssa-loop-manip.h      |   4 +-
>  gcc/tree-vect-data-refs.cc     |   8 +-
>  gcc/tree-vect-loop-manip.cc    | 280 +++++++++++++++++++++++++++++++--
>  gcc/tree-vect-loop.cc          |  29 +++-
>  gcc/tree-vect-stmts.cc         |  95 ++++++++++-
>  gcc/tree-vectorizer.h          |   4 +-
>  14 files changed, 445 insertions(+), 37 deletions(-)
> 
> diff --git a/gcc/cfgloopmanip.cc b/gcc/cfgloopmanip.cc
> index 0e3ad8ed742..6e09dcbb0b1 100644
> --- a/gcc/cfgloopmanip.cc
> +++ b/gcc/cfgloopmanip.cc
> @@ -826,7 +826,7 @@ create_empty_loop_on_edge (edge entry_edge,
>      }
>  
>    gsi = gsi_last_bb (loop_header);
> -  create_iv (initial_value, stride, iv, loop, &gsi, false,
> +  create_iv (initial_value, PLUS_EXPR, stride, iv, loop, &gsi, false,
>  	     iv_before, iv_after);
>  
>    /* Insert loop exit condition.  */
> diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
> index 07bf8bdebff..307d3c407d8 100644
> --- a/gcc/doc/md.texi
> +++ b/gcc/doc/md.texi
> @@ -4965,6 +4965,40 @@ for (i = 1; i < operand3; i++)
>    operand0[i] = operand0[i - 1] && (operand1 + i < operand2);
>  @end smallexample
>  
> +@cindex @code{while_len@var{m}@var{n}} instruction pattern
> +@item @code{while_len@var{m}@var{n}}
> +Set operand 0 to the number of active elements in vector will be updated value.
> +operand 1 is the total elements need to be updated value.
> +operand 2 is the vectorization factor.
> +The value of operand 0 is target dependent and flexible in each iteration.

I can't really parse the above - there's information missing on what
the modes M and N are and how they relate.  I suppose you take the
while_ult description and mangle it to the while_len semantics?

> +The operation of this pattern can be:
> +
> +@smallexample
> +Case 1:
> +operand0 = MIN (operand1, operand2);
> +operand2 can be const_poly_int or poly_int related to vector mode size.
> +Some target like RISC-V has a standalone instruction to get MIN (n, MODE SIZE) so
> +that we can reduce a use of general purpose register.
> +
> +In this case, only the last iteration of the loop is partial iteration.
> +@end smallexample
> +
> +@smallexample
> +Case 2:
> +if (operand1 <= operand2)
> +  operand0 = operand1;
> +else if (operand1 < 2 * operand2)
> +  operand0 = IN_RANGE (ceil (operand1 / 2), operand2);
> +else
> +  operand0 = operand2;
> +
> +This case will evenly distribute work over the last 2 iterations of a stripmine loop.
> +@end smallexample
> +
> +The output of this pattern is not only used as IV of loop control counter, but also
> +is used as the IV of address calculation with multiply/shift operation. This allow
> +us dynamic adjust the number of elements is processed in each iteration of the loop.
> +
>  @cindex @code{check_raw_ptrs@var{m}} instruction pattern
>  @item @samp{check_raw_ptrs@var{m}}
>  Check whether, given two pointers @var{a} and @var{b} and a length @var{len},
> diff --git a/gcc/gimple-loop-interchange.cc b/gcc/gimple-loop-interchange.cc
> index 1b77bfd46b2..e5590374e59 100644
> --- a/gcc/gimple-loop-interchange.cc
> +++ b/gcc/gimple-loop-interchange.cc
> @@ -1185,7 +1185,7 @@ tree_loop_interchange::map_inductions_to_loop (loop_cand &src, loop_cand &tgt)
>  	  tree var_before, var_after;
>  	  tree base = unshare_expr (iv->init_expr);
>  	  tree step = unshare_expr (iv->step);
> -	  create_iv (base, step, SSA_NAME_VAR (iv->var),
> +	  create_iv (base, PLUS_EXPR, step, SSA_NAME_VAR (iv->var),
>  		     tgt.m_loop, &incr_pos, false, &var_before, &var_after);
>  	  bitmap_set_bit (m_dce_seeds, SSA_NAME_VERSION (var_before));
>  	  bitmap_set_bit (m_dce_seeds, SSA_NAME_VERSION (var_after));
> diff --git a/gcc/internal-fn.def b/gcc/internal-fn.def
> index 7fe742c2ae7..9f7269f8f60 100644
> --- a/gcc/internal-fn.def
> +++ b/gcc/internal-fn.def
> @@ -153,6 +153,7 @@ DEF_INTERNAL_OPTAB_FN (VEC_SET, 0, vec_set, vec_set)
>  DEF_INTERNAL_OPTAB_FN (LEN_STORE, 0, len_store, len_store)
>  
>  DEF_INTERNAL_OPTAB_FN (WHILE_ULT, ECF_CONST | ECF_NOTHROW, while_ult, while)
> +DEF_INTERNAL_OPTAB_FN (WHILE_LEN, ECF_CONST | ECF_NOTHROW, while_len, binary)
>  DEF_INTERNAL_OPTAB_FN (CHECK_RAW_PTRS, ECF_CONST | ECF_NOTHROW,
>  		       check_raw_ptrs, check_ptrs)
>  DEF_INTERNAL_OPTAB_FN (CHECK_WAR_PTRS, ECF_CONST | ECF_NOTHROW,
> diff --git a/gcc/optabs.def b/gcc/optabs.def
> index 695f5911b30..f5938bd2c24 100644
> --- a/gcc/optabs.def
> +++ b/gcc/optabs.def
> @@ -476,3 +476,4 @@ OPTAB_DC (vec_series_optab, "vec_series$a", VEC_SERIES)
>  OPTAB_D (vec_shl_insert_optab, "vec_shl_insert_$a")
>  OPTAB_D (len_load_optab, "len_load_$a")
>  OPTAB_D (len_store_optab, "len_store_$a")
> +OPTAB_D (while_len_optab, "while_len$a")

But your description mentions two modes?

> diff --git a/gcc/tree-ssa-loop-ivcanon.cc b/gcc/tree-ssa-loop-ivcanon.cc
> index 9f72d534b7c..6941d0a4d1d 100644
> --- a/gcc/tree-ssa-loop-ivcanon.cc
> +++ b/gcc/tree-ssa-loop-ivcanon.cc
> @@ -113,7 +113,7 @@ create_canonical_iv (class loop *loop, edge exit, tree niter,
>  		       niter,
>  		       build_int_cst (type, 1));
>    incr_at = gsi_last_bb (in->src);
> -  create_iv (niter,
> +  create_iv (niter, PLUS_EXPR,
>  	     build_int_cst (type, -1),
>  	     NULL_TREE, loop,
>  	     &incr_at, false, var_before, &var);
> diff --git a/gcc/tree-ssa-loop-ivopts.cc b/gcc/tree-ssa-loop-ivopts.cc
> index 0dd47910f97..a7377f91f9e 100644
> --- a/gcc/tree-ssa-loop-ivopts.cc
> +++ b/gcc/tree-ssa-loop-ivopts.cc
> @@ -7267,7 +7267,7 @@ create_new_iv (struct ivopts_data *data, struct iv_cand *cand)
>  
>    base = unshare_expr (cand->iv->base);
>  
> -  create_iv (base, unshare_expr (cand->iv->step),
> +  create_iv (base, PLUS_EXPR, unshare_expr (cand->iv->step),
>  	     cand->var_before, data->current_loop,
>  	     &incr_pos, after, &cand->var_before, &cand->var_after);
>  }
> diff --git a/gcc/tree-ssa-loop-manip.cc b/gcc/tree-ssa-loop-manip.cc
> index a52277abdbf..b0d254b0c89 100644
> --- a/gcc/tree-ssa-loop-manip.cc
> +++ b/gcc/tree-ssa-loop-manip.cc
> @@ -47,7 +47,9 @@ along with GCC; see the file COPYING3.  If not see
>     so that we can free them all at once.  */
>  static bitmap_obstack loop_renamer_obstack;
>  
> -/* Creates an induction variable with value BASE + STEP * iteration in LOOP.
> +/* Creates an induction variable with value BASE (+/-) STEP * iteration in LOOP.
> +   If CODE is PLUS_EXPR, the induction variable is BASE + STEP * iteration.
> +   If CODE is MINUS_EXPR, the induction variable is BASE - STEP * iteration.
>     It is expected that neither BASE nor STEP are shared with other expressions
>     (unless the sharing rules allow this).  Use VAR as a base var_decl for it
>     (if NULL, a new temporary will be created).  The increment will occur at
> @@ -57,8 +59,8 @@ static bitmap_obstack loop_renamer_obstack;
>     VAR_AFTER (unless they are NULL).  */
>  
>  void
> -create_iv (tree base, tree step, tree var, class loop *loop,
> -	   gimple_stmt_iterator *incr_pos, bool after,
> +create_iv (tree base, tree_code code, tree step, tree var,
> +	   class loop *loop, gimple_stmt_iterator *incr_pos, bool after,
>  	   tree *var_before, tree *var_after)
>  {
>    gassign *stmt;
> @@ -66,7 +68,9 @@ create_iv (tree base, tree step, tree var, class loop *loop,
>    tree initial, step1;
>    gimple_seq stmts;
>    tree vb, va;
> -  enum tree_code incr_op = PLUS_EXPR;
> +  /* The code can only be PLUS_EXPR or MINUS_EXPR.  */
> +  gcc_assert (code == PLUS_EXPR || code == MINUS_EXPR);

Ah, you already have an assert.

> +  tree_code incr_op = code;

If you name the parameter incr_op you can get rid of this assignment.

>    edge pe = loop_preheader_edge (loop);
>  
>    if (var != NULL_TREE)
> @@ -1368,7 +1372,7 @@ tree_transform_and_unroll_loop (class loop *loop, unsigned factor,
>        tree ctr_before, ctr_after;
>        gimple_stmt_iterator bsi = gsi_last_nondebug_bb (new_exit->src);
>        exit_if = as_a <gcond *> (gsi_stmt (bsi));
> -      create_iv (exit_base, exit_step, NULL_TREE, loop,
> +      create_iv (exit_base, PLUS_EXPR, exit_step, NULL_TREE, loop,
>  		 &bsi, false, &ctr_before, &ctr_after);
>        gimple_cond_set_code (exit_if, exit_cmp);
>        gimple_cond_set_lhs (exit_if, ctr_after);
> @@ -1583,8 +1587,8 @@ canonicalize_loop_ivs (class loop *loop, tree *nit, bool bump_in_latch)
>      gsi = gsi_last_bb (loop->latch);
>    else
>      gsi = gsi_last_nondebug_bb (loop->header);
> -  create_iv (build_int_cst_type (type, 0), build_int_cst (type, 1), NULL_TREE,
> -	     loop, &gsi, bump_in_latch, &var_before, NULL);
> +  create_iv (build_int_cst_type (type, 0), PLUS_EXPR, build_int_cst (type, 1),
> +	     NULL_TREE, loop, &gsi, bump_in_latch, &var_before, NULL);
>  
>    rewrite_all_phi_nodes_with_iv (loop, var_before);
>  
> diff --git a/gcc/tree-ssa-loop-manip.h b/gcc/tree-ssa-loop-manip.h
> index d49273a3987..bda09f51d56 100644
> --- a/gcc/tree-ssa-loop-manip.h
> +++ b/gcc/tree-ssa-loop-manip.h
> @@ -22,8 +22,8 @@ along with GCC; see the file COPYING3.  If not see
>  
>  typedef void (*transform_callback)(class loop *, void *);
>  
> -extern void create_iv (tree, tree, tree, class loop *, gimple_stmt_iterator *,
> -		       bool, tree *, tree *);
> +extern void create_iv (tree, tree_code, tree, tree, class loop *,
> +		       gimple_stmt_iterator *, bool, tree *, tree *);
>  extern void rewrite_into_loop_closed_ssa (bitmap, unsigned);
>  extern void verify_loop_closed_ssa (bool, class loop * = NULL);
>  
> diff --git a/gcc/tree-vect-data-refs.cc b/gcc/tree-vect-data-refs.cc
> index c03ffb3aaf1..355ede5c0d9 100644
> --- a/gcc/tree-vect-data-refs.cc
> +++ b/gcc/tree-vect-data-refs.cc
> @@ -5101,7 +5101,7 @@ vect_create_data_ref_ptr (vec_info *vinfo, stmt_vec_info stmt_info,
>  
>        standard_iv_increment_position (loop, &incr_gsi, &insert_after);
>  
> -      create_iv (aggr_ptr_init,
> +      create_iv (aggr_ptr_init, PLUS_EXPR,
>  		 fold_convert (aggr_ptr_type, iv_step),
>  		 aggr_ptr, loop, &incr_gsi, insert_after,
>  		 &indx_before_incr, &indx_after_incr);
> @@ -5131,9 +5131,9 @@ vect_create_data_ref_ptr (vec_info *vinfo, stmt_vec_info stmt_info,
>      {
>        standard_iv_increment_position (containing_loop, &incr_gsi,
>  				      &insert_after);
> -      create_iv (aptr, fold_convert (aggr_ptr_type, DR_STEP (dr)), aggr_ptr,
> -		 containing_loop, &incr_gsi, insert_after, &indx_before_incr,
> -		 &indx_after_incr);
> +      create_iv (aptr, PLUS_EXPR, fold_convert (aggr_ptr_type, DR_STEP (dr)),
> +		 aggr_ptr, containing_loop, &incr_gsi, insert_after,
> +		 &indx_before_incr, &indx_after_incr);
>        incr = gsi_stmt (incr_gsi);
>  
>        /* Copy the points-to information if it exists. */
> diff --git a/gcc/tree-vect-loop-manip.cc b/gcc/tree-vect-loop-manip.cc
> index f60fa50e8f4..6973314adb5 100644
> --- a/gcc/tree-vect-loop-manip.cc
> +++ b/gcc/tree-vect-loop-manip.cc
> @@ -468,8 +468,9 @@ vect_set_loop_controls_directly (class loop *loop, loop_vec_info loop_vinfo,
>    gimple_stmt_iterator incr_gsi;
>    bool insert_after;
>    standard_iv_increment_position (loop, &incr_gsi, &insert_after);
> -  create_iv (build_int_cst (iv_type, 0), nitems_step, NULL_TREE, loop,
> -	     &incr_gsi, insert_after, &index_before_incr, &index_after_incr);
> +  create_iv (build_int_cst (iv_type, 0), PLUS_EXPR, nitems_step, NULL_TREE,
> +	     loop, &incr_gsi, insert_after, &index_before_incr,
> +	     &index_after_incr);
>  
>    tree zero_index = build_int_cst (compare_type, 0);
>    tree test_index, test_limit, first_limit;
> @@ -682,6 +683,260 @@ vect_set_loop_controls_directly (class loop *loop, loop_vec_info loop_vinfo,
>    return next_ctrl;
>  }
>  
> +/* Helper for vect_set_loop_condition_partial_vectors.  Generate definitions
> +   for all the rgroup controls in RGC and return a control that is nonzero
> +   when the loop needs to iterate.  Add any new preheader statements to
> +   PREHEADER_SEQ.  Use LOOP_COND_GSI to insert code before the exit gcond.
> +
> +   RGC belongs to loop LOOP.  The loop originally iterated NITERS
> +   times and has been vectorized according to LOOP_VINFO.
> +
> +   Unlike vect_set_loop_controls_directly which is iterating from 0-based IV
> +   to TEST_LIMIT - bias.
> +
> +   In vect_set_loop_controls_by_while_len, we are iterating from start at
> +   IV = TEST_LIMIT - bias and keep subtract IV by the length calculated by
> +   IFN_WHILE_LEN pattern.
> +
> +   1. Single rgroup, the Gimple IR should be:
> +
> +	# vectp_B.6_8 = PHI <vectp_B.6_13(6), &B(5)>
> +	# vectp_B.8_16 = PHI <vectp_B.8_17(6), &B(5)>
> +	# vectp_A.11_19 = PHI <vectp_A.11_20(6), &A(5)>
> +	# vectp_A.13_22 = PHI <vectp_A.13_23(6), &A(5)>
> +	# ivtmp_26 = PHI <ivtmp_27(6), _25(5)>
> +	_28 = .WHILE_LEN (ivtmp_26, POLY_INT_CST [4, 4]);
> +	ivtmp_15 = _28 * 4;
> +	vect__1.10_18 = .LEN_LOAD (vectp_B.8_16, 128B, _28, 0);
> +	_1 = B[i_10];
> +	.LEN_STORE (vectp_A.13_22, 128B, _28, vect__1.10_18, 0);
> +	i_7 = i_10 + 1;
> +	vectp_B.8_17 = vectp_B.8_16 + ivtmp_15;
> +	vectp_A.13_23 = vectp_A.13_22 + ivtmp_15;
> +	ivtmp_27 = ivtmp_26 - _28;
> +	if (ivtmp_27 != 0)
> +	  goto <bb 6>; [83.33%]
> +	else
> +	  goto <bb 7>; [16.67%]
> +
> +   Note: We use the outcome of .WHILE_LEN to adjust both loop control IV and
> +   data reference pointer IV.
> +
> +   1). The result of .WHILE_LEN:
> +       _28 = .WHILE_LEN (ivtmp_26, POLY_INT_CST [4, 4]);
> +       The _28 is not necessary to be VF in any iteration, instead, we allow
> +       _28 to be any value as long as it it <= VF. Such flexible WHILE_LEN

"it is", but I don't really understand what you are trying to say?  Is it
that loop control cannot be implemented as

        if (_28 != POLY_INT_CST [4, 4])
          next iteration will be the last;

but the result can only be used to decrement the remaining scalar
iterations?  That is, .WHILE_LEN (a, b) doesn't actually subtract
'b' but may choose to subtract any value in [1, b]?

How can this work when the vectorizer "knows" that 'b' covers two
vectors and thus it emits two vector stmts?

> +       pattern allows target have various flexible optimizations in vector
> +       loop iterations. Target like RISC-V has special application vector
> +       length calculation instruction which will distribute even workload
> +       in the last 2 iterations.
> +
> +       Other example is that we can allow even generate _28 <= VF / 2 so
> +       that some machine can run vector codes in low power mode.
> +
> +   2). Loop control IV:
> +       ivtmp_27 = ivtmp_26 - _28;
> +       if (ivtmp_27 != 0)
> +	 goto <bb 6>; [83.33%]
> +       else
> +	 goto <bb 7>; [16.67%]
> +
> +       This is the saturating-subtraction towards zero, the outcome of
> +       .WHILE_LEN wil make ivtmp_27 never underflow zero.

So it's not actually saturating.  The saturating operation is
done by .WHILE_LEN?

> +   3). Data reference pointer IV:
> +       ivtmp_15 = _28 * 4;
> +       vectp_B.8_17 = vectp_B.8_16 + ivtmp_15;
> +       vectp_A.13_23 = vectp_A.13_22 + ivtmp_15;
> +
> +       The pointer IV is adjusted accurately according to the .WHILE_LEN.

Note that this will be bad for IVOPTs, it would  be _way_ better if
the increment were (poly-)constant.  For powerpc this would mean
complex addressing modes cannot be used (not sure if they have those).

> +   2. Multiple rgroup, the Gimple IR should be:
> +
> +	# i_23 = PHI <i_20(6), 0(11)>
> +	# vectp_f.8_51 = PHI <vectp_f.8_52(6), f_15(D)(11)>
> +	# vectp_d.10_59 = PHI <vectp_d.10_60(6), d_18(D)(11)>
> +	# ivtmp_70 = PHI <ivtmp_71(6), _69(11)>
> +	# ivtmp_73 = PHI <ivtmp_74(6), _67(11)>
> +	_72 = MIN_EXPR <ivtmp_70, 16>;
> +	_75 = MIN_EXPR <ivtmp_73, 16>;
> +	_1 = i_23 * 2;
> +	_2 = (long unsigned int) _1;
> +	_3 = _2 * 2;
> +	_4 = f_15(D) + _3;
> +	_5 = _2 + 1;
> +	_6 = _5 * 2;
> +	_7 = f_15(D) + _6;
> +	.LEN_STORE (vectp_f.8_51, 128B, _75, { 1, 2, 1, 2, 1, 2, 1, 2 }, 0);
> +	vectp_f.8_56 = vectp_f.8_51 + 16;
> +	.LEN_STORE (vectp_f.8_56, 128B, _72, { 1, 2, 1, 2, 1, 2, 1, 2 }, 0);
> +	_8 = (long unsigned int) i_23;
> +	_9 = _8 * 4;
> +	_10 = d_18(D) + _9;
> +	_61 = _75 / 2;
> +	.LEN_STORE (vectp_d.10_59, 128B, _61, { 3, 3, 3, 3 }, 0);
> +	vectp_d.10_63 = vectp_d.10_59 + 16;
> +	_64 = _72 / 2;
> +	.LEN_STORE (vectp_d.10_63, 128B, _64, { 3, 3, 3, 3 }, 0);
> +	i_20 = i_23 + 1;
> +	vectp_f.8_52 = vectp_f.8_56 + 16;
> +	vectp_d.10_60 = vectp_d.10_63 + 16;
> +	ivtmp_74 = ivtmp_73 - _75;
> +	ivtmp_71 = ivtmp_70 - _72;
> +	if (ivtmp_74 != 0)
> +	  goto <bb 6>; [83.33%]
> +	else
> +	  goto <bb 13>; [16.67%]
> +
> +   Note: We DO NOT use .WHILE_LEN in SLP auto-vectorization for multiple
> +   rgroups. Instead, we use MIN_EXPR to guarantee we always use VF as the
> +   iteration amount for mutiple rgroups.

I see.  I wonder if it makes sense to leave .WHILE_LEN aside for a start,
the above scheme should also work for single rgroups, no?

> +   The analysis of the flow of multiple rgroups:
> +	_72 = MIN_EXPR <ivtmp_70, 16>;
> +	_75 = MIN_EXPR <ivtmp_73, 16>;
> +	...
> +	.LEN_STORE (vectp_f.8_51, 128B, _75, { 1, 2, 1, 2, 1, 2, 1, 2 }, 0);
> +	vectp_f.8_56 = vectp_f.8_51 + 16;
> +	.LEN_STORE (vectp_f.8_56, 128B, _72, { 1, 2, 1, 2, 1, 2, 1, 2 }, 0);
> +	...
> +	_61 = _75 / 2;
> +	.LEN_STORE (vectp_d.10_59, 128B, _61, { 3, 3, 3, 3 }, 0);
> +	vectp_d.10_63 = vectp_d.10_59 + 16;
> +	_64 = _72 / 2;
> +	.LEN_STORE (vectp_d.10_63, 128B, _64, { 3, 3, 3, 3 }, 0);
> +
> +  We use _72 = MIN_EXPR <ivtmp_70, 16>; to generate the number of the elements
> +  to be processed in each iteration.
> +
> +  The related STOREs:
> +    _72 = MIN_EXPR <ivtmp_70, 16>;
> +    .LEN_STORE (vectp_f.8_56, 128B, _72, { 1, 2, 1, 2, 1, 2, 1, 2 }, 0);
> +    _64 = _72 / 2;
> +    .LEN_STORE (vectp_d.10_63, 128B, _64, { 3, 3, 3, 3 }, 0);
> +  Since these 2 STOREs store 2 vectors that the second vector is half elements
> +  of the first vector. So the length of second STORE will be _64 = _72 / 2;
> +  It's similar to the VIEW_CONVERT of handling masks in SLP.
> +*/
> +
> +static tree
> +vect_set_loop_controls_by_while_len (class loop *loop, loop_vec_info loop_vinfo,
> +				     gimple_seq *preheader_seq,
> +				     gimple_seq *header_seq,
> +				     rgroup_controls *rgc, tree niters)
> +{
> +  tree compare_type = LOOP_VINFO_RGROUP_COMPARE_TYPE (loop_vinfo);
> +  tree iv_type = LOOP_VINFO_RGROUP_IV_TYPE (loop_vinfo);
> +  /* We are not allowing masked approach in WHILE_LEN.  */
> +  gcc_assert (!LOOP_VINFO_FULLY_MASKED_P (loop_vinfo));
> +
> +  tree ctrl_type = rgc->type;
> +  unsigned int nitems_per_iter = rgc->max_nscalars_per_iter * rgc->factor;
> +  poly_uint64 nitems_per_ctrl = TYPE_VECTOR_SUBPARTS (ctrl_type) * rgc->factor;
> +  poly_uint64 vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
> +
> +  /* Calculate the maximum number of item values that the rgroup
> +     handles in total, the number that it handles for each iteration
> +     of the vector loop.  */
> +  tree nitems_total = niters;
> +  if (nitems_per_iter != 1)
> +    {
> +      /* We checked before setting LOOP_VINFO_USING_PARTIAL_VECTORS_P that
> +	 these multiplications don't overflow.  */
> +      tree compare_factor = build_int_cst (compare_type, nitems_per_iter);
> +      nitems_total = gimple_build (preheader_seq, MULT_EXPR, compare_type,
> +				   nitems_total, compare_factor);
> +    }
> +
> +  /* Convert the comparison value to the IV type (either a no-op or
> +     a promotion).  */
> +  nitems_total = gimple_convert (preheader_seq, iv_type, nitems_total);
> +
> +  /* Create an induction variable that counts the number of items
> +     processed.  */
> +  tree index_before_incr, index_after_incr;
> +  gimple_stmt_iterator incr_gsi;
> +  bool insert_after;
> +  standard_iv_increment_position (loop, &incr_gsi, &insert_after);
> +
> +  /* Test the decremented IV, which will never underflow 0 since we have
> +     IFN_WHILE_LEN to gurantee that.  */
> +  tree test_limit = nitems_total;
> +
> +  /* Provide a definition of each control in the group.  */
> +  tree ctrl;
> +  unsigned int i;
> +  FOR_EACH_VEC_ELT_REVERSE (rgc->controls, i, ctrl)
> +    {
> +      /* Previous controls will cover BIAS items.  This control covers the
> +	 next batch.  */
> +      poly_uint64 bias = nitems_per_ctrl * i;
> +      tree bias_tree = build_int_cst (iv_type, bias);
> +
> +      /* Rather than have a new IV that starts at TEST_LIMIT and goes down to
> +	 BIAS, prefer to use the same TEST_LIMIT - BIAS based IV for each
> +	 control and adjust the bound down by BIAS.  */
> +      tree this_test_limit = test_limit;
> +      if (i != 0)
> +	{
> +	  this_test_limit = gimple_build (preheader_seq, MAX_EXPR, iv_type,
> +					  this_test_limit, bias_tree);
> +	  this_test_limit = gimple_build (preheader_seq, MINUS_EXPR, iv_type,
> +					  this_test_limit, bias_tree);
> +	}
> +
> +      /* Create decrement IV.  */
> +      create_iv (this_test_limit, MINUS_EXPR, ctrl, NULL_TREE, loop, &incr_gsi,
> +		 insert_after, &index_before_incr, &index_after_incr);
> +
> +      poly_uint64 final_vf = vf * nitems_per_iter;
> +      tree vf_step = build_int_cst (iv_type, final_vf);
> +      tree res_len;
> +      if (nitems_per_iter != 1)
> +	{
> +	  /* For SLP, we can't allow non-VF number of elements to be processed
> +	     in non-final iteration. We force the number of elements to be
> +	   processed in each non-final iteration is VF elements. If we allow
> +	   non-VF elements processing in non-final iteration will make SLP too
> +	   complicated and produce inferior codegen.
> +
> +	       For example:
> +
> +		If non-final iteration process VF elements.
> +
> +		  ...
> +		  .LEN_STORE (vectp_f.8_51, 128B, _71, { 1, 2, 1, 2 }, 0);
> +		  .LEN_STORE (vectp_f.8_56, 128B, _72, { 1, 2, 1, 2 }, 0);
> +		  ...
> +
> +		If non-final iteration process non-VF elements.
> +
> +		  ...
> +		  .LEN_STORE (vectp_f.8_51, 128B, _71, { 1, 2, 1, 2 }, 0);
> +		  if (_71 % 2 == 0)
> +		   .LEN_STORE (vectp_f.8_56, 128B, _72, { 1, 2, 1, 2 }, 0);
> +		  else
> +		   .LEN_STORE (vectp_f.8_56, 128B, _72, { 2, 1, 2, 1 }, 0);
> +		  ...
> +
> +	   This is the simple case of 2-elements interleaved vector SLP. We
> +	   consider other interleave vector, the situation will become more
> +	   complicated.  */
> +	  res_len = gimple_build (header_seq, MIN_EXPR, iv_type,
> +				  index_before_incr, vf_step);
> +	}
> +      else
> +	{
> +	  res_len = gimple_build (header_seq, IFN_WHILE_LEN, iv_type,
> +				  index_before_incr, vf_step);
> +	}
> +      gassign *assign = gimple_build_assign (ctrl, res_len);
> +      gimple_seq_add_stmt (header_seq, assign);
> +    }
> +
> +  return index_after_incr;
> +}
> +
>  /* Set up the iteration condition and rgroup controls for LOOP, given
>     that LOOP_VINFO_USING_PARTIAL_VECTORS_P is true for the vectorized
>     loop.  LOOP_VINFO describes the vectorization of LOOP.  NITERS is
> @@ -703,6 +958,7 @@ vect_set_loop_condition_partial_vectors (class loop *loop,
>  
>    bool use_masks_p = LOOP_VINFO_FULLY_MASKED_P (loop_vinfo);
>    tree compare_type = LOOP_VINFO_RGROUP_COMPARE_TYPE (loop_vinfo);
> +  tree iv_type = LOOP_VINFO_RGROUP_IV_TYPE (loop_vinfo);
>    unsigned int compare_precision = TYPE_PRECISION (compare_type);
>    tree orig_niters = niters;
>  
> @@ -757,12 +1013,18 @@ vect_set_loop_condition_partial_vectors (class loop *loop,
>  	bool might_wrap_p = vect_rgroup_iv_might_wrap_p (loop_vinfo, rgc);
>  
>  	/* Set up all controls for this group.  */
> -	test_ctrl = vect_set_loop_controls_directly (loop, loop_vinfo,
> -						     &preheader_seq,
> -						     &header_seq,
> -						     loop_cond_gsi, rgc,
> -						     niters, niters_skip,
> -						     might_wrap_p);
> +	if (direct_internal_fn_supported_p (IFN_WHILE_LEN, iv_type,
> +					    OPTIMIZE_FOR_SPEED))
> +	  test_ctrl
> +	    = vect_set_loop_controls_by_while_len (loop, loop_vinfo,
> +						   &preheader_seq, &header_seq,
> +						   rgc, niters);
> +	else
> +	  test_ctrl
> +	    = vect_set_loop_controls_directly (loop, loop_vinfo, &preheader_seq,
> +					       &header_seq, loop_cond_gsi, rgc,
> +					       niters, niters_skip,
> +					       might_wrap_p);
>        }
>  
>    /* Emit all accumulated statements.  */
> @@ -893,7 +1155,7 @@ vect_set_loop_condition_normal (class loop *loop, tree niters, tree step,
>      }
>  
>    standard_iv_increment_position (loop, &incr_gsi, &insert_after);
> -  create_iv (init, step, NULL_TREE, loop,
> +  create_iv (init, PLUS_EXPR, step, NULL_TREE, loop,
>               &incr_gsi, insert_after, &indx_before_incr, &indx_after_incr);
>    indx_after_incr = force_gimple_operand_gsi (&loop_cond_gsi, indx_after_incr,
>  					      true, NULL_TREE, true,
> diff --git a/gcc/tree-vect-loop.cc b/gcc/tree-vect-loop.cc
> index 6ea0f21fd13..d6e90ed93cd 100644
> --- a/gcc/tree-vect-loop.cc
> +++ b/gcc/tree-vect-loop.cc
> @@ -5567,7 +5567,7 @@ vect_create_epilog_for_reduction (loop_vec_info loop_vinfo,
>        gimple_stmt_iterator incr_gsi;
>        bool insert_after;
>        standard_iv_increment_position (loop, &incr_gsi, &insert_after);
> -      create_iv (series_vect, vec_step, NULL_TREE, loop, &incr_gsi,
> +      create_iv (series_vect, PLUS_EXPR, vec_step, NULL_TREE, loop, &incr_gsi,
>  		 insert_after, &indx_before_incr, &indx_after_incr);
>  
>        /* Next create a new phi node vector (NEW_PHI_TREE) which starts
> @@ -10364,12 +10364,14 @@ vect_record_loop_len (loop_vec_info loop_vinfo, vec_loop_lens *lens,
>     rgroup that operates on NVECTORS vectors, where 0 <= INDEX < NVECTORS.  */
>  
>  tree
> -vect_get_loop_len (loop_vec_info loop_vinfo, vec_loop_lens *lens,
> -		   unsigned int nvectors, unsigned int index)
> +vect_get_loop_len (gimple_stmt_iterator *gsi, loop_vec_info loop_vinfo,

please keep loop_vinfo as first parameter and update the comment for the
new.

> +		   vec_loop_lens *lens, unsigned int nvectors, tree vectype,
> +		   unsigned int index)
>  {
>    rgroup_controls *rgl = &(*lens)[nvectors - 1];
>    bool use_bias_adjusted_len =
>      LOOP_VINFO_PARTIAL_LOAD_STORE_BIAS (loop_vinfo) != 0;
> +  tree iv_type = LOOP_VINFO_RGROUP_IV_TYPE (loop_vinfo);
>  
>    /* Populate the rgroup's len array, if this is the first time we've
>       used it.  */
> @@ -10400,6 +10402,27 @@ vect_get_loop_len (loop_vec_info loop_vinfo, vec_loop_lens *lens,
>  
>    if (use_bias_adjusted_len)
>      return rgl->bias_adjusted_ctrl;
> +  else if (direct_internal_fn_supported_p (IFN_WHILE_LEN, iv_type,
> +					   OPTIMIZE_FOR_SPEED))

It's odd to see this test here - why's the .WHILE_LEN sufficiently
different from "other" loop_len cases?

> +    {
> +      tree loop_len = rgl->controls[index];
> +      poly_int64 nunits1 = TYPE_VECTOR_SUBPARTS (rgl->type);
> +      poly_int64 nunits2 = TYPE_VECTOR_SUBPARTS (vectype);
> +      if (maybe_ne (nunits1, nunits2))
> +	{
> +	  /* A loop len for data type X can be reused for data type Y
> +	     if X has N times more elements than Y and if Y's elements
> +	     are N times bigger than X's.  */
> +	  gcc_assert (multiple_p (nunits1, nunits2));
> +	  unsigned int factor = exact_div (nunits1, nunits2).to_constant ();
> +	  gimple_seq seq = NULL;
> +	  loop_len = gimple_build (&seq, RDIV_EXPR, iv_type, loop_len,
> +				   build_int_cst (iv_type, factor));
> +	  if (seq)
> +	    gsi_insert_seq_before (gsi, seq, GSI_SAME_STMT);
> +	}
> +      return loop_len;

You are going to re-emit this division for each load/store, shouldn't
this be integrated in the loop populating controls[] at the first call?

> +    }
>    else
>      return rgl->controls[index];
>  }
> diff --git a/gcc/tree-vect-stmts.cc b/gcc/tree-vect-stmts.cc
> index 6b7dbfd4a23..98bef8f4fc9 100644
> --- a/gcc/tree-vect-stmts.cc
> +++ b/gcc/tree-vect-stmts.cc
> @@ -3144,6 +3144,70 @@ vect_get_data_ptr_increment (vec_info *vinfo,
>    return iv_step;
>  }
>  
> +/* Prepare the pointer IVs which needs to be updated by a variable amount.
> +   Such variable amount is the outcome of .WHILE_LEN. In this case, we can
> +   allow each iteration process the flexible number of elements as long as
> +   the number <= vf elments.
> +
> +   Return data reference according to WHILE_LEN.
> +   If new statements are needed, insert them before GSI.  */
> +
> +static tree
> +get_while_len_data_ref_ptr (vec_info *vinfo, stmt_vec_info stmt_info,
> +			    tree aggr_type, class loop *at_loop, tree offset,
> +			    tree *dummy, gimple_stmt_iterator *gsi,
> +			    bool simd_lane_access_p, vec_loop_lens *loop_lens,
> +			    dr_vec_info *dr_info,
> +			    vect_memory_access_type memory_access_type)
> +{
> +  if (!loop_lens || loop_lens->length () != 1)
> +    return NULL_TREE;
> +  loop_vec_info loop_vinfo = dyn_cast<loop_vec_info> (vinfo);
> +  tree iv_type = LOOP_VINFO_RGROUP_IV_TYPE (loop_vinfo);
> +  tree step = vect_dr_behavior (vinfo, dr_info)->step;
> +  if (!direct_internal_fn_supported_p (IFN_WHILE_LEN, iv_type,
> +				       OPTIMIZE_FOR_SPEED))
> +    return NULL_TREE;
> +
> +  if (memory_access_type == VMAT_INVARIANT)
> +    return NULL_TREE;
> +
> +  /* TODO: We don't support gather/scatter or load_lanes/store_lanes for pointer
> +     IVs are updated by variable amount but we will support them in the future.
> +   */
> +  gcc_assert (memory_access_type != VMAT_GATHER_SCATTER
> +	      && memory_access_type != VMAT_LOAD_STORE_LANES);

All this should be gated at the caller site.  Quoting one:

@@ -9798,6 +9872,15 @@ vectorizable_load (vec_info *vinfo,
                                          at_loop,
                                          offset, &dummy, gsi, &ptr_incr,
                                          simd_lane_access_p, bump);
+         
+         tree while_len_data_ref_ptr
+           = get_while_len_data_ref_ptr (vinfo, stmt_info, aggr_type, 
at_loop,
+                                         offset, &dummy, gsi,
+                                         simd_lane_access_p, loop_lens,
+                                         dr_info, memory_access_type);
+         if (while_len_data_ref_ptr)
+           dataref_ptr = while_len_data_ref_ptr;

you call this after other methods have computed a dataref_ptr, that's
very ugly.  Instead you should compute the "update method" during
load/store _analysis_ and then conciously decide on using this
function (which then better should succeed).  That will also be required
to properly apply costing during analysis.

I also wonder how you can get away without touching the j == 1 case.
I suppose we'll ever use the .WHILE_LEN path then.

> +  /* When we support WHILE_LEN pattern, we dynamic adjust
> +     the memory address by .WHILE_LEN result.
> +
> +     The result of .WHILE_LEN is the number of elements to
> +     be processed of each iteration. So the memory address
> +     adjustment operation should be:
> +
> +     bytesize = GET_MODE_SIZE (element_mode (aggr_type));
> +     addr = addr + .WHILE_LEN (ARG..) * bytesize;
> +  */
> +  gimple *ptr_incr;
> +  tree loop_len
> +    = vect_get_loop_len (gsi, loop_vinfo, loop_lens, 1, aggr_type, 0);
> +  tree len_type = TREE_TYPE (loop_len);
> +  poly_uint64 bytesize = GET_MODE_SIZE (element_mode (aggr_type));
> +  /* Since the outcome of .WHILE_LEN is element size, we should adjust
> +     it into bytesize so that it can be used in address pointer variable
> +     amount IVs adjustment.  */
> +  tree tmp = fold_build2 (MULT_EXPR, len_type, loop_len,
> +			  build_int_cst (len_type, bytesize));
> +  if (tree_int_cst_sgn (step) == -1)
> +    tmp = fold_build1 (NEGATE_EXPR, len_type, tmp);
> +  tree bump = make_temp_ssa_name (len_type, NULL, "ivtmp");
> +  gassign *assign = gimple_build_assign (bump, tmp);
> +  gsi_insert_before (gsi, assign, GSI_SAME_STMT);
> +  return vect_create_data_ref_ptr (vinfo, stmt_info, aggr_type, at_loop, offset,
> +				   dummy, gsi, &ptr_incr, simd_lane_access_p,
> +				   bump);
> +}
> +
>  /* Check and perform vectorization of BUILT_IN_BSWAP{16,32,64,128}.  */
>  
>  static bool
> @@ -8195,7 +8259,7 @@ vectorizable_store (vec_info *vinfo,
>  
>        stride_base = cse_and_gimplify_to_preheader (loop_vinfo, stride_base);
>        ivstep = cse_and_gimplify_to_preheader (loop_vinfo, ivstep);
> -      create_iv (stride_base, ivstep, NULL,
> +      create_iv (stride_base, PLUS_EXPR, ivstep, NULL,
>  		 loop, &incr_gsi, insert_after,
>  		 &offvar, NULL);
>        incr = gsi_stmt (incr_gsi);
> @@ -8465,6 +8529,15 @@ vectorizable_store (vec_info *vinfo,
>  					  simd_lane_access_p ? loop : NULL,
>  					  offset, &dummy, gsi, &ptr_incr,
>  					  simd_lane_access_p, bump);
> +
> +	  tree while_len_data_ref_ptr
> +	    = get_while_len_data_ref_ptr (vinfo, stmt_info, aggr_type,
> +					  simd_lane_access_p ? loop : NULL,
> +					  offset, &dummy, gsi,
> +					  simd_lane_access_p, loop_lens,
> +					  dr_info, memory_access_type);
> +	  if (while_len_data_ref_ptr)
> +	    dataref_ptr = while_len_data_ref_ptr;
>  	}
>        else
>  	{
> @@ -8652,8 +8725,9 @@ vectorizable_store (vec_info *vinfo,
>  	      else if (loop_lens)
>  		{
>  		  tree final_len
> -		    = vect_get_loop_len (loop_vinfo, loop_lens,
> -					 vec_num * ncopies, vec_num * j + i);
> +		    = vect_get_loop_len (gsi, loop_vinfo, loop_lens,
> +					 vec_num * ncopies, vectype,
> +					 vec_num * j + i);
>  		  tree ptr = build_int_cst (ref_type, align * BITS_PER_UNIT);
>  		  machine_mode vmode = TYPE_MODE (vectype);
>  		  opt_machine_mode new_ovmode
> @@ -9314,7 +9388,7 @@ vectorizable_load (vec_info *vinfo,
>  
>        stride_base = cse_and_gimplify_to_preheader (loop_vinfo, stride_base);
>        ivstep = cse_and_gimplify_to_preheader (loop_vinfo, ivstep);
> -      create_iv (stride_base, ivstep, NULL,
> +      create_iv (stride_base, PLUS_EXPR, ivstep, NULL,
>  		 loop, &incr_gsi, insert_after,
>  		 &offvar, NULL);
>  
> @@ -9798,6 +9872,15 @@ vectorizable_load (vec_info *vinfo,
>  					  at_loop,
>  					  offset, &dummy, gsi, &ptr_incr,
>  					  simd_lane_access_p, bump);
> +
> +	  tree while_len_data_ref_ptr
> +	    = get_while_len_data_ref_ptr (vinfo, stmt_info, aggr_type, at_loop,
> +					  offset, &dummy, gsi,
> +					  simd_lane_access_p, loop_lens,
> +					  dr_info, memory_access_type);
> +	  if (while_len_data_ref_ptr)
> +	    dataref_ptr = while_len_data_ref_ptr;
> +
>  	  if (mask)
>  	    vec_mask = vec_masks[0];
>  	}
> @@ -10008,8 +10091,8 @@ vectorizable_load (vec_info *vinfo,
>  		    else if (loop_lens && memory_access_type != VMAT_INVARIANT)
>  		      {
>  			tree final_len
> -			  = vect_get_loop_len (loop_vinfo, loop_lens,
> -					       vec_num * ncopies,
> +			  = vect_get_loop_len (gsi, loop_vinfo, loop_lens,
> +					       vec_num * ncopies, vectype,
>  					       vec_num * j + i);
>  			tree ptr = build_int_cst (ref_type,
>  						  align * BITS_PER_UNIT);
> diff --git a/gcc/tree-vectorizer.h b/gcc/tree-vectorizer.h
> index 9cf2fb23fe3..811c17e1d3d 100644
> --- a/gcc/tree-vectorizer.h
> +++ b/gcc/tree-vectorizer.h
> @@ -2293,8 +2293,8 @@ extern tree vect_get_loop_mask (gimple_stmt_iterator *, vec_loop_masks *,
>  				unsigned int, tree, unsigned int);
>  extern void vect_record_loop_len (loop_vec_info, vec_loop_lens *, unsigned int,
>  				  tree, unsigned int);
> -extern tree vect_get_loop_len (loop_vec_info, vec_loop_lens *, unsigned int,
> -			       unsigned int);
> +extern tree vect_get_loop_len (gimple_stmt_iterator *, loop_vec_info, vec_loop_lens *,
> +			       unsigned int, tree, unsigned int);
>  extern gimple_seq vect_gen_len (tree, tree, tree, tree);
>  extern stmt_vec_info info_for_reduction (vec_info *, stmt_vec_info);
>  extern bool reduction_fn_for_scalar_code (code_helper, internal_fn *);

Otherwise looks reasonable.

As said, it _looks_ like you can progress without .WHILE_LEN and using
.WHILE_LEN is a pure optimization?

Thanks,
Richard.