summaryrefslogtreecommitdiff
path: root/src/math/fmaf.c
diff options
context:
space:
mode:
authorRich Felker <dalias@aerifal.cx>2012-03-13 01:17:53 -0400
committerRich Felker <dalias@aerifal.cx>2012-03-13 01:17:53 -0400
commitb69f695acedd4ce2798ef9ea28d834ceccc789bd (patch)
treeeafd98b9b75160210f3295ac074d699f863d958e /src/math/fmaf.c
parentd46cf2e14cc4df7cc75e77d7009fcb6df1f48a33 (diff)
downloadmusl-b69f695acedd4ce2798ef9ea28d834ceccc789bd.tar.gz
musl-b69f695acedd4ce2798ef9ea28d834ceccc789bd.tar.bz2
musl-b69f695acedd4ce2798ef9ea28d834ceccc789bd.tar.xz
musl-b69f695acedd4ce2798ef9ea28d834ceccc789bd.zip
first commit of the new libm!
thanks to the hard work of Szabolcs Nagy (nsz), identifying the best (from correctness and license standpoint) implementations from freebsd and openbsd and cleaning them up! musl should now fully support c99 float and long double math functions, and has near-complete complex math support. tgmath should also work (fully on gcc-compatible compilers, and mostly on any c99 compiler). based largely on commit 0376d44a890fea261506f1fc63833e7a686dca19 from nsz's libm git repo, with some additions (dummy versions of a few missing long double complex functions, etc.) by me. various cleanups still need to be made, including re-adding (if they're correct) some asm functions that were dropped.
Diffstat (limited to 'src/math/fmaf.c')
-rw-r--r--src/math/fmaf.c64
1 files changed, 64 insertions, 0 deletions
diff --git a/src/math/fmaf.c b/src/math/fmaf.c
new file mode 100644
index 00000000..0dccf108
--- /dev/null
+++ b/src/math/fmaf.c
@@ -0,0 +1,64 @@
+/* origin: FreeBSD /usr/src/lib/msun/src/s_fmaf.c */
+/*-
+ * Copyright (c) 2005-2011 David Schultz <das@FreeBSD.ORG>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <fenv.h>
+#include "libm.h"
+
+/*
+ * Fused multiply-add: Compute x * y + z with a single rounding error.
+ *
+ * A double has more than twice as much precision than a float, so
+ * direct double-precision arithmetic suffices, except where double
+ * rounding occurs.
+ */
+float fmaf(float x, float y, float z)
+{
+ double xy, result;
+ uint32_t hr, lr;
+
+ xy = (double)x * y;
+ result = xy + z;
+ EXTRACT_WORDS(hr, lr, result);
+ /* Common case: The double precision result is fine. */
+ if ((lr & 0x1fffffff) != 0x10000000 || /* not a halfway case */
+ (hr & 0x7ff00000) == 0x7ff00000 || /* NaN */
+ result - xy == z || /* exact */
+ fegetround() != FE_TONEAREST) /* not round-to-nearest */
+ return (result);
+
+ /*
+ * If result is inexact, and exactly halfway between two float values,
+ * we need to adjust the low-order bit in the direction of the error.
+ */
+ fesetround(FE_TOWARDZERO);
+ volatile double vxy = xy; /* XXX work around gcc CSE bug */
+ double adjusted_result = vxy + z;
+ fesetround(FE_TONEAREST);
+ if (result == adjusted_result)
+ SET_LOW_WORD(adjusted_result, lr + 1);
+ return (adjusted_result);
+}