[patch] Enhance conditional store sinking

[Richard Guenther]

On Tue, Mar 22, 2011 at 2:28 PM, Ira Rosen <ira.rosen@linaro.org> wrote:
> On 17 March 2011 16:48, Richard Guenther <richard.guenther@gmail.com> wrote:
>
>>> +  then_datarefs = VEC_alloc (data_reference_p, heap, 1);
>>> +  else_datarefs = VEC_alloc (data_reference_p, heap, 1);
>>> +  then_ddrs = VEC_alloc (ddr_p, heap, 1);
>>> +  else_ddrs = VEC_alloc (ddr_p, heap, 1);
>>> +  if (!compute_data_dependences_for_bb (then_bb, false, &then_datarefs,
>>> +                                        &then_ddrs)
>>
>> Can we avoid computing dependencies if the other BB would have no
>> data-refs?  Thus, split collecting datarefs and computing dependences?
>
> Done.
>
>>
>>> +      || !compute_data_dependences_for_bb (else_bb, false, &else_datarefs,
>>> +                                           &else_ddrs)
>>> +      || !VEC_length (data_reference_p, then_datarefs)
>>> +      || !VEC_length (data_reference_p, else_datarefs))
>>> +    {
>>> +      free_data_refs (then_datarefs);
>>> +      free_data_refs (else_datarefs);
>>> +      free_dependence_relations (then_ddrs);
>>> +      free_dependence_relations (else_ddrs);
>>> +      return false;
>>> +    }
>>> +
>>> +  /* Check that there are no read-after-write or write-after-write dependencies
>>> +     in THEN_BB.  */
>>> +  FOR_EACH_VEC_ELT (ddr_p, then_ddrs, i, ddr)
>>> +    {
>>> +      struct data_reference *dra = DDR_A (ddr);
>>> +      struct data_reference *drb = DDR_B (ddr);
>>> +
>>> +      if (DDR_ARE_DEPENDENT (ddr) != chrec_known
>>> +          && ((DR_IS_READ (dra) && DR_IS_WRITE (drb)
>>> +               && gimple_uid (DR_STMT (dra)) > gimple_uid (DR_STMT (drb)))
>>> +              || (DR_IS_READ (drb) && DR_IS_WRITE (dra)
>>> +                  && gimple_uid (DR_STMT (drb)) > gimple_uid (DR_STMT (dra)))
>>> +              || (DR_IS_WRITE (dra) && DR_IS_WRITE (drb))))
>>
>> The gimple_uids are not initialized here, you need to make sure to
>> call renumber_gimple_stmt_uids () before starting.  Note that phiopt
>> incrementally changes the IL, so I'm not sure those uids will stay
>> valid as stmts are moved across blocks.
>
> I added a call to renumber_gimple_stmt_uids_in_blocks() before data
> dependence checks, and there are no code changes between that and the
> checks, so, I think, it should be OK.
>
>>
>>> +        {
>>> +          free_data_refs (then_datarefs);
>>> +          free_data_refs (else_datarefs);
>>> +          free_dependence_relations (then_ddrs);
>>> +          free_dependence_relations (else_ddrs);
>>> +          return false;
>>> +        }
>>> +    }
>>> +
>>> +  /* Check that there are no read-after-write or write-after-write dependencies
>>> +     in ELSE_BB.  */
>>> +  FOR_EACH_VEC_ELT (ddr_p, else_ddrs, i, ddr)
>>> +    {
>>> +      struct data_reference *dra = DDR_A (ddr);
>>> +      struct data_reference *drb = DDR_B (ddr);
>>> +
>>> +      if (DDR_ARE_DEPENDENT (ddr) != chrec_known
>>> +          && ((DR_IS_READ (dra) && DR_IS_WRITE (drb)
>>> +               && gimple_uid (DR_STMT (dra)) > gimple_uid (DR_STMT (drb)))
>>> +              || (DR_IS_READ (drb) && DR_IS_WRITE (dra)
>>> +                  && gimple_uid (DR_STMT (drb)) > gimple_uid (DR_STMT (dra)))
>>> +              || (DR_IS_WRITE (dra) && DR_IS_WRITE (drb))))
>>> +        {
>>> +          free_data_refs (then_datarefs);
>>> +          free_data_refs (else_datarefs);
>>> +          free_dependence_relations (then_ddrs);
>>> +          free_dependence_relations (else_ddrs);
>>> +          return false;
>>> +        }
>>> +    }
>>> +
>>> +  /* Found pairs of stores with equal LHS.  */
>>> +  FOR_EACH_VEC_ELT (data_reference_p, then_datarefs, i, then_dr)
>>> +    {
>>> +      if (DR_IS_READ (then_dr))
>>> +        continue;
>>> +
>>> +      then_store = DR_STMT (then_dr);
>>> +      then_lhs = gimple_assign_lhs (then_store);
>>> +      found = false;
>>> +
>>> +      FOR_EACH_VEC_ELT (data_reference_p, else_datarefs, j, else_dr)
>>> +        {
>>> +          if (DR_IS_READ (else_dr))
>>> +            continue;
>>> +
>>> +          else_store = DR_STMT (else_dr);
>>> +          else_lhs = gimple_assign_lhs (else_store);
>>> +
>>> +          if (operand_equal_p (then_lhs, else_lhs, 0))
>>> +            {
>>> +              found = true;
>>> +              break;
>>> +            }
>>> +        }
>>> +
>>> +      if (!found)
>>> +        continue;
>>> +
>>> +      res = cond_if_else_store_replacement_1 (then_bb, else_bb, join_bb,
>>> +                                              then_store, else_store);
>>
>> So you are executing if-else store replacement for common data reference
>> pairs only.  I think it's cheaper to collect those pairs before computing
>> dependences and only if there is at least one pair perform the optimization.
>
> OK, I changed the order.
>
>>
>> You basically perform store sinking, creating a PHI node for each store
>> you sink (that's then probably if-converted by if-conversion later, eventually
>> redundant with -ftree-loop-if-convert-stores?)
>>
>> I am concerned that having no bound on the number of stores sunk will
>> increase register pressure and does not allow scheduling of the stores
>> in an optimal way.  Consider two BBs similar to
>>
>>  t = a + b;
>>  *p = t;
>>  t = c + d;
>>  *q = t;
>>
>> where the transformation undoes a good schedule and makes fixing it
>> impossible if the remaining statements are not if-convertible.
>>
>> Thus, I'd rather make this transformation only if in the end the conditional
>> can be completely if-converted.
>>
>> I realize that we already do unbound and very aggressive if-conversion
>> in tree-ifcvt.c regardless of whether the loop will be vectorized or not
>> (including leaking the if-converted loops to the various loop versions
>> we create during vectorization, causing only code-size bloat).  But it's
>> not a good reason to continue down this road ;)
>>
>> I suppose a simple maximum on the number of stores to sink
>> controllable by a param should do, eventually disabling this
>> extended transformation when vectorization is disabled?
>
> I added a param, and it's set to 0 if either vectorization or if-conversion
> is disabled.
>
>>
>> Otherwise the implementation looks good.
>
> Bootstrapped and tested on powerpc64-suse-linux.
> OK to apply?

Ok with

+  if (find_data_references_in_bb (NULL, then_bb, &then_datarefs)
+       == chrec_dont_know
+      || !VEC_length (data_reference_p, then_datarefs)
+      || find_data_references_in_bb (NULL, else_bb, &else_datarefs)
+       == chrec_dont_know

paranteses added around the call & comparison to make auto-indent
work.

Thanks,
Richard.

> Thanks,
> Ira
>
> ChangeLog:
>
>     * doc/invoke.texi (max-stores-to-sink): Document.
>     * params.h (MAX_STORES_TO_SINK): Define.
>     * opts.c (finish_options): Set MAX_STORES_TO_SINK to 0
>     if either vectorization or if-conversion is disabled.
>     * tree-data-ref.c (dr_equal_offsets_p1): Moved and renamed from
>     tree-vect-data-refs.c vect_equal_offsets.
>     (dr_equal_offsets_p): New function.
>     (find_data_references_in_bb): Remove static.
>     * tree-data-ref.h (find_data_references_in_bb): Declare.
>     (dr_equal_offsets_p): Likewise.
>     * tree-vect-data-refs.c (vect_equal_offsets): Move to tree-data-ref.c.
>     (vect_drs_dependent_in_basic_block): Update calls to vect_equal_offsets.
>     (vect_check_interleaving): Likewise.
>     * tree-ssa-phiopt.c: Include cfgloop.h and tree-data-ref.h.
>     (cond_if_else_store_replacement): Rename to...
>     (cond_if_else_store_replacement_1): ... this. Change arguments and
>     documentation.
>     (cond_if_else_store_replacement): New function.
>     * Makefile.in (tree-ssa-phiopt.o): Adjust dependencies.
>     * params.def (PARAM_MAX_STORES_TO_SINK): Define.
>
> testsuite/ChangeLog:
>
>     * gcc.dg/vect/vect-cselim-1.c: New test.
>     * gcc.dg/vect/vect-cselim-2.c: New test.
>