[PATCH] Inline IBM long double __gcc_qsub

[PATCH] Inline IBM long double __gcc_qsub

[David Edelsohn]

    rs6000: inline ldouble __gcc_qsub

    While performing some tests of IEEE 128 float for PPC64LE, Michael
    Meissner noticed that __gcc_qsub is substantially slower than
    __gcc_qadd.  __gcc_qsub valls __gcc_add with the second operand
    negated.  Because the functions normally are invoked through
    libgcc shared object, the extra PLT overhead has a large impact
    on the overall time of the function.  Instead of trying to be
    fancy with function decorations to prevent interposition, this
    patch inlines the definition of __gcc_qadd into __gcc_qsub with
    the negation propagated through the function.

    libgcc/ChangeLog:

            * config/rs6000/ibm-ldouble.c (__gcc_qsub): Inline negated
__gcc_qadd.

diff --git a/libgcc/config/rs6000/ibm-ldouble.c
b/libgcc/config/rs6000/ibm-ldouble.c
index 4c13453f975..ed74900e5c3 100644
--- a/libgcc/config/rs6000/ibm-ldouble.c
+++ b/libgcc/config/rs6000/ibm-ldouble.c
@@ -158,9 +158,42 @@ __gcc_qadd (double a, double aa, double c, double cc)
 }

 IBM128_TYPE
-__gcc_qsub (double a, double b, double c, double d)
+__gcc_qsub (double a, double aa, double c, double cc)
 {
-  return __gcc_qadd (a, b, -c, -d);
+  double xh, xl, z, q, zz;
+
+  z = a - c;
+
+  if (nonfinite (z))
+    {
+      if (fabs (z) != inf())
+       return z;
+      z = -cc + aa - c + a;
+      if (nonfinite (z))
+       return z;
+      xh = z;  /* Will always be DBL_MAX.  */
+      zz = aa - cc;
+      if (fabs(a) > fabs(c))
+       xl = a - z - c + zz;
+      else
+       xl = -c - z + a + zz;
+    }
+  else
+    {
+      q = a - z;
+      zz = q - c + (a - (q + z)) + aa - cc;
+
+      /* Keep -0 result.  */
+      if (zz == 0.0)
+       return z;
+
+      xh = z + zz;
+      if (nonfinite (xh))
+       return xh;
+
+      xl = z - xh + zz;
+    }
+  return pack_ldouble (xh, xl);
 }

 #ifdef __NO_FPRS__

Re: [PATCH] Inline IBM long double __gcc_qsub

[Andreas Schwab]

On Aug 25 2021, David Edelsohn via Gcc-patches wrote:

>     rs6000: inline ldouble __gcc_qsub
>
>     While performing some tests of IEEE 128 float for PPC64LE, Michael
>     Meissner noticed that __gcc_qsub is substantially slower than
>     __gcc_qadd.  __gcc_qsub valls __gcc_add with the second operand
                                    __gcc_qadd

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

Re: [PATCH] Inline IBM long double __gcc_qsub

[Andreas Schwab]

On Aug 25 2021, David Edelsohn via Gcc-patches wrote:

>     rs6000: inline ldouble __gcc_qsub
>
>     While performing some tests of IEEE 128 float for PPC64LE, Michael
>     Meissner noticed that __gcc_qsub is substantially slower than
>     __gcc_qadd.  __gcc_qsub valls __gcc_add with the second operand
>     negated.  Because the functions normally are invoked through
>     libgcc shared object, the extra PLT overhead has a large impact
>     on the overall time of the function.  Instead of trying to be
>     fancy with function decorations to prevent interposition, this
>     patch inlines the definition of __gcc_qadd into __gcc_qsub with
>     the negation propagated through the function.
>
>     libgcc/ChangeLog:
>
>             * config/rs6000/ibm-ldouble.c (__gcc_qsub): Inline negated
> __gcc_qadd.

How about defining a static function that is used by both?

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

[PATCH v2] Inline IBM long double __gcc_qsub

[David Edelsohn]

    While performing some tests of IEEE 128 float for PPC64LE, Michael
    Meissner noticed that __gcc_qsub is substantially slower than
    __gcc_qadd.  __gcc_qsub calls __gcc_add with the second operand
    negated.  Because the functions normally are invoked through
    libgcc shared object, the extra PLT overhead has a large impact
    on the overall time of the function.  This patch converts
    __gcc_qadd to a static inline function invoked by __gcc_qadd
    and __gcc_qsub.

    libgcc/ChangeLog:

            * config/rs6000/ibm-ldouble.c (ldouble_qadd_internal): Rename from
            __gcc_qadd.
            (__gcc_qadd): Call ldouble_qadd_internal.
            (__gcc_qsub): Call ldouble_qadd_internal with second long double
            argument negated.

diff --git a/libgcc/config/rs6000/ibm-ldouble.c
b/libgcc/config/rs6000/ibm-ldouble.c
index 4c13453f975..0b385aa940b 100644
--- a/libgcc/config/rs6000/ibm-ldouble.c
+++ b/libgcc/config/rs6000/ibm-ldouble.c
@@ -118,8 +118,8 @@ pack_ldouble (double dh, double dl)
 }

 /* Add two 'IBM128_TYPE' values and return the result. */
-IBM128_TYPE
-__gcc_qadd (double a, double aa, double c, double cc)
+static inline IBM128_TYPE
+ldouble_qadd_internal (double a, double aa, double c, double cc)
 {
   double xh, xl, z, q, zz;

@@ -158,9 +158,15 @@ __gcc_qadd (double a, double aa, double c, double cc)
 }

 IBM128_TYPE
-__gcc_qsub (double a, double b, double c, double d)
+__gcc_qadd (double a, double aa, double c, double cc)
+{
+  return ldouble_qadd_internal (a, aa, c, cc);
+}
+
+IBM128_TYPE
+__gcc_qsub (double a, double aa, double c, double cc)
 {
-  return __gcc_qadd (a, b, -c, -d);
+  return ldouble_qadd_internal (a, aa, -c, -cc);
 }

 #ifdef __NO_FPRS__

Re: [PATCH v2] Inline IBM long double __gcc_qsub

[Segher Boessenkool]

Hi!

On Thu, Aug 26, 2021 at 02:57:35PM -0400, David Edelsohn wrote:
>             * config/rs6000/ibm-ldouble.c (ldouble_qadd_internal): Rename from
>             __gcc_qadd.
>             (__gcc_qadd): Call ldouble_qadd_internal.
>             (__gcc_qsub): Call ldouble_qadd_internal with second long double
>             argument negated.

Still looks good, please commit.  Thanks :-)

> +static inline IBM128_TYPE
> +ldouble_qadd_internal (double a, double aa, double c, double cc)

Does it end up actually inlined, or as one static function that both
__gcc_qadd and __gcc_qsub use?  This is fine for complexity, it is just
a simple tail-call jump, just wondering what the compiler thinks is best
here (it matters in other cases, if the inline function has conditional
branches for example).


Segher

Re: [PATCH v2] Inline IBM long double __gcc_qsub

[David Edelsohn]

On Thu, Aug 26, 2021 at 6:53 PM Segher Boessenkool
<segher@kernel.crashing.org> wrote:
>
> Hi!
>
> On Thu, Aug 26, 2021 at 02:57:35PM -0400, David Edelsohn wrote:
> >             * config/rs6000/ibm-ldouble.c (ldouble_qadd_internal): Rename from
> >             __gcc_qadd.
> >             (__gcc_qadd): Call ldouble_qadd_internal.
> >             (__gcc_qsub): Call ldouble_qadd_internal with second long double
> >             argument negated.
>
> Still looks good, please commit.  Thanks :-)
>
> > +static inline IBM128_TYPE
> > +ldouble_qadd_internal (double a, double aa, double c, double cc)
>
> Does it end up actually inlined, or as one static function that both
> __gcc_qadd and __gcc_qsub use?  This is fine for complexity, it is just
> a simple tail-call jump, just wondering what the compiler thinks is best
> here (it matters in other cases, if the inline function has conditional
> branches for example).

I confirmed that the implementation is inlined in both functions when
compiled with optimization, and the negation is propagated into qsub.

Thanks, David

Re: [PATCH] Inline IBM long double __gcc_qsub

[Segher Boessenkool]

Hi!

On Wed, Aug 25, 2021 at 08:23:32PM -0400, David Edelsohn wrote:
>     rs6000: inline ldouble __gcc_qsub
> 
>     While performing some tests of IEEE 128 float for PPC64LE, Michael
>     Meissner noticed that __gcc_qsub is substantially slower than
>     __gcc_qadd.  __gcc_qsub valls __gcc_add with the second operand

("calls", "__gcc_qadd")

>     negated.  Because the functions normally are invoked through
>     libgcc shared object, the extra PLT overhead has a large impact
>     on the overall time of the function.  Instead of trying to be
>     fancy with function decorations to prevent interposition, this
>     patch inlines the definition of __gcc_qadd into __gcc_qsub with
>     the negation propagated through the function.

Looks good to me, and it is a good way to resolve this.  This code is
too old (and unimportant) to do serious engineering on.  If we want
any serious optimisation on it we should do that at tree level (why does
that not happen yet anyway?), and inline all of this.  This patch is
really just to make benchmark results saner ;-)

Thanks David!


Segher