On Tue, May 21, 2024 at 5:28 AM Li, Pan2 <pan2...@intel.com> wrote:
>
> Thanks Andrew for comments.
>
>
>
> > I think you need to make sure type and @0's type matches.
>
>
>
> Oh, yes, we need that, will update in v2.
>
>
>
> > Also I don't think you need :c here since you don't match @0 nor @1 more
> > than once.
>
>
>
> You mean the :c from (IFN_ADD_OVERFLOW:c@2 @0 @1)), right?
>
> My initial idea is to catch both the (IFN_ADD_OVERFLOW @0 @1) and
> (IFN_ADD_OVERFLOW @1 @0).
>
> It is unnecessary if IFN_ADD_OVERFLOW takes care of this already.
Since in this case there is Canonical form/order here (at least there
should be).
> + (cond (ne (imagpart (IFN_ADD_OVERFLOW:c@2 @0 @1)) integer_zerop)
> + integer_minus_onep (realpart @2))
Since you matching @2 for the realpart rather than `(IFN_ADD_OVERFLOW
@0 @1)` directly the :c is not needed and genmatch will just generate
extra matching code that cannot be not get reached
Thanks,
Andrew.
>
>
>
> Pan
>
>
>
>
>
> From: Andrew Pinski <pins...@gmail.com>
> Sent: Tuesday, May 21, 2024 7:40 PM
> To: Li, Pan2 <pan2...@intel.com>
> Cc: GCC Patches <gcc-patches@gcc.gnu.org>; 钟居哲 <juzhe.zh...@rivai.ai>; Kito
> Cheng <kito.ch...@gmail.com>; Tamar Christina <tamar.christ...@arm.com>;
> Richard Guenther <richard.guent...@gmail.com>
> Subject: Re: [PATCH v1 1/2] Match: Support __builtin_add_overflow branch form
> for unsigned SAT_ADD
>
>
>
>
>
> On Tue, May 21, 2024, 3:55 AM <pan2...@intel.com> wrote:
>
> From: Pan Li <pan2...@intel.com>
>
> This patch would like to support the __builtin_add_overflow branch form for
> unsigned SAT_ADD. For example as below:
>
> uint64_t
> sat_add (uint64_t x, uint64_t y)
> {
> uint64_t ret;
> return __builtin_add_overflow (x, y, &ret) ? -1 : ret;
> }
>
> Different to the branchless version, we leverage the simplify to
> convert the branch version of SAT_ADD into branchless if and only
> if the backend has supported the IFN_SAT_ADD. Thus, the backend has
> the ability to choose branch or branchless implementation of .SAT_ADD.
> For example, some target can take care of branches code more optimally.
>
> When the target implement the IFN_SAT_ADD for unsigned and before this
> patch:
>
> uint64_t sat_add (uint64_t x, uint64_t y)
> {
> long unsigned int _1;
> long unsigned int _2;
> uint64_t _3;
> __complex__ long unsigned int _6;
>
> ;; basic block 2, loop depth 0
> ;; pred: ENTRY
> _6 = .ADD_OVERFLOW (x_4(D), y_5(D));
> _2 = IMAGPART_EXPR <_6>;
> if (_2 != 0)
> goto <bb 4>; [35.00%]
> else
> goto <bb 3>; [65.00%]
> ;; succ: 4
> ;; 3
>
> ;; basic block 3, loop depth 0
> ;; pred: 2
> _1 = REALPART_EXPR <_6>;
> ;; succ: 4
>
> ;; basic block 4, loop depth 0
> ;; pred: 3
> ;; 2
> # _3 = PHI <_1(3), 18446744073709551615(2)>
> return _3;
> ;; succ: EXIT
> }
>
> After this patch:
> uint64_t sat_add (uint64_t x, uint64_t y)
> {
> long unsigned int _12;
>
> ;; basic block 2, loop depth 0
> ;; pred: ENTRY
> _12 = .SAT_ADD (x_4(D), y_5(D)); [tail call]
> return _12;
> ;; succ: EXIT
> }
>
> The below test suites are passed for this patch:
> * The x86 bootstrap test.
> * The x86 fully regression test.
> * The riscv fully regression test.
>
> gcc/ChangeLog:
>
> * match.pd: Add new simplify to convert branch SAT_ADD into
> branchless, if and only if backend implement the IFN.
>
> Signed-off-by: Pan Li <pan2...@intel.com>
> ---
> gcc/match.pd | 10 ++++++++++
> 1 file changed, 10 insertions(+)
>
> diff --git a/gcc/match.pd b/gcc/match.pd
> index 0f9c34fa897..8b9ded98323 100644
> --- a/gcc/match.pd
> +++ b/gcc/match.pd
> @@ -3094,6 +3094,16 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
> (match (unsigned_integer_sat_add @0 @1)
> (bit_ior:c (usadd_left_part_2 @0 @1) (usadd_right_part_2 @0 @1)))
>
> +#if GIMPLE
> +
> +(simplify
> + (cond (ne (imagpart (IFN_ADD_OVERFLOW:c@2 @0 @1)) integer_zerop)
> + integer_minus_onep (realpart @2))
> + (if (direct_internal_fn_supported_p (IFN_SAT_ADD, type, OPTIMIZE_FOR_BOTH))
> + (bit_ior (plus@3 @0 @1) (negate (convert (lt @3 @0))))))
>
>
>
> I think you need to make sure type and @0's type matches.
>
>
>
> Also I don't think you need :c here since you don't match @0 nor @1 more than
> once.
>
>
>
> Thanks,
>
> Andrew
>
>
>
>
>
> +
> +#endif
> +
> /* x > y && x != XXX_MIN --> x > y
> x > y && x == XXX_MIN --> false . */
> (for eqne (eq ne)
> --
> 2.34.1
