Trigonometric optimizations for POWER

1 files changed, 118 insertions, 0 deletions

diff --git a/sysdeps/powerpc/fpu/e_hypotf.c b/sysdeps/powerpc/fpu/e_hypotf.c
new file mode 100644
index 0000000000..3e6c597cc9
--- /dev/null
+++ b/sysdeps/powerpc/fpu/e_hypotf.c
@@ -0,0 +1,118 @@
+/* Pythagorean addition using floats
+   Copyright (C) 2011 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Adhemerval Zanella <azanella@br.ibm.com>, 2011
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Library General Public License as
+   published by the Free Software Foundation; either version 2 of the
+   License, or (at your option) any later version.
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Library General Public License for more details.
+
+   You should have received a copy of the GNU Library General Public
+   License along with the GNU C Library; see the file COPYING.LIB.  If not,
+   write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+   Boston, MA 02111-1307, USA.  */
+
+#include "math.h"
+#include "math_private.h"
+
+
+static const float two30  = 1.0737418e09;
+static const float two50  = 1.1259000e15;
+static const float two60  = 1.1529221e18;
+static const float two126 = 8.5070592e+37;
+static const float twoM50 = 8.8817842e-16;
+static const float twoM60 = 6.7762644e-21;
+static const float pdnum  = 1.1754939e-38;
+
+
+/* __ieee754_hypotf(x,y)
+ *
+ * This a FP only version without any FP->INT conversion.
+ * It is similar to default C version, making appropriates
+ * overflow and underflows checks as well scaling when it
+ * is needed.
+ */
+
+#ifdef _ARCH_PWR7
+/* POWER7 isinf and isnan optimizations are fast. */
+# define TEST_INF_NAN(x, y)                                      \
+   if (isinff(x) || isinff(y))                                   \
+     return INFINITY;                                            \
+   if (isnanf(x) || isnanf(y))                                   \
+     return NAN;
+# else
+/* For POWER6 and below isinf/isnan triggers LHS and PLT calls are
+ * costly (especially for POWER6). */
+# define GET_TWO_FLOAT_WORD(f1,f2,i1,i2)                         \
+ do {                                                            \
+   ieee_float_shape_type gf_u1;                                  \
+   ieee_float_shape_type gf_u2;                                  \
+   gf_u1.value = (f1);                                           \
+   gf_u2.value = (f2);                                           \
+   (i1) = gf_u1.word;                                            \
+   (i2) = gf_u2.word;                                            \
+ } while (0)
+
+# define TEST_INF_NAN(x, y)                                      \
+ do {                                                            \
+   int32_t hx, hy;                                               \
+   GET_TWO_FLOAT_WORD(x, y, hx, hy);                             \
+   if (hy > hx) {                                                \
+     uint32_t ht = hx; hx = hy; hy = ht;                         \
+   }                                                             \
+   if (hx >= 0x7f800000) {                                       \
+     if (hx == 0x7f800000 || hy == 0x7f800000)                   \
+       return INFINITY;                                          \
+     return NAN;                                                 \
+   }                                                             \
+ } while (0)
+#endif
+
+
+float
+__ieee754_hypotf (float x, float y)
+{
+  x = fabsf (x);
+  y = fabsf (y);
+
+  TEST_INF_NAN (x, y);
+
+  if (y > x)
+    {
+      float t = y;
+      y = x;
+      x = t;
+    }
+  if (y == 0.0 || (x / y) > two30)
+    {
+      return x + y;
+    }
+  if (x > two50)
+    {
+      x *= twoM60;
+      y *= twoM60;
+      return sqrtf (x * x + y * y) / twoM60;
+    }
+  if (y < twoM50)
+    {
+      if (y <= pdnum)
+	{
+	  x *= two126;
+	  y *= two126;
+	  return sqrtf (x * x + y * y) / two126;
+	}
+      else
+	{
+	  x *= two60;
+	  y *= two60;
+	  return sqrtf (x * x + y * y) / two60;
+	}
+    }
+  return sqrtf (x * x + y * y);
+}