[14/n] PR85694: Rework overwidening detection

This patch is the main part of PR85694.  The aim is to recognise at least:

  signed char *a, *b, *c;
  ...
  for (int i = 0; i < 2048; i++)
    c[i] = (a[i] + b[i]) >> 1;

as an over-widening pattern, since the addition and shift can be done
on shorts rather than ints.  However, it ended up being a lot more
general than that.

The current over-widening pattern detection is limited to a few simple
cases: logical ops with immediate second operands, and shifts by a
constant.  These cases are enough for common pixel-format conversion
and can be detected in a peephole way.

The loop above requires two generalisations of the current code: support
for addition as well as logical ops, and support for non-constant second
operands.  These are harder to detect in the same peephole way, so the
patch tries to take a more global approach.

The idea is to get information about the minimum operation width
in two ways:

(1) by using the range information attached to the SSA_NAMEs
    (effectively a forward walk, since the range info is
    context-independent).

(2) by back-propagating the number of output bits required by
    users of the result.

As explained in the comments, there's a balance to be struck between
narrowing an individual operation and fitting in with the surrounding
code.  The approach is pretty conservative: if we could narrow an
operation to N bits without changing its semantics, it's OK to do that if:

- no operations later in the chain require more than N bits; or

- all internally-defined inputs are extended from N bits or fewer,
  and at least one of them is single-use.

See the comments for the rationale.

I didn't bother adding STMT_VINFO_* wrappers for the new fields
since the code seemed more readable without.

2018-06-20  Richard Sandiford  <richard.sandiford@arm.com>

gcc/
	* poly-int.h (print_hex): New function.
	* dumpfile.h (dump_dec, dump_hex): Declare.
	* dumpfile.c (dump_dec, dump_hex): New poly_wide_int functions.
	* tree-vectorizer.h (_stmt_vec_info): Add min_output_precision,
	min_input_precision, operation_precision and operation_sign.
	* tree-vect-patterns.c (vect_get_range_info): New function.
	(vect_same_loop_or_bb_p, vect_single_imm_use)
	(vect_operation_fits_smaller_type): Delete.
	(vect_look_through_possible_promotion): Add an optional
	single_use_p parameter.
	(vect_recog_over_widening_pattern): Rewrite to use new
	stmt_vec_info infomration.  Handle one operation at a time.
	(vect_recog_cast_forwprop_pattern, vect_narrowable_type_p)
	(vect_truncatable_operation_p, vect_set_operation_type)
	(vect_set_min_input_precision): New functions.
	(vect_determine_min_output_precision_1): Likewise.
	(vect_determine_min_output_precision): Likewise.
	(vect_determine_precisions_from_range): Likewise.
	(vect_determine_precisions_from_users): Likewise.
	(vect_determine_stmt_precisions, vect_determine_precisions): Likewise.
	(vect_vect_recog_func_ptrs): Put over_widening first.
	Add cast_forwprop.
	(vect_pattern_recog): Call vect_determine_precisions.

gcc/testsuite/
	* gcc.dg/vect/vect-widen-mult-u8-u32.c: Check specifically for a
	widen_mult pattern.
	* gcc.dg/vect/vect-over-widen-1.c: Update the scan tests for new
	over-widening messages.
	* gcc.dg/vect/vect-over-widen-1-big-array.c: Likewise.
	* gcc.dg/vect/vect-over-widen-2.c: Likewise.
	* gcc.dg/vect/vect-over-widen-2-big-array.c: Likewise.
	* gcc.dg/vect/vect-over-widen-3.c: Likewise.
	* gcc.dg/vect/vect-over-widen-3-big-array.c: Likewise.
	* gcc.dg/vect/vect-over-widen-4.c: Likewise.
	* gcc.dg/vect/vect-over-widen-4-big-array.c: Likewise.
	* gcc.dg/vect/bb-slp-over-widen-1.c: New test.
	* gcc.dg/vect/bb-slp-over-widen-2.c: Likewise.
	* gcc.dg/vect/vect-over-widen-5.c: Likewise.
	* gcc.dg/vect/vect-over-widen-6.c: Likewise.
	* gcc.dg/vect/vect-over-widen-7.c: Likewise.
	* gcc.dg/vect/vect-over-widen-8.c: Likewise.
	* gcc.dg/vect/vect-over-widen-9.c: Likewise.
	* gcc.dg/vect/vect-over-widen-10.c: Likewise.
	* gcc.dg/vect/vect-over-widen-11.c: Likewise.
	* gcc.dg/vect/vect-over-widen-12.c: Likewise.
	* gcc.dg/vect/vect-over-widen-13.c: Likewise.
	* gcc.dg/vect/vect-over-widen-14.c: Likewise.
	* gcc.dg/vect/vect-over-widen-15.c: Likewise.
	* gcc.dg/vect/vect-over-widen-16.c: Likewise.
	* gcc.dg/vect/vect-over-widen-17.c: Likewise.
	* gcc.dg/vect/vect-over-widen-18.c: Likewise.
	* gcc.dg/vect/vect-over-widen-19.c: Likewise.
	* gcc.dg/vect/vect-over-widen-20.c: Likewise.
	* gcc.dg/vect/vect-over-widen-21.c: Likewise.

From-SVN: r262333

1 files changed, 15 insertions, 0 deletions

diff --git a/gcc/tree-vectorizer.h b/gcc/tree-vectorizer.h
index 2dac54e..28be41f 100644
--- a/gcc/tree-vectorizer.h
+++ b/gcc/tree-vectorizer.h
@@ -899,6 +899,21 @@ typedef struct _stmt_vec_info {
 
   /* The number of scalar stmt references from active SLP instances.  */
   unsigned int num_slp_uses;
+
+  /* If nonzero, the lhs of the statement could be truncated to this
+     many bits without affecting any users of the result.  */
+  unsigned int min_output_precision;
+
+  /* If nonzero, all non-boolean input operands have the same precision,
+     and they could each be truncated to this many bits without changing
+     the result.  */
+  unsigned int min_input_precision;
+
+  /* If OPERATION_BITS is nonzero, the statement could be performed on
+     an integer with the sign and number of bits given by OPERATION_SIGN
+     and OPERATION_BITS without changing the result.  */
+  unsigned int operation_precision;
+  signop operation_sign;
 } *stmt_vec_info;
 
 /* Information about a gather/scatter call.  */