libclc/generic/lib/math/cos.cl

   1 /*
   2  * Copyright (c) 2014 Advanced Micro Devices, Inc.
   3  *
   4  * Permission is hereby granted, free of charge, to any person obtaining a copy
   5  * of this software and associated documentation files (the "Software"), to deal
   6  * in the Software without restriction, including without limitation the rights
   7  * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
   8  * copies of the Software, and to permit persons to whom the Software is
   9  * furnished to do so, subject to the following conditions:
  10  *
  11  * The above copyright notice and this permission notice shall be included in
  12  * all copies or substantial portions of the Software.
  13  *
  14  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  15  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  16  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  17  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  18  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  19  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  20  * THE SOFTWARE.
  21  */
  22
  23 #include <clc/clc.h>
  24 #include <clc/clcmacro.h>
  25
  26 #include "math.h"
  27 #include "sincos_helpers.h"
  28
  29 _CLC_OVERLOAD _CLC_DEF float cos(float x)
  30 {
  31     int ix = as_int(x);
  32     int ax = ix & 0x7fffffff;
  33     float dx = as_float(ax);
  34
  35     float r0, r1;
  36     int regn = __clc_argReductionS(&r0, &r1, dx);
  37
  38     float ss = -__clc_sinf_piby4(r0, r1);
  39     float cc =  __clc_cosf_piby4(r0, r1);
  40
  41     float c =  (regn & 1) != 0 ? ss : cc;
  42     c = as_float(as_int(c) ^ ((regn > 1) << 31));
  43
  44     c = ax >= PINFBITPATT_SP32 ? as_float(QNANBITPATT_SP32) : c;
  45
  46     return c;
  47 }
  48
  49 _CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, float, cos, float);
  50
  51 #ifdef cl_khr_fp64
  52
  53 #pragma OPENCL EXTENSION cl_khr_fp64 : enable
  54
  55 _CLC_OVERLOAD _CLC_DEF double cos(double x) {
  56     x = fabs(x);
  57
  58     double r, rr;
  59     int regn;
  60
  61     if (x < 0x1.0p+47)
  62         __clc_remainder_piby2_medium(x, &r, &rr, &regn);
  63     else
  64         __clc_remainder_piby2_large(x, &r, &rr, &regn);
  65
  66     double2 sc = __clc_sincos_piby4(r, rr);
  67     sc.lo = -sc.lo;
  68
  69     int2 c = as_int2(regn & 1 ? sc.lo : sc.hi);
  70     c.hi ^= (regn > 1) << 31;
  71
  72     return isnan(x) | isinf(x) ? as_double(QNANBITPATT_DP64) : as_double(c);
  73 }
  74
  75 _CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, double, cos, double);
  76
  77 #endif
  78
  79 _CLC_DEFINE_UNARY_BUILTIN_FP16(cos)