1vsincospi_(3MVEC) Vector Math Library Functions vsincospi_(3MVEC)
2
6 vsincospi_, vsincospif_ - vector sincospi functions
7
9 cc [ flag... ] file... -lmvec [ library... ]
10 void vsincospi_(int *n, double * restrict x, int *stridex,
12 double * restrict s, int *strides, double * restrict c,
13 int *stridec);
14 void vsincospif_(int *n, float * restrict x, int *stridex,
17 float * restrict s, int *strides, float * restrict c,
18 int *stridec);
19
22 These functions evaluate both sinpi(x) and cospi(x), defined by
23 sinpi(x) = sin(pi * x) and cospi(x) = cos(pi * x), for an entire vector
24 of values at once. The first parameter specifies the number of values
25 to compute. Subsequent parameters specify the argument and result vec‐
26 tors. Each vector is described by a pointer to the first element and a
27 stride, which is the increment between successive elements.
28 Specifically, vsincospi_(n, x, sx, s, ss, c, sc) simultaneously com‐
31 putes s[i * *ss] = sinpi(x[i * *sx]) and c[i * *sc] = cospi(x[i * *sx])
32 for each i = 0, 1, ..., *n - 1. The vsincosf_() function performs the
33 same computation for single precision data.
34 Non-exceptional results are accurate to within a unit in the last
37 place.
38
40 The element count *n must be greater than zero. The strides for the
41 argument and result arrays can be arbitrary integers, but the arrays
42 themselves must not be the same or overlap. A zero stride effectively
43 collapses an entire vector into a single element. A negative stride
44 causes a vector to be accessed in descending memory order, but note
45 that the corresponding pointer must still point to the first element of
46 the vector to be used; if the stride is negative, this will be the
47 highest-addressed element in memory. This convention differs from the
48 Level 1 BLAS, in which array parameters always refer to the lowest-
49 addressed element in memory even when negative increments are used.
50 These functions assume that the default round-to-nearest rounding
53 direction mode is in effect. On x86, these functions also assume that
54 the default round-to-64-bit rounding precision mode is in effect. The
55 result of calling a vector function with a non-default rounding mode in
56 effect is undefined.
57 These functions handle special cases and exceptions in the spirit of
60 IEEE 754. In particular,
61 o sinpi(NaN), cospi(NaN) are NaN,
63 o sinpi(±0) is ±0,
65 o sinpi(±Inf), cospi(±Inf) are NaN, and an invalid operation
67 exception is raised.
68 An application wanting to check for exceptions should call feclearex‐
71 cept(FE_ALL_EXCEPT) before calling these functions. On return, if
72 fetestexcept(FE_INVALID | FE_DIVBYZERO | FE_OVERFLOW | FE_UNDERFLOW) is
73 non-zero, an exception has been raised. The application can then exam‐
74 ine the result or argument vectors for exceptional values. Some vector
75 functions can raise the inexact exception even if all elements of the
76 argument array are such that the numerical results are exact.
77
79 See attributes(5) for descriptions of the following attributes:
80 ┌─────────────────────────────┬─────────────────────────────┐
85 │ ATTRIBUTE TYPE │ ATTRIBUTE VALUE │
86 ├─────────────────────────────┼─────────────────────────────┤
87 │Interface Stability │Committed │
88 ├─────────────────────────────┼─────────────────────────────┤
89 │MT-Level │MT-Safe │
90 └─────────────────────────────┴─────────────────────────────┘
91
93 feclearexcept(3M), fetestexcept(3M), attributes(5)
94SunOS 5.11 14 Dec 2007 vsincospi_(3MVEC)