Skip to content
Projects
Groups
Snippets
Help
Loading...
Help
Support
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
C
cpython
Project overview
Project overview
Details
Activity
Releases
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Issues
0
Issues
0
List
Boards
Labels
Milestones
Merge Requests
0
Merge Requests
0
Analytics
Analytics
Repository
Value Stream
Wiki
Wiki
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Create a new issue
Commits
Issue Boards
Open sidebar
Kirill Smelkov
cpython
Commits
5e65cd39
Commit
5e65cd39
authored
Jan 21, 2017
by
Mark Dickinson
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
Issue #29282: Backed out changeset b33012ef1417
parent
d1b230e4
Changes
6
Expand all
Hide whitespace changes
Inline
Side-by-side
Showing
6 changed files
with
1 addition
and
338 deletions
+1
-338
Doc/library/math.rst
Doc/library/math.rst
+0
-15
Doc/whatsnew/3.7.rst
Doc/whatsnew/3.7.rst
+0
-9
Lib/test/test_math.py
Lib/test/test_math.py
+0
-234
Misc/NEWS
Misc/NEWS
+0
-3
Modules/clinic/mathmodule.c.h
Modules/clinic/mathmodule.c.h
+1
-35
Modules/mathmodule.c
Modules/mathmodule.c
+0
-42
No files found.
Doc/library/math.rst
View file @
5e65cd39
...
...
@@ -57,21 +57,6 @@ Number-theoretic and representation functions
If *x* is not a float, delegates to ``x.__floor__()``, which should return an
:class:`~numbers.Integral` value.
.. function:: fma(x, y, z)
Fused multiply-add operation. Return ``(x * y) + z``, computed as though with
infinite precision and range followed by a single round to the ``float``
format. This operation often provides better accuracy than the direct
expression ``(x * y) + z``.
This function follows the specification of the fusedMultiplyAdd operation
described in the IEEE 754-2008 standard. The standard leaves one case
implementation-defined, namely the result of ``fma(0, inf, nan)``
and ``fma(inf, 0, nan)``. In these cases, ``math.fma`` returns a NaN,
and does not raise any exception.
.. versionadded:: 3.7
.. function:: fmod(x, y)
...
...
Doc/whatsnew/3.7.rst
View file @
5e65cd39
...
...
@@ -100,15 +100,6 @@ The :const:`~unittest.mock.sentinel` attributes now preserve their identity
when they are :mod:`copied <copy>` or :mod:`pickled <pickle>`.
(Contributed by Serhiy Storchaka in :issue:`20804`.)
math module
-----------
A new function :func:`~math.fma` for fused multiply-add operations has been
added. This function computes ``x * y + z`` with only a single round, and so
avoids any intermediate loss of precision. It wraps the ``fma`` function
provided by C99, and follows the specification of the IEEE 754-2008
"fusedMultiplyAdd" operation for special cases.
Optimizations
=============
...
...
Lib/test/test_math.py
View file @
5e65cd39
This diff is collapsed.
Click to expand it.
Misc/NEWS
View file @
5e65cd39
...
...
@@ -215,9 +215,6 @@ Core and Builtins
Library
-------
-
Issue
#
29282
:
Added
new
math
.
fma
function
,
wrapping
C99
's fma
operation.
-
Issue
#
29197
:
Removed
deprecated
function
ntpath
.
splitunc
().
-
Issue
#
29210
:
Removed
support
of
deprecated
argument
"exclude"
in
...
...
Modules/clinic/mathmodule.c.h
View file @
5e65cd39
...
...
@@ -80,40 +80,6 @@ PyDoc_STRVAR(math_factorial__doc__,
#define MATH_FACTORIAL_METHODDEF \
{"factorial", (PyCFunction)math_factorial, METH_O, math_factorial__doc__},
PyDoc_STRVAR
(
math_fma__doc__
,
"fma($module, x, y, z, /)
\n
"
"--
\n
"
"
\n
"
"Fused multiply-add operation. Compute (x * y) + z with a single round."
);
#define MATH_FMA_METHODDEF \
{"fma", (PyCFunction)math_fma, METH_FASTCALL, math_fma__doc__},
static
PyObject
*
math_fma_impl
(
PyObject
*
module
,
double
x
,
double
y
,
double
z
);
static
PyObject
*
math_fma
(
PyObject
*
module
,
PyObject
**
args
,
Py_ssize_t
nargs
,
PyObject
*
kwnames
)
{
PyObject
*
return_value
=
NULL
;
double
x
;
double
y
;
double
z
;
if
(
!
_PyArg_ParseStack
(
args
,
nargs
,
"ddd:fma"
,
&
x
,
&
y
,
&
z
))
{
goto
exit
;
}
if
(
!
_PyArg_NoStackKeywords
(
"fma"
,
kwnames
))
{
goto
exit
;
}
return_value
=
math_fma_impl
(
module
,
x
,
y
,
z
);
exit:
return
return_value
;
}
PyDoc_STRVAR
(
math_trunc__doc__
,
"trunc($module, x, /)
\n
"
"--
\n
"
...
...
@@ -570,4 +536,4 @@ math_isclose(PyObject *module, PyObject **args, Py_ssize_t nargs, PyObject *kwna
exit:
return
return_value
;
}
/*[clinic end generated code: output=
f428e1075d00c334
input=a9049054013a1b77]*/
/*[clinic end generated code: output=
71806f73a5c4bf0b
input=a9049054013a1b77]*/
Modules/mathmodule.c
View file @
5e65cd39
...
...
@@ -1595,47 +1595,6 @@ math_factorial(PyObject *module, PyObject *arg)
}
/*[clinic input]
math.fma
x: double
y: double
z: double
/
Fused multiply-add operation. Compute (x * y) + z with a single round.
[clinic start generated code]*/
static
PyObject
*
math_fma_impl
(
PyObject
*
module
,
double
x
,
double
y
,
double
z
)
/*[clinic end generated code: output=4fc8626dbc278d17 input=2ae8bb2a6e0f8b77]*/
{
double
r
;
r
=
fma
(
x
,
y
,
z
);
/* Fast path: if we got a finite result, we're done. */
if
(
Py_IS_FINITE
(
r
))
{
return
PyFloat_FromDouble
(
r
);
}
/* Non-finite result. Raise an exception if appropriate, else return r. */
if
(
Py_IS_NAN
(
r
))
{
if
(
!
Py_IS_NAN
(
x
)
&&
!
Py_IS_NAN
(
y
)
&&
!
Py_IS_NAN
(
z
))
{
/* NaN result from non-NaN inputs. */
PyErr_SetString
(
PyExc_ValueError
,
"invalid operation in fma"
);
return
NULL
;
}
}
else
if
(
Py_IS_FINITE
(
x
)
&&
Py_IS_FINITE
(
y
)
&&
Py_IS_FINITE
(
z
))
{
/* Infinite result from finite inputs. */
PyErr_SetString
(
PyExc_OverflowError
,
"overflow in fma"
);
return
NULL
;
}
return
PyFloat_FromDouble
(
r
);
}
/*[clinic input]
math.trunc
...
...
@@ -2265,7 +2224,6 @@ static PyMethodDef math_methods[] = {
{
"fabs"
,
math_fabs
,
METH_O
,
math_fabs_doc
},
MATH_FACTORIAL_METHODDEF
MATH_FLOOR_METHODDEF
MATH_FMA_METHODDEF
MATH_FMOD_METHODDEF
MATH_FREXP_METHODDEF
MATH_FSUM_METHODDEF
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment