mathops_arm.h (3884B)
1 /* Copyright (c) 2024 Arm Limited */ 2 /* 3 Redistribution and use in source and binary forms, with or without 4 modification, are permitted provided that the following conditions 5 are met: 6 7 - Redistributions of source code must retain the above copyright 8 notice, this list of conditions and the following disclaimer. 9 10 - Redistributions in binary form must reproduce the above copyright 11 notice, this list of conditions and the following disclaimer in the 12 documentation and/or other materials provided with the distribution. 13 14 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 15 ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 16 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 17 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER 18 OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, 19 EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, 20 PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR 21 PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF 22 LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING 23 NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS 24 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 25 */ 26 27 #if !defined(MATHOPS_ARM_H) 28 # define MATHOPS_ARM_H 29 30 #include "armcpu.h" 31 #include "cpu_support.h" 32 #include "opus_defines.h" 33 34 # if !defined(DISABLE_FLOAT_API) && defined(OPUS_ARM_MAY_HAVE_NEON_INTR) 35 36 #include <arm_neon.h> 37 38 static inline int32x4_t vroundf(float32x4_t x) 39 { 40 # if defined(__aarch64__) || (defined(__ARM_ARCH) && __ARM_ARCH >= 8) 41 return vcvtaq_s32_f32(x); 42 # else 43 uint32x4_t sign = vandq_u32(vreinterpretq_u32_f32(x), vdupq_n_u32(0x80000000)); 44 uint32x4_t bias = vdupq_n_u32(0x3F000000); 45 return vcvtq_s32_f32(vaddq_f32(x, vreinterpretq_f32_u32(vorrq_u32(bias, sign)))); 46 # endif 47 } 48 49 static inline float vminvf(float32x4_t a) 50 { 51 #if defined(__aarch64__) 52 return vminvq_f32(a); 53 #else 54 float32x2_t xy = vmin_f32(vget_low_f32(a), vget_high_f32(a)); 55 float x = vget_lane_f32(xy, 0); 56 float y = vget_lane_f32(xy, 1); 57 return x < y ? x : y; 58 #endif 59 } 60 61 static inline float vmaxvf(float32x4_t a) 62 { 63 #if defined(__aarch64__) 64 return vmaxvq_f32(a); 65 #else 66 float32x2_t xy = vmax_f32(vget_low_f32(a), vget_high_f32(a)); 67 float x = vget_lane_f32(xy, 0); 68 float y = vget_lane_f32(xy, 1); 69 return x > y ? x : y; 70 #endif 71 } 72 73 void celt_float2int16_neon(const float * OPUS_RESTRICT in, short * OPUS_RESTRICT out, int cnt); 74 # if defined(OPUS_HAVE_RTCD) && \ 75 (defined(OPUS_ARM_MAY_HAVE_NEON_INTR) && !defined(OPUS_ARM_PRESUME_NEON_INTR)) 76 extern void 77 (*const CELT_FLOAT2INT16_IMPL[OPUS_ARCHMASK+1])(const float * OPUS_RESTRICT in, short * OPUS_RESTRICT out, int cnt); 78 79 # define OVERRIDE_FLOAT2INT16 (1) 80 # define celt_float2int16(in, out, cnt, arch) \ 81 ((*CELT_FLOAT2INT16_IMPL[(arch)&OPUS_ARCHMASK])(in, out, cnt)) 82 83 # elif defined(OPUS_ARM_PRESUME_NEON_INTR) 84 # define OVERRIDE_FLOAT2INT16 (1) 85 # define celt_float2int16(in, out, cnt, arch) ((void)(arch), celt_float2int16_neon(in, out, cnt)) 86 # endif 87 88 int opus_limit2_checkwithin1_neon(float * samples, int cnt); 89 # if defined(OPUS_HAVE_RTCD) && \ 90 (defined(OPUS_ARM_MAY_HAVE_NEON_INTR) && !defined(OPUS_ARM_PRESUME_NEON_INTR)) 91 extern int (*const OPUS_LIMIT2_CHECKWITHIN1_IMPL[OPUS_ARCHMASK+1])(float * samples, int cnt); 92 93 # define OVERRIDE_LIMIT2_CHECKWITHIN1 (1) 94 # define opus_limit2_checkwithin1(samples, cnt, arch) \ 95 ((*OPUS_LIMIT2_CHECKWITHIN1_IMPL[(arch)&OPUS_ARCHMASK])(samples, cnt)) 96 97 # elif defined(OPUS_ARM_PRESUME_NEON_INTR) 98 # define OVERRIDE_LIMIT2_CHECKWITHIN1 (1) 99 # define opus_limit2_checkwithin1(samples, cnt, arch) ((void)(arch), opus_limit2_checkwithin1_neon(samples, cnt)) 100 # endif 101 # endif 102 103 #endif /* MATHOPS_ARM_H */