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 "sincos_helpers.h" 24#include <clc/clc.h> 25#include <clc/clcmacro.h> 26#include <clc/math/math.h> 27 28_CLC_OVERLOAD _CLC_DEF float cos(float x) 29{ 30 int ix = as_int(x); 31 int ax = ix & 0x7fffffff; 32 float dx = as_float(ax); 33 34 float r0, r1; 35 int regn = __clc_argReductionS(&r0, &r1, dx); 36 37 float ss = -__clc_sinf_piby4(r0, r1); 38 float cc = __clc_cosf_piby4(r0, r1); 39 40 float c = (regn & 1) != 0 ? ss : cc; 41 c = as_float(as_int(c) ^ ((regn > 1) << 31)); 42 43 c = ax >= PINFBITPATT_SP32 ? as_float(QNANBITPATT_SP32) : c; 44 45 return c; 46} 47 48_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, float, cos, float); 49 50#ifdef cl_khr_fp64 51 52#pragma OPENCL EXTENSION cl_khr_fp64 : enable 53 54_CLC_OVERLOAD _CLC_DEF double cos(double x) { 55 x = fabs(x); 56 57 double r, rr; 58 int regn; 59 60 if (x < 0x1.0p+47) 61 __clc_remainder_piby2_medium(x, &r, &rr, ®n); 62 else 63 __clc_remainder_piby2_large(x, &r, &rr, ®n); 64 65 double2 sc = __clc_sincos_piby4(r, rr); 66 sc.lo = -sc.lo; 67 68 int2 c = as_int2(regn & 1 ? sc.lo : sc.hi); 69 c.hi ^= (regn > 1) << 31; 70 71 return isnan(x) | isinf(x) ? as_double(QNANBITPATT_DP64) : as_double(c); 72} 73 74_CLC_UNARY_VECTORIZE(_CLC_OVERLOAD _CLC_DEF, double, cos, double); 75 76#endif 77 78#ifdef cl_khr_fp16 79 80#pragma OPENCL EXTENSION cl_khr_fp16 : enable 81 82_CLC_DEFINE_UNARY_BUILTIN_FP16(cos) 83 84#endif 85