[tree-optimization/63387] Recognize isunordered

[tree-optimization/63387] Recognize isunordered

[Marc Glisse]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 430 bytes --]

Hello,

just a simple pattern for match.pd. I am ignoring the issue of whether 
isnan is the same as isunordered, I am only combining isunordered 
together.

2015-04-13  Marc Glisse  <marc.glisse@inria.fr>

 	PR tree-optimization/63387
gcc/
 	* match.pd ((x unord x) | (y unord y) -> (x unord y),
 	(x unord x) | (x unord y) -> (x unord y)): New simplifications.
gcc/testsuite/
 	* gcc.dg/pr63387.c: New testcase.

-- 
Marc Glisse

[-- Attachment #2: Type: TEXT/PLAIN, Size: 2146 bytes --]

Index: match.pd
===================================================================
--- match.pd	(revision 222041)
+++ match.pd	(working copy)
@@ -925,20 +925,27 @@ along with GCC; see the file COPYING3.
      (ncmp @0 @1)))))
  (simplify
   (bit_xor (cmp @0 @1) integer_truep)
   (with { enum tree_code ic = invert_tree_comparison
             (cmp, HONOR_NANS (@0)); }
    (if (ic == icmp)
     (icmp @0 @1))
    (if (ic == ncmp)
     (ncmp @0 @1)))))
 
+/* Unordered tests if either argument is a NaN.  */
+(simplify
+ (bit_ior (unordered @0 @0) (unordered @1 @1))
+ (unordered @0 @1))
+(simplify
+ (bit_ior:c (unordered @0 @0) (unordered:c@2 @0 @1))
+ @2)
 
 /* Simplification of math builtins.  */
 
 (define_operator_list LOG BUILT_IN_LOGF BUILT_IN_LOG BUILT_IN_LOGL)
 (define_operator_list EXP BUILT_IN_EXPF BUILT_IN_EXP BUILT_IN_EXPL)
 (define_operator_list LOG2 BUILT_IN_LOG2F BUILT_IN_LOG2 BUILT_IN_LOG2L)
 (define_operator_list EXP2 BUILT_IN_EXP2F BUILT_IN_EXP2 BUILT_IN_EXP2L)
 (define_operator_list LOG10 BUILT_IN_LOG10F BUILT_IN_LOG10 BUILT_IN_LOG10L)
 (define_operator_list EXP10 BUILT_IN_EXP10F BUILT_IN_EXP10 BUILT_IN_EXP10L)
 (define_operator_list POW BUILT_IN_POWF BUILT_IN_POW BUILT_IN_POWL)
Index: testsuite/gcc.dg/pr63387.c
===================================================================
--- testsuite/gcc.dg/pr63387.c	(revision 0)
+++ testsuite/gcc.dg/pr63387.c	(working copy)
@@ -0,0 +1,19 @@
+/* { dg-do compile } */
+/* { dg-options "-O -fdump-tree-optimized" } */
+
+int f(double aaa, double bbb){
+  int xa = __builtin_isunordered(aaa, aaa);
+  int xb = __builtin_isunordered(bbb, bbb);
+  return xa | xb;
+}
+
+int g(double aaa, double bbb){
+  int xa = __builtin_isunordered(aaa, bbb);
+  int xb = __builtin_isunordered(bbb, bbb);
+  return xa | xb;
+}
+
+/* { dg-final { scan-tree-dump-not "aaa\[^\n\r\]* unord aaa" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "bbb\[^\n\r\]* unord bbb" "optimized" } } */
+/* { dg-final { scan-tree-dump-times "aaa\[^\n\r\]* unord bbb" 2 "optimized" } } */
+/* { dg-final { cleanup-tree-dump "optimized" } } */

Re: [tree-optimization/63387] Recognize isunordered

[Richard Biener]

On Mon, Apr 13, 2015 at 2:23 PM, Marc Glisse <marc.glisse@inria.fr> wrote:
> Hello,
>
> just a simple pattern for match.pd. I am ignoring the issue of whether isnan
> is the same as isunordered, I am only combining isunordered together.

Ok.

Thanks,
Richard.

> 2015-04-13  Marc Glisse  <marc.glisse@inria.fr>
>
>         PR tree-optimization/63387
> gcc/
>         * match.pd ((x unord x) | (y unord y) -> (x unord y),
>         (x unord x) | (x unord y) -> (x unord y)): New simplifications.
> gcc/testsuite/
>         * gcc.dg/pr63387.c: New testcase.
>
> --
> Marc Glisse
> Index: match.pd
> ===================================================================
> --- match.pd    (revision 222041)
> +++ match.pd    (working copy)
> @@ -925,20 +925,27 @@ along with GCC; see the file COPYING3.
>       (ncmp @0 @1)))))
>   (simplify
>    (bit_xor (cmp @0 @1) integer_truep)
>    (with { enum tree_code ic = invert_tree_comparison
>              (cmp, HONOR_NANS (@0)); }
>     (if (ic == icmp)
>      (icmp @0 @1))
>     (if (ic == ncmp)
>      (ncmp @0 @1)))))
>
> +/* Unordered tests if either argument is a NaN.  */
> +(simplify
> + (bit_ior (unordered @0 @0) (unordered @1 @1))
> + (unordered @0 @1))
> +(simplify
> + (bit_ior:c (unordered @0 @0) (unordered:c@2 @0 @1))
> + @2)
>
>  /* Simplification of math builtins.  */
>
>  (define_operator_list LOG BUILT_IN_LOGF BUILT_IN_LOG BUILT_IN_LOGL)
>  (define_operator_list EXP BUILT_IN_EXPF BUILT_IN_EXP BUILT_IN_EXPL)
>  (define_operator_list LOG2 BUILT_IN_LOG2F BUILT_IN_LOG2 BUILT_IN_LOG2L)
>  (define_operator_list EXP2 BUILT_IN_EXP2F BUILT_IN_EXP2 BUILT_IN_EXP2L)
>  (define_operator_list LOG10 BUILT_IN_LOG10F BUILT_IN_LOG10 BUILT_IN_LOG10L)
>  (define_operator_list EXP10 BUILT_IN_EXP10F BUILT_IN_EXP10 BUILT_IN_EXP10L)
>  (define_operator_list POW BUILT_IN_POWF BUILT_IN_POW BUILT_IN_POWL)
> Index: testsuite/gcc.dg/pr63387.c
> ===================================================================
> --- testsuite/gcc.dg/pr63387.c  (revision 0)
> +++ testsuite/gcc.dg/pr63387.c  (working copy)
> @@ -0,0 +1,19 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O -fdump-tree-optimized" } */
> +
> +int f(double aaa, double bbb){
> +  int xa = __builtin_isunordered(aaa, aaa);
> +  int xb = __builtin_isunordered(bbb, bbb);
> +  return xa | xb;
> +}
> +
> +int g(double aaa, double bbb){
> +  int xa = __builtin_isunordered(aaa, bbb);
> +  int xb = __builtin_isunordered(bbb, bbb);
> +  return xa | xb;
> +}
> +
> +/* { dg-final { scan-tree-dump-not "aaa\[^\n\r\]* unord aaa" "optimized" }
> } */
> +/* { dg-final { scan-tree-dump-not "bbb\[^\n\r\]* unord bbb" "optimized" }
> } */
> +/* { dg-final { scan-tree-dump-times "aaa\[^\n\r\]* unord bbb" 2
> "optimized" } } */
> +/* { dg-final { cleanup-tree-dump "optimized" } } */
>

Re: [tree-optimization/63387] Recognize isunordered

[Marc Glisse]

On Mon, 13 Apr 2015, Richard Biener wrote:

> On Mon, Apr 13, 2015 at 2:23 PM, Marc Glisse <marc.glisse@inria.fr> wrote:
>> Hello,
>>
>> just a simple pattern for match.pd. I am ignoring the issue of whether isnan
>> is the same as isunordered, I am only combining isunordered together.
>
> Ok.

Oups, I am an idiot. My own comment in bugzilla shows why the patch has a 
bug, I'll post a fixed version soon. Sorry,

-- 
Marc Glisse

Re: [tree-optimization/63387] Recognize isunordered

[Marc Glisse]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 706 bytes --]

On Mon, 13 Apr 2015, Marc Glisse wrote:

> On Mon, 13 Apr 2015, Richard Biener wrote:
>
>> On Mon, Apr 13, 2015 at 2:23 PM, Marc Glisse <marc.glisse@inria.fr> wrote:
>>> Hello,
>>> 
>>> just a simple pattern for match.pd. I am ignoring the issue of whether 
>>> isnan
>>> is the same as isunordered, I am only combining isunordered together.
>> 
>> Ok.
>
> Oups, I am an idiot. My own comment in bugzilla shows why the patch has a 
> bug, I'll post a fixed version soon. Sorry,

Here is the fixed patch (same ChangeLog).

I copied the type compatibility check from other places in match.pd, it 
would be a good idea to provide a helper to avoid repeating this code.

Is the new version ok?

-- 
Marc Glisse

[-- Attachment #2: Type: TEXT/PLAIN, Size: 2535 bytes --]

Index: gcc/match.pd
===================================================================
--- gcc/match.pd	(revision 222062)
+++ gcc/match.pd	(working copy)
@@ -925,20 +925,29 @@ along with GCC; see the file COPYING3.
      (ncmp @0 @1)))))
  (simplify
   (bit_xor (cmp @0 @1) integer_truep)
   (with { enum tree_code ic = invert_tree_comparison
             (cmp, HONOR_NANS (@0)); }
    (if (ic == icmp)
     (icmp @0 @1))
    (if (ic == ncmp)
     (ncmp @0 @1)))))
 
+/* Unordered tests if either argument is a NaN.  */
+(simplify
+ (bit_ior (unordered @0 @0) (unordered @1 @1))
+ (if ((GIMPLE && types_compatible_p (TREE_TYPE (@0), TREE_TYPE (@1)))
+      || (GENERIC && TREE_TYPE (@0) == TREE_TYPE (@1)))
+  (unordered @0 @1)))
+(simplify
+ (bit_ior:c (unordered @0 @0) (unordered:c@2 @0 @1))
+ @2)
 
 /* Simplification of math builtins.  */
 
 (define_operator_list LOG BUILT_IN_LOGF BUILT_IN_LOG BUILT_IN_LOGL)
 (define_operator_list EXP BUILT_IN_EXPF BUILT_IN_EXP BUILT_IN_EXPL)
 (define_operator_list LOG2 BUILT_IN_LOG2F BUILT_IN_LOG2 BUILT_IN_LOG2L)
 (define_operator_list EXP2 BUILT_IN_EXP2F BUILT_IN_EXP2 BUILT_IN_EXP2L)
 (define_operator_list LOG10 BUILT_IN_LOG10F BUILT_IN_LOG10 BUILT_IN_LOG10L)
 (define_operator_list EXP10 BUILT_IN_EXP10F BUILT_IN_EXP10 BUILT_IN_EXP10L)
 (define_operator_list POW BUILT_IN_POWF BUILT_IN_POW BUILT_IN_POWL)
Index: gcc/testsuite/gcc.dg/pr63387.c
===================================================================
--- gcc/testsuite/gcc.dg/pr63387.c	(revision 0)
+++ gcc/testsuite/gcc.dg/pr63387.c	(working copy)
@@ -0,0 +1,26 @@
+/* { dg-do compile } */
+/* { dg-options "-O -fdump-tree-optimized" } */
+
+int f(double aaa, double bbb){
+  int xa = __builtin_isunordered(aaa, aaa);
+  int xb = __builtin_isunordered(bbb, bbb);
+  return xa | xb;
+}
+
+int g(double aaa, double bbb){
+  int xa = __builtin_isunordered(aaa, bbb);
+  int xb = __builtin_isunordered(bbb, bbb);
+  return xa | xb;
+}
+
+int h(double ccc, float ddd){
+  int xc = __builtin_isunordered(ccc, ccc);
+  int xd = __builtin_isunordered(ddd, ddd);
+  return xc | xd;
+}
+
+/* { dg-final { scan-tree-dump-not "aaa\[^\n\r\]* unord aaa" "optimized" } } */
+/* { dg-final { scan-tree-dump-not "bbb\[^\n\r\]* unord bbb" "optimized" } } */
+/* { dg-final { scan-tree-dump-times "aaa\[^\n\r\]* unord bbb" 2 "optimized" } } */
+/* { dg-final { scan-tree-dump-not "ccc\[^\n\r\]* unord ddd" "optimized" } } */
+/* { dg-final { cleanup-tree-dump "optimized" } } */

Re: [tree-optimization/63387] Recognize isunordered

[Richard Biener]

On Mon, Apr 13, 2015 at 11:37 PM, Marc Glisse <marc.glisse@inria.fr> wrote:
> On Mon, 13 Apr 2015, Marc Glisse wrote:
>
>> On Mon, 13 Apr 2015, Richard Biener wrote:
>>
>>> On Mon, Apr 13, 2015 at 2:23 PM, Marc Glisse <marc.glisse@inria.fr>
>>> wrote:
>>>>
>>>> Hello,
>>>>
>>>> just a simple pattern for match.pd. I am ignoring the issue of whether
>>>> isnan
>>>> is the same as isunordered, I am only combining isunordered together.
>>>
>>>
>>> Ok.
>>
>>
>> Oups, I am an idiot. My own comment in bugzilla shows why the patch has a
>> bug, I'll post a fixed version soon. Sorry,
>
>
> Here is the fixed patch (same ChangeLog).
>
> I copied the type compatibility check from other places in match.pd, it
> would be a good idea to provide a helper to avoid repeating this code.

Agreed and on my list of cleanups ...

> Is the new version ok?

Ok if bootstrapped / tested.

Thanks,
Richard.

> --
> Marc Glisse
> Index: gcc/match.pd
> ===================================================================
> --- gcc/match.pd        (revision 222062)
> +++ gcc/match.pd        (working copy)
> @@ -925,20 +925,29 @@ along with GCC; see the file COPYING3.
>       (ncmp @0 @1)))))
>   (simplify
>    (bit_xor (cmp @0 @1) integer_truep)
>    (with { enum tree_code ic = invert_tree_comparison
>              (cmp, HONOR_NANS (@0)); }
>     (if (ic == icmp)
>      (icmp @0 @1))
>     (if (ic == ncmp)
>      (ncmp @0 @1)))))
>
> +/* Unordered tests if either argument is a NaN.  */
> +(simplify
> + (bit_ior (unordered @0 @0) (unordered @1 @1))
> + (if ((GIMPLE && types_compatible_p (TREE_TYPE (@0), TREE_TYPE (@1)))
> +      || (GENERIC && TREE_TYPE (@0) == TREE_TYPE (@1)))
> +  (unordered @0 @1)))
> +(simplify
> + (bit_ior:c (unordered @0 @0) (unordered:c@2 @0 @1))
> + @2)
>
>  /* Simplification of math builtins.  */
>
>  (define_operator_list LOG BUILT_IN_LOGF BUILT_IN_LOG BUILT_IN_LOGL)
>  (define_operator_list EXP BUILT_IN_EXPF BUILT_IN_EXP BUILT_IN_EXPL)
>  (define_operator_list LOG2 BUILT_IN_LOG2F BUILT_IN_LOG2 BUILT_IN_LOG2L)
>  (define_operator_list EXP2 BUILT_IN_EXP2F BUILT_IN_EXP2 BUILT_IN_EXP2L)
>  (define_operator_list LOG10 BUILT_IN_LOG10F BUILT_IN_LOG10 BUILT_IN_LOG10L)
>  (define_operator_list EXP10 BUILT_IN_EXP10F BUILT_IN_EXP10 BUILT_IN_EXP10L)
>  (define_operator_list POW BUILT_IN_POWF BUILT_IN_POW BUILT_IN_POWL)
> Index: gcc/testsuite/gcc.dg/pr63387.c
> ===================================================================
> --- gcc/testsuite/gcc.dg/pr63387.c      (revision 0)
> +++ gcc/testsuite/gcc.dg/pr63387.c      (working copy)
> @@ -0,0 +1,26 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O -fdump-tree-optimized" } */
> +
> +int f(double aaa, double bbb){
> +  int xa = __builtin_isunordered(aaa, aaa);
> +  int xb = __builtin_isunordered(bbb, bbb);
> +  return xa | xb;
> +}
> +
> +int g(double aaa, double bbb){
> +  int xa = __builtin_isunordered(aaa, bbb);
> +  int xb = __builtin_isunordered(bbb, bbb);
> +  return xa | xb;
> +}
> +
> +int h(double ccc, float ddd){
> +  int xc = __builtin_isunordered(ccc, ccc);
> +  int xd = __builtin_isunordered(ddd, ddd);
> +  return xc | xd;
> +}
> +
> +/* { dg-final { scan-tree-dump-not "aaa\[^\n\r\]* unord aaa" "optimized" }
> } */
> +/* { dg-final { scan-tree-dump-not "bbb\[^\n\r\]* unord bbb" "optimized" }
> } */
> +/* { dg-final { scan-tree-dump-times "aaa\[^\n\r\]* unord bbb" 2
> "optimized" } } */
> +/* { dg-final { scan-tree-dump-not "ccc\[^\n\r\]* unord ddd" "optimized" }
> } */
> +/* { dg-final { cleanup-tree-dump "optimized" } } */
>