/* ============================================================================== This file is part of the JUCE library. Copyright (c) 2022 - Raw Material Software Limited JUCE is an open source library subject to commercial or open-source licensing. The code included in this file is provided under the terms of the ISC license http://www.isc.org/downloads/software-support-policy/isc-license. Permission To use, copy, modify, and/or distribute this software for any purpose with or without fee is hereby granted provided that the above copyright notice and this permission notice appear in all copies. JUCE IS PROVIDED "AS IS" WITHOUT ANY WARRANTY, AND ALL WARRANTIES, WHETHER EXPRESSED OR IMPLIED, INCLUDING MERCHANTABILITY AND FITNESS FOR PURPOSE, ARE DISCLAIMED. ============================================================================== */ namespace juce { namespace FloatVectorHelpers { #define JUCE_INCREMENT_SRC_DEST dest += (16 / sizeof (*dest)); src += (16 / sizeof (*dest)); #define JUCE_INCREMENT_SRC1_SRC2_DEST dest += (16 / sizeof (*dest)); src1 += (16 / sizeof (*dest)); src2 += (16 / sizeof (*dest)); #define JUCE_INCREMENT_DEST dest += (16 / sizeof (*dest)); #if JUCE_USE_SSE_INTRINSICS static bool isAligned (const void* p) noexcept { return (((pointer_sized_int) p) & 15) == 0; } struct BasicOps32 { using Type = float; using ParallelType = __m128; using IntegerType = __m128; enum { numParallel = 4 }; // Integer and parallel types are the same for SSE. On neon they have different types static forcedinline IntegerType toint (ParallelType v) noexcept { return v; } static forcedinline ParallelType toflt (IntegerType v) noexcept { return v; } static forcedinline ParallelType load1 (Type v) noexcept { return _mm_load1_ps (&v); } static forcedinline ParallelType loadA (const Type* v) noexcept { return _mm_load_ps (v); } static forcedinline ParallelType loadU (const Type* v) noexcept { return _mm_loadu_ps (v); } static forcedinline void storeA (Type* dest, ParallelType a) noexcept { _mm_store_ps (dest, a); } static forcedinline void storeU (Type* dest, ParallelType a) noexcept { _mm_storeu_ps (dest, a); } static forcedinline ParallelType add (ParallelType a, ParallelType b) noexcept { return _mm_add_ps (a, b); } static forcedinline ParallelType sub (ParallelType a, ParallelType b) noexcept { return _mm_sub_ps (a, b); } static forcedinline ParallelType mul (ParallelType a, ParallelType b) noexcept { return _mm_mul_ps (a, b); } static forcedinline ParallelType max (ParallelType a, ParallelType b) noexcept { return _mm_max_ps (a, b); } static forcedinline ParallelType min (ParallelType a, ParallelType b) noexcept { return _mm_min_ps (a, b); } static forcedinline ParallelType bit_and (ParallelType a, ParallelType b) noexcept { return _mm_and_ps (a, b); } static forcedinline ParallelType bit_not (ParallelType a, ParallelType b) noexcept { return _mm_andnot_ps (a, b); } static forcedinline ParallelType bit_or (ParallelType a, ParallelType b) noexcept { return _mm_or_ps (a, b); } static forcedinline ParallelType bit_xor (ParallelType a, ParallelType b) noexcept { return _mm_xor_ps (a, b); } static forcedinline Type max (ParallelType a) noexcept { Type v[numParallel]; storeU (v, a); return jmax (v[0], v[1], v[2], v[3]); } static forcedinline Type min (ParallelType a) noexcept { Type v[numParallel]; storeU (v, a); return jmin (v[0], v[1], v[2], v[3]); } }; struct BasicOps64 { using Type = double; using ParallelType = __m128d; using IntegerType = __m128d; enum { numParallel = 2 }; // Integer and parallel types are the same for SSE. On neon they have different types static forcedinline IntegerType toint (ParallelType v) noexcept { return v; } static forcedinline ParallelType toflt (IntegerType v) noexcept { return v; } static forcedinline ParallelType load1 (Type v) noexcept { return _mm_load1_pd (&v); } static forcedinline ParallelType loadA (const Type* v) noexcept { return _mm_load_pd (v); } static forcedinline ParallelType loadU (const Type* v) noexcept { return _mm_loadu_pd (v); } static forcedinline void storeA (Type* dest, ParallelType a) noexcept { _mm_store_pd (dest, a); } static forcedinline void storeU (Type* dest, ParallelType a) noexcept { _mm_storeu_pd (dest, a); } static forcedinline ParallelType add (ParallelType a, ParallelType b) noexcept { return _mm_add_pd (a, b); } static forcedinline ParallelType sub (ParallelType a, ParallelType b) noexcept { return _mm_sub_pd (a, b); } static forcedinline ParallelType mul (ParallelType a, ParallelType b) noexcept { return _mm_mul_pd (a, b); } static forcedinline ParallelType max (ParallelType a, ParallelType b) noexcept { return _mm_max_pd (a, b); } static forcedinline ParallelType min (ParallelType a, ParallelType b) noexcept { return _mm_min_pd (a, b); } static forcedinline ParallelType bit_and (ParallelType a, ParallelType b) noexcept { return _mm_and_pd (a, b); } static forcedinline ParallelType bit_not (ParallelType a, ParallelType b) noexcept { return _mm_andnot_pd (a, b); } static forcedinline ParallelType bit_or (ParallelType a, ParallelType b) noexcept { return _mm_or_pd (a, b); } static forcedinline ParallelType bit_xor (ParallelType a, ParallelType b) noexcept { return _mm_xor_pd (a, b); } static forcedinline Type max (ParallelType a) noexcept { Type v[numParallel]; storeU (v, a); return jmax (v[0], v[1]); } static forcedinline Type min (ParallelType a) noexcept { Type v[numParallel]; storeU (v, a); return jmin (v[0], v[1]); } }; #define JUCE_BEGIN_VEC_OP \ using Mode = FloatVectorHelpers::ModeType::Mode; \ { \ const auto numLongOps = num / Mode::numParallel; #define JUCE_FINISH_VEC_OP(normalOp) \ num &= (Mode::numParallel - 1); \ if (num == 0) return; \ } \ for (auto i = (decltype (num)) 0; i < num; ++i) normalOp; #define JUCE_PERFORM_VEC_OP_DEST(normalOp, vecOp, locals, setupOp) \ JUCE_BEGIN_VEC_OP \ setupOp \ if (FloatVectorHelpers::isAligned (dest)) JUCE_VEC_LOOP (vecOp, dummy, Mode::loadA, Mode::storeA, locals, JUCE_INCREMENT_DEST) \ else JUCE_VEC_LOOP (vecOp, dummy, Mode::loadU, Mode::storeU, locals, JUCE_INCREMENT_DEST) \ JUCE_FINISH_VEC_OP (normalOp) #define JUCE_PERFORM_VEC_OP_SRC_DEST(normalOp, vecOp, locals, increment, setupOp) \ JUCE_BEGIN_VEC_OP \ setupOp \ if (FloatVectorHelpers::isAligned (dest)) \ { \ if (FloatVectorHelpers::isAligned (src)) JUCE_VEC_LOOP (vecOp, Mode::loadA, Mode::loadA, Mode::storeA, locals, increment) \ else JUCE_VEC_LOOP (vecOp, Mode::loadU, Mode::loadA, Mode::storeA, locals, increment) \ }\ else \ { \ if (FloatVectorHelpers::isAligned (src)) JUCE_VEC_LOOP (vecOp, Mode::loadA, Mode::loadU, Mode::storeU, locals, increment) \ else JUCE_VEC_LOOP (vecOp, Mode::loadU, Mode::loadU, Mode::storeU, locals, increment) \ } \ JUCE_FINISH_VEC_OP (normalOp) #define JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST(normalOp, vecOp, locals, increment, setupOp) \ JUCE_BEGIN_VEC_OP \ setupOp \ if (FloatVectorHelpers::isAligned (dest)) \ { \ if (FloatVectorHelpers::isAligned (src1)) \ { \ if (FloatVectorHelpers::isAligned (src2)) JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadA, Mode::loadA, Mode::storeA, locals, increment) \ else JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadA, Mode::loadU, Mode::storeA, locals, increment) \ } \ else \ { \ if (FloatVectorHelpers::isAligned (src2)) JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadU, Mode::loadA, Mode::storeA, locals, increment) \ else JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadU, Mode::loadU, Mode::storeA, locals, increment) \ } \ } \ else \ { \ if (FloatVectorHelpers::isAligned (src1)) \ { \ if (FloatVectorHelpers::isAligned (src2)) JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadA, Mode::loadA, Mode::storeU, locals, increment) \ else JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadA, Mode::loadU, Mode::storeU, locals, increment) \ } \ else \ { \ if (FloatVectorHelpers::isAligned (src2)) JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadU, Mode::loadA, Mode::storeU, locals, increment) \ else JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadU, Mode::loadU, Mode::storeU, locals, increment) \ } \ } \ JUCE_FINISH_VEC_OP (normalOp) #define JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST_DEST(normalOp, vecOp, locals, increment, setupOp) \ JUCE_BEGIN_VEC_OP \ setupOp \ if (FloatVectorHelpers::isAligned (dest)) \ { \ if (FloatVectorHelpers::isAligned (src1)) \ { \ if (FloatVectorHelpers::isAligned (src2)) JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadA, Mode::loadA, Mode::loadA, Mode::storeA, locals, increment) \ else JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadA, Mode::loadU, Mode::loadA, Mode::storeA, locals, increment) \ } \ else \ { \ if (FloatVectorHelpers::isAligned (src2)) JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadU, Mode::loadA, Mode::loadA, Mode::storeA, locals, increment) \ else JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadU, Mode::loadU, Mode::loadA, Mode::storeA, locals, increment) \ } \ } \ else \ { \ if (FloatVectorHelpers::isAligned (src1)) \ { \ if (FloatVectorHelpers::isAligned (src2)) JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadA, Mode::loadA, Mode::loadU, Mode::storeU, locals, increment) \ else JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadA, Mode::loadU, Mode::loadU, Mode::storeU, locals, increment) \ } \ else \ { \ if (FloatVectorHelpers::isAligned (src2)) JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadU, Mode::loadA, Mode::loadU, Mode::storeU, locals, increment) \ else JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadU, Mode::loadU, Mode::loadU, Mode::storeU, locals, increment) \ } \ } \ JUCE_FINISH_VEC_OP (normalOp) //============================================================================== #elif JUCE_USE_ARM_NEON struct BasicOps32 { using Type = float; using ParallelType = float32x4_t; using IntegerType = uint32x4_t; union signMaskUnion { ParallelType f; IntegerType i; }; enum { numParallel = 4 }; static forcedinline IntegerType toint (ParallelType v) noexcept { signMaskUnion u; u.f = v; return u.i; } static forcedinline ParallelType toflt (IntegerType v) noexcept { signMaskUnion u; u.i = v; return u.f; } static forcedinline ParallelType load1 (Type v) noexcept { return vld1q_dup_f32 (&v); } static forcedinline ParallelType loadA (const Type* v) noexcept { return vld1q_f32 (v); } static forcedinline ParallelType loadU (const Type* v) noexcept { return vld1q_f32 (v); } static forcedinline void storeA (Type* dest, ParallelType a) noexcept { vst1q_f32 (dest, a); } static forcedinline void storeU (Type* dest, ParallelType a) noexcept { vst1q_f32 (dest, a); } static forcedinline ParallelType add (ParallelType a, ParallelType b) noexcept { return vaddq_f32 (a, b); } static forcedinline ParallelType sub (ParallelType a, ParallelType b) noexcept { return vsubq_f32 (a, b); } static forcedinline ParallelType mul (ParallelType a, ParallelType b) noexcept { return vmulq_f32 (a, b); } static forcedinline ParallelType max (ParallelType a, ParallelType b) noexcept { return vmaxq_f32 (a, b); } static forcedinline ParallelType min (ParallelType a, ParallelType b) noexcept { return vminq_f32 (a, b); } static forcedinline ParallelType bit_and (ParallelType a, ParallelType b) noexcept { return toflt (vandq_u32 (toint (a), toint (b))); } static forcedinline ParallelType bit_not (ParallelType a, ParallelType b) noexcept { return toflt (vbicq_u32 (toint (a), toint (b))); } static forcedinline ParallelType bit_or (ParallelType a, ParallelType b) noexcept { return toflt (vorrq_u32 (toint (a), toint (b))); } static forcedinline ParallelType bit_xor (ParallelType a, ParallelType b) noexcept { return toflt (veorq_u32 (toint (a), toint (b))); } static forcedinline Type max (ParallelType a) noexcept { Type v[numParallel]; storeU (v, a); return jmax (v[0], v[1], v[2], v[3]); } static forcedinline Type min (ParallelType a) noexcept { Type v[numParallel]; storeU (v, a); return jmin (v[0], v[1], v[2], v[3]); } }; struct BasicOps64 { using Type = double; using ParallelType = double; using IntegerType = uint64; union signMaskUnion { ParallelType f; IntegerType i; }; enum { numParallel = 1 }; static forcedinline IntegerType toint (ParallelType v) noexcept { signMaskUnion u; u.f = v; return u.i; } static forcedinline ParallelType toflt (IntegerType v) noexcept { signMaskUnion u; u.i = v; return u.f; } static forcedinline ParallelType load1 (Type v) noexcept { return v; } static forcedinline ParallelType loadA (const Type* v) noexcept { return *v; } static forcedinline ParallelType loadU (const Type* v) noexcept { return *v; } static forcedinline void storeA (Type* dest, ParallelType a) noexcept { *dest = a; } static forcedinline void storeU (Type* dest, ParallelType a) noexcept { *dest = a; } static forcedinline ParallelType add (ParallelType a, ParallelType b) noexcept { return a + b; } static forcedinline ParallelType sub (ParallelType a, ParallelType b) noexcept { return a - b; } static forcedinline ParallelType mul (ParallelType a, ParallelType b) noexcept { return a * b; } static forcedinline ParallelType max (ParallelType a, ParallelType b) noexcept { return jmax (a, b); } static forcedinline ParallelType min (ParallelType a, ParallelType b) noexcept { return jmin (a, b); } static forcedinline ParallelType bit_and (ParallelType a, ParallelType b) noexcept { return toflt (toint (a) & toint (b)); } static forcedinline ParallelType bit_not (ParallelType a, ParallelType b) noexcept { return toflt ((~toint (a)) & toint (b)); } static forcedinline ParallelType bit_or (ParallelType a, ParallelType b) noexcept { return toflt (toint (a) | toint (b)); } static forcedinline ParallelType bit_xor (ParallelType a, ParallelType b) noexcept { return toflt (toint (a) ^ toint (b)); } static forcedinline Type max (ParallelType a) noexcept { return a; } static forcedinline Type min (ParallelType a) noexcept { return a; } }; #define JUCE_BEGIN_VEC_OP \ using Mode = FloatVectorHelpers::ModeType::Mode; \ if (Mode::numParallel > 1) \ { \ const auto numLongOps = num / Mode::numParallel; #define JUCE_FINISH_VEC_OP(normalOp) \ num &= (Mode::numParallel - 1); \ if (num == 0) return; \ } \ for (auto i = (decltype (num)) 0; i < num; ++i) normalOp; #define JUCE_PERFORM_VEC_OP_DEST(normalOp, vecOp, locals, setupOp) \ JUCE_BEGIN_VEC_OP \ setupOp \ JUCE_VEC_LOOP (vecOp, dummy, Mode::loadU, Mode::storeU, locals, JUCE_INCREMENT_DEST) \ JUCE_FINISH_VEC_OP (normalOp) #define JUCE_PERFORM_VEC_OP_SRC_DEST(normalOp, vecOp, locals, increment, setupOp) \ JUCE_BEGIN_VEC_OP \ setupOp \ JUCE_VEC_LOOP (vecOp, Mode::loadU, Mode::loadU, Mode::storeU, locals, increment) \ JUCE_FINISH_VEC_OP (normalOp) #define JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST(normalOp, vecOp, locals, increment, setupOp) \ JUCE_BEGIN_VEC_OP \ setupOp \ JUCE_VEC_LOOP_TWO_SOURCES (vecOp, Mode::loadU, Mode::loadU, Mode::storeU, locals, increment) \ JUCE_FINISH_VEC_OP (normalOp) #define JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST_DEST(normalOp, vecOp, locals, increment, setupOp) \ JUCE_BEGIN_VEC_OP \ setupOp \ JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD (vecOp, Mode::loadU, Mode::loadU, Mode::loadU, Mode::storeU, locals, increment) \ JUCE_FINISH_VEC_OP (normalOp) //============================================================================== #else #define JUCE_PERFORM_VEC_OP_DEST(normalOp, vecOp, locals, setupOp) \ for (auto i = (decltype (num)) 0; i < num; ++i) normalOp; #define JUCE_PERFORM_VEC_OP_SRC_DEST(normalOp, vecOp, locals, increment, setupOp) \ for (auto i = (decltype (num)) 0; i < num; ++i) normalOp; #define JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST(normalOp, vecOp, locals, increment, setupOp) \ for (auto i = (decltype (num)) 0; i < num; ++i) normalOp; #define JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST_DEST(normalOp, vecOp, locals, increment, setupOp) \ for (auto i = (decltype (num)) 0; i < num; ++i) normalOp; #endif //============================================================================== #define JUCE_VEC_LOOP(vecOp, srcLoad, dstLoad, dstStore, locals, increment) \ for (auto i = (decltype (numLongOps)) 0; i < numLongOps; ++i) \ { \ locals (srcLoad, dstLoad); \ dstStore (dest, vecOp); \ increment; \ } #define JUCE_VEC_LOOP_TWO_SOURCES(vecOp, src1Load, src2Load, dstStore, locals, increment) \ for (auto i = (decltype (numLongOps)) 0; i < numLongOps; ++i) \ { \ locals (src1Load, src2Load); \ dstStore (dest, vecOp); \ increment; \ } #define JUCE_VEC_LOOP_TWO_SOURCES_WITH_DEST_LOAD(vecOp, src1Load, src2Load, dstLoad, dstStore, locals, increment) \ for (auto i = (decltype (numLongOps)) 0; i < numLongOps; ++i) \ { \ locals (src1Load, src2Load, dstLoad); \ dstStore (dest, vecOp); \ increment; \ } #define JUCE_LOAD_NONE(srcLoad, dstLoad) #define JUCE_LOAD_DEST(srcLoad, dstLoad) const Mode::ParallelType d = dstLoad (dest); #define JUCE_LOAD_SRC(srcLoad, dstLoad) const Mode::ParallelType s = srcLoad (src); #define JUCE_LOAD_SRC1_SRC2(src1Load, src2Load) const Mode::ParallelType s1 = src1Load (src1), s2 = src2Load (src2); #define JUCE_LOAD_SRC1_SRC2_DEST(src1Load, src2Load, dstLoad) const Mode::ParallelType d = dstLoad (dest), s1 = src1Load (src1), s2 = src2Load (src2); #define JUCE_LOAD_SRC_DEST(srcLoad, dstLoad) const Mode::ParallelType d = dstLoad (dest), s = srcLoad (src); union signMask32 { float f; uint32 i; }; union signMask64 { double d; uint64 i; }; #if JUCE_USE_SSE_INTRINSICS || JUCE_USE_ARM_NEON template struct ModeType { using Mode = BasicOps32; }; template <> struct ModeType<8> { using Mode = BasicOps64; }; template struct MinMax { using Type = typename Mode::Type; using ParallelType = typename Mode::ParallelType; template static Type findMinOrMax (const Type* src, Size num, const bool isMinimum) noexcept { auto numLongOps = num / Mode::numParallel; if (numLongOps > 1) { ParallelType val; #if ! JUCE_USE_ARM_NEON if (isAligned (src)) { val = Mode::loadA (src); if (isMinimum) { while (--numLongOps > 0) { src += Mode::numParallel; val = Mode::min (val, Mode::loadA (src)); } } else { while (--numLongOps > 0) { src += Mode::numParallel; val = Mode::max (val, Mode::loadA (src)); } } } else #endif { val = Mode::loadU (src); if (isMinimum) { while (--numLongOps > 0) { src += Mode::numParallel; val = Mode::min (val, Mode::loadU (src)); } } else { while (--numLongOps > 0) { src += Mode::numParallel; val = Mode::max (val, Mode::loadU (src)); } } } Type result = isMinimum ? Mode::min (val) : Mode::max (val); num &= (Mode::numParallel - 1); src += Mode::numParallel; for (auto i = (decltype (num)) 0; i < num; ++i) result = isMinimum ? jmin (result, src[i]) : jmax (result, src[i]); return result; } if (num <= 0) return 0; return isMinimum ? *std::min_element (src, src + num) : *std::max_element (src, src + num); } template static Range findMinAndMax (const Type* src, Size num) noexcept { auto numLongOps = num / Mode::numParallel; if (numLongOps > 1) { ParallelType mn, mx; #if ! JUCE_USE_ARM_NEON if (isAligned (src)) { mn = Mode::loadA (src); mx = mn; while (--numLongOps > 0) { src += Mode::numParallel; const ParallelType v = Mode::loadA (src); mn = Mode::min (mn, v); mx = Mode::max (mx, v); } } else #endif { mn = Mode::loadU (src); mx = mn; while (--numLongOps > 0) { src += Mode::numParallel; const ParallelType v = Mode::loadU (src); mn = Mode::min (mn, v); mx = Mode::max (mx, v); } } Range result (Mode::min (mn), Mode::max (mx)); num &= (Mode::numParallel - 1); src += Mode::numParallel; for (auto i = (decltype (num)) 0; i < num; ++i) result = result.getUnionWith (src[i]); return result; } return Range::findMinAndMax (src, num); } }; #endif //============================================================================== namespace { template void clear (float* dest, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vclr (dest, 1, (vDSP_Length) num); #else zeromem (dest, (size_t) num * sizeof (float)); #endif } template void clear (double* dest, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vclrD (dest, 1, (vDSP_Length) num); #else zeromem (dest, (size_t) num * sizeof (double)); #endif } template void fill (float* dest, float valueToFill, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vfill (&valueToFill, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_DEST (dest[i] = valueToFill, val, JUCE_LOAD_NONE, const Mode::ParallelType val = Mode::load1 (valueToFill);) #endif } template void fill (double* dest, double valueToFill, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vfillD (&valueToFill, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_DEST (dest[i] = valueToFill, val, JUCE_LOAD_NONE, const Mode::ParallelType val = Mode::load1 (valueToFill);) #endif } template void copyWithMultiply (float* dest, const float* src, float multiplier, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsmul (src, 1, &multiplier, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = src[i] * multiplier, Mode::mul (mult, s), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) #endif } template void copyWithMultiply (double* dest, const double* src, double multiplier, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsmulD (src, 1, &multiplier, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = src[i] * multiplier, Mode::mul (mult, s), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) #endif } template void add (float* dest, float amount, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsadd (dest, 1, &amount, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_DEST (dest[i] += amount, Mode::add (d, amountToAdd), JUCE_LOAD_DEST, const Mode::ParallelType amountToAdd = Mode::load1 (amount);) #endif } template void add (double* dest, double amount, Size num) noexcept { JUCE_PERFORM_VEC_OP_DEST (dest[i] += amount, Mode::add (d, amountToAdd), JUCE_LOAD_DEST, const Mode::ParallelType amountToAdd = Mode::load1 (amount);) } template void add (float* dest, const float* src, float amount, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsadd (src, 1, &amount, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = src[i] + amount, Mode::add (am, s), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType am = Mode::load1 (amount);) #endif } template void add (double* dest, const double* src, double amount, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsaddD (src, 1, &amount, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = src[i] + amount, Mode::add (am, s), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType am = Mode::load1 (amount);) #endif } template void add (float* dest, const float* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vadd (src, 1, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] += src[i], Mode::add (d, s), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, ) #endif } template void add (double* dest, const double* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vaddD (src, 1, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] += src[i], Mode::add (d, s), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, ) #endif } template void add (float* dest, const float* src1, const float* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vadd (src1, 1, src2, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = src1[i] + src2[i], Mode::add (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void add (double* dest, const double* src1, const double* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vaddD (src1, 1, src2, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = src1[i] + src2[i], Mode::add (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void subtract (float* dest, const float* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsub (src, 1, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] -= src[i], Mode::sub (d, s), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, ) #endif } template void subtract (double* dest, const double* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsubD (src, 1, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] -= src[i], Mode::sub (d, s), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, ) #endif } template void subtract (float* dest, const float* src1, const float* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsub (src2, 1, src1, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = src1[i] - src2[i], Mode::sub (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void subtract (double* dest, const double* src1, const double* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsubD (src2, 1, src1, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = src1[i] - src2[i], Mode::sub (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void addWithMultiply (float* dest, const float* src, float multiplier, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsma (src, 1, &multiplier, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] += src[i] * multiplier, Mode::add (d, Mode::mul (mult, s)), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) #endif } template void addWithMultiply (double* dest, const double* src, double multiplier, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsmaD (src, 1, &multiplier, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] += src[i] * multiplier, Mode::add (d, Mode::mul (mult, s)), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) #endif } template void addWithMultiply (float* dest, const float* src1, const float* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vma ((float*) src1, 1, (float*) src2, 1, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST_DEST (dest[i] += src1[i] * src2[i], Mode::add (d, Mode::mul (s1, s2)), JUCE_LOAD_SRC1_SRC2_DEST, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void addWithMultiply (double* dest, const double* src1, const double* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vmaD ((double*) src1, 1, (double*) src2, 1, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST_DEST (dest[i] += src1[i] * src2[i], Mode::add (d, Mode::mul (s1, s2)), JUCE_LOAD_SRC1_SRC2_DEST, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void subtractWithMultiply (float* dest, const float* src, float multiplier, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] -= src[i] * multiplier, Mode::sub (d, Mode::mul (mult, s)), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) } template void subtractWithMultiply (double* dest, const double* src, double multiplier, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] -= src[i] * multiplier, Mode::sub (d, Mode::mul (mult, s)), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) } template void subtractWithMultiply (float* dest, const float* src1, const float* src2, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST_DEST (dest[i] -= src1[i] * src2[i], Mode::sub (d, Mode::mul (s1, s2)), JUCE_LOAD_SRC1_SRC2_DEST, JUCE_INCREMENT_SRC1_SRC2_DEST, ) } template void subtractWithMultiply (double* dest, const double* src1, const double* src2, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST_DEST (dest[i] -= src1[i] * src2[i], Mode::sub (d, Mode::mul (s1, s2)), JUCE_LOAD_SRC1_SRC2_DEST, JUCE_INCREMENT_SRC1_SRC2_DEST, ) } template void multiply (float* dest, const float* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vmul (src, 1, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] *= src[i], Mode::mul (d, s), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, ) #endif } template void multiply (double* dest, const double* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vmulD (src, 1, dest, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] *= src[i], Mode::mul (d, s), JUCE_LOAD_SRC_DEST, JUCE_INCREMENT_SRC_DEST, ) #endif } template void multiply (float* dest, const float* src1, const float* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vmul (src1, 1, src2, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = src1[i] * src2[i], Mode::mul (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void multiply (double* dest, const double* src1, const double* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vmulD (src1, 1, src2, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = src1[i] * src2[i], Mode::mul (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void multiply (float* dest, float multiplier, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsmul (dest, 1, &multiplier, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_DEST (dest[i] *= multiplier, Mode::mul (d, mult), JUCE_LOAD_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) #endif } template void multiply (double* dest, double multiplier, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vsmulD (dest, 1, &multiplier, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_DEST (dest[i] *= multiplier, Mode::mul (d, mult), JUCE_LOAD_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) #endif } template void multiply (float* dest, const float* src, float multiplier, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = src[i] * multiplier, Mode::mul (mult, s), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) } template void multiply (double* dest, const double* src, double multiplier, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = src[i] * multiplier, Mode::mul (mult, s), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) } template void negate (float* dest, const float* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vneg ((float*) src, 1, dest, 1, (vDSP_Length) num); #else copyWithMultiply (dest, src, -1.0f, num); #endif } template void negate (double* dest, const double* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vnegD ((double*) src, 1, dest, 1, (vDSP_Length) num); #else copyWithMultiply (dest, src, -1.0f, num); #endif } template void abs (float* dest, const float* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vabs ((float*) src, 1, dest, 1, (vDSP_Length) num); #else FloatVectorHelpers::signMask32 signMask; signMask.i = 0x7fffffffUL; JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = std::abs (src[i]), Mode::bit_and (s, mask), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mask = Mode::load1 (signMask.f);) ignoreUnused (signMask); #endif } template void abs (double* dest, const double* src, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vabsD ((double*) src, 1, dest, 1, (vDSP_Length) num); #else FloatVectorHelpers::signMask64 signMask; signMask.i = 0x7fffffffffffffffULL; JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = std::abs (src[i]), Mode::bit_and (s, mask), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mask = Mode::load1 (signMask.d);) ignoreUnused (signMask); #endif } template void min (float* dest, const float* src, float comp, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = jmin (src[i], comp), Mode::min (s, cmp), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType cmp = Mode::load1 (comp);) } template void min (double* dest, const double* src, double comp, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = jmin (src[i], comp), Mode::min (s, cmp), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType cmp = Mode::load1 (comp);) } template void min (float* dest, const float* src1, const float* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vmin ((float*) src1, 1, (float*) src2, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = jmin (src1[i], src2[i]), Mode::min (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void min (double* dest, const double* src1, const double* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vminD ((double*) src1, 1, (double*) src2, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = jmin (src1[i], src2[i]), Mode::min (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void max (float* dest, const float* src, float comp, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = jmax (src[i], comp), Mode::max (s, cmp), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType cmp = Mode::load1 (comp);) } template void max (double* dest, const double* src, double comp, Size num) noexcept { JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = jmax (src[i], comp), Mode::max (s, cmp), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType cmp = Mode::load1 (comp);) } template void max (float* dest, const float* src1, const float* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vmax ((float*) src1, 1, (float*) src2, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = jmax (src1[i], src2[i]), Mode::max (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void max (double* dest, const double* src1, const double* src2, Size num) noexcept { #if JUCE_USE_VDSP_FRAMEWORK vDSP_vmaxD ((double*) src1, 1, (double*) src2, 1, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC1_SRC2_DEST (dest[i] = jmax (src1[i], src2[i]), Mode::max (s1, s2), JUCE_LOAD_SRC1_SRC2, JUCE_INCREMENT_SRC1_SRC2_DEST, ) #endif } template void clip (float* dest, const float* src, float low, float high, Size num) noexcept { jassert (high >= low); #if JUCE_USE_VDSP_FRAMEWORK vDSP_vclip ((float*) src, 1, &low, &high, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = jmax (jmin (src[i], high), low), Mode::max (Mode::min (s, hi), lo), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType lo = Mode::load1 (low); const Mode::ParallelType hi = Mode::load1 (high);) #endif } template void clip (double* dest, const double* src, double low, double high, Size num) noexcept { jassert (high >= low); #if JUCE_USE_VDSP_FRAMEWORK vDSP_vclipD ((double*) src, 1, &low, &high, dest, 1, (vDSP_Length) num); #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = jmax (jmin (src[i], high), low), Mode::max (Mode::min (s, hi), lo), JUCE_LOAD_SRC, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType lo = Mode::load1 (low); const Mode::ParallelType hi = Mode::load1 (high);) #endif } template Range findMinAndMax (const float* src, Size num) noexcept { #if JUCE_USE_SSE_INTRINSICS || JUCE_USE_ARM_NEON return FloatVectorHelpers::MinMax::findMinAndMax (src, num); #else return Range::findMinAndMax (src, num); #endif } template Range findMinAndMax (const double* src, Size num) noexcept { #if JUCE_USE_SSE_INTRINSICS || JUCE_USE_ARM_NEON return FloatVectorHelpers::MinMax::findMinAndMax (src, num); #else return Range::findMinAndMax (src, num); #endif } template float findMinimum (const float* src, Size num) noexcept { #if JUCE_USE_SSE_INTRINSICS || JUCE_USE_ARM_NEON return FloatVectorHelpers::MinMax::findMinOrMax (src, num, true); #else return juce::findMinimum (src, num); #endif } template double findMinimum (const double* src, Size num) noexcept { #if JUCE_USE_SSE_INTRINSICS || JUCE_USE_ARM_NEON return FloatVectorHelpers::MinMax::findMinOrMax (src, num, true); #else return juce::findMinimum (src, num); #endif } template float findMaximum (const float* src, Size num) noexcept { #if JUCE_USE_SSE_INTRINSICS || JUCE_USE_ARM_NEON return FloatVectorHelpers::MinMax::findMinOrMax (src, num, false); #else return juce::findMaximum (src, num); #endif } template double findMaximum (const double* src, Size num) noexcept { #if JUCE_USE_SSE_INTRINSICS || JUCE_USE_ARM_NEON return FloatVectorHelpers::MinMax::findMinOrMax (src, num, false); #else return juce::findMaximum (src, num); #endif } template void convertFixedToFloat (float* dest, const int* src, float multiplier, Size num) noexcept { #if JUCE_USE_ARM_NEON JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = (float) src[i] * multiplier, vmulq_n_f32 (vcvtq_f32_s32 (vld1q_s32 (src)), multiplier), JUCE_LOAD_NONE, JUCE_INCREMENT_SRC_DEST, ) #else JUCE_PERFORM_VEC_OP_SRC_DEST (dest[i] = (float) src[i] * multiplier, Mode::mul (mult, _mm_cvtepi32_ps (_mm_loadu_si128 (reinterpret_cast (src)))), JUCE_LOAD_NONE, JUCE_INCREMENT_SRC_DEST, const Mode::ParallelType mult = Mode::load1 (multiplier);) #endif } } // namespace } // namespace FloatVectorHelpers //============================================================================== template void JUCE_CALLTYPE FloatVectorOperationsBase::clear (FloatType* dest, CountType numValues) noexcept { FloatVectorHelpers::clear (dest, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::fill (FloatType* dest, FloatType valueToFill, CountType numValues) noexcept { FloatVectorHelpers::fill (dest, valueToFill, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::copy (FloatType* dest, const FloatType* src, CountType numValues) noexcept { memcpy (dest, src, (size_t) numValues * sizeof (FloatType)); } template void JUCE_CALLTYPE FloatVectorOperationsBase::copyWithMultiply (FloatType* dest, const FloatType* src, FloatType multiplier, CountType numValues) noexcept { FloatVectorHelpers::copyWithMultiply (dest, src, multiplier, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::add (FloatType* dest, FloatType amountToAdd, CountType numValues) noexcept { FloatVectorHelpers::add (dest, amountToAdd, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::add (FloatType* dest, const FloatType* src, FloatType amount, CountType numValues) noexcept { FloatVectorHelpers::add (dest, src, amount, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::add (FloatType* dest, const FloatType* src, CountType numValues) noexcept { FloatVectorHelpers::add (dest, src, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::add (FloatType* dest, const FloatType* src1, const FloatType* src2, CountType num) noexcept { FloatVectorHelpers::add (dest, src1, src2, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::subtract (FloatType* dest, const FloatType* src, CountType numValues) noexcept { FloatVectorHelpers::subtract (dest, src, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::subtract (FloatType* dest, const FloatType* src1, const FloatType* src2, CountType num) noexcept { FloatVectorHelpers::subtract (dest, src1, src2, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::addWithMultiply (FloatType* dest, const FloatType* src, FloatType multiplier, CountType numValues) noexcept { FloatVectorHelpers::addWithMultiply (dest, src, multiplier, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::addWithMultiply (FloatType* dest, const FloatType* src1, const FloatType* src2, CountType num) noexcept { FloatVectorHelpers::addWithMultiply (dest, src1, src2, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::subtractWithMultiply (FloatType* dest, const FloatType* src, FloatType multiplier, CountType numValues) noexcept { FloatVectorHelpers::subtractWithMultiply (dest, src, multiplier, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::subtractWithMultiply (FloatType* dest, const FloatType* src1, const FloatType* src2, CountType num) noexcept { FloatVectorHelpers::subtractWithMultiply (dest, src1, src2, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::multiply (FloatType* dest, const FloatType* src, CountType numValues) noexcept { FloatVectorHelpers::multiply (dest, src, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::multiply (FloatType* dest, const FloatType* src1, const FloatType* src2, CountType numValues) noexcept { FloatVectorHelpers::multiply (dest, src1, src2, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::multiply (FloatType* dest, FloatType multiplier, CountType numValues) noexcept { FloatVectorHelpers::multiply (dest, multiplier, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::multiply (FloatType* dest, const FloatType* src, FloatType multiplier, CountType num) noexcept { FloatVectorHelpers::multiply (dest, src, multiplier, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::negate (FloatType* dest, const FloatType* src, CountType numValues) noexcept { FloatVectorHelpers::negate (dest, src, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::abs (FloatType* dest, const FloatType* src, CountType numValues) noexcept { FloatVectorHelpers::abs (dest, src, numValues); } template void JUCE_CALLTYPE FloatVectorOperationsBase::min (FloatType* dest, const FloatType* src, FloatType comp, CountType num) noexcept { FloatVectorHelpers::min (dest, src, comp, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::min (FloatType* dest, const FloatType* src1, const FloatType* src2, CountType num) noexcept { FloatVectorHelpers::min (dest, src1, src2, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::max (FloatType* dest, const FloatType* src, FloatType comp, CountType num) noexcept { FloatVectorHelpers::max (dest, src, comp, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::max (FloatType* dest, const FloatType* src1, const FloatType* src2, CountType num) noexcept { FloatVectorHelpers::max (dest, src1, src2, num); } template void JUCE_CALLTYPE FloatVectorOperationsBase::clip (FloatType* dest, const FloatType* src, FloatType low, FloatType high, CountType num) noexcept { FloatVectorHelpers::clip (dest, src, low, high, num); } template Range JUCE_CALLTYPE FloatVectorOperationsBase::findMinAndMax (const FloatType* src, CountType numValues) noexcept { return FloatVectorHelpers::findMinAndMax (src, numValues); } template FloatType JUCE_CALLTYPE FloatVectorOperationsBase::findMinimum (const FloatType* src, CountType numValues) noexcept { return FloatVectorHelpers::findMinimum (src, numValues); } template FloatType JUCE_CALLTYPE FloatVectorOperationsBase::findMaximum (const FloatType* src, CountType numValues) noexcept { return FloatVectorHelpers::findMaximum (src, numValues); } template struct FloatVectorOperationsBase; template struct FloatVectorOperationsBase; template struct FloatVectorOperationsBase; template struct FloatVectorOperationsBase; void JUCE_CALLTYPE FloatVectorOperations::convertFixedToFloat (float* dest, const int* src, float multiplier, size_t num) noexcept { FloatVectorHelpers::convertFixedToFloat (dest, src, multiplier, num); } void JUCE_CALLTYPE FloatVectorOperations::convertFixedToFloat (float* dest, const int* src, float multiplier, int num) noexcept { FloatVectorHelpers::convertFixedToFloat (dest, src, multiplier, num); } intptr_t JUCE_CALLTYPE FloatVectorOperations::getFpStatusRegister() noexcept { intptr_t fpsr = 0; #if JUCE_INTEL && JUCE_USE_SSE_INTRINSICS fpsr = static_cast (_mm_getcsr()); #elif defined(__arm64__) || defined(__aarch64__) || JUCE_USE_ARM_NEON #if defined(__arm64__) || defined(__aarch64__) asm volatile("mrs %0, fpcr" : "=r"(fpsr)); #elif JUCE_USE_ARM_NEON asm volatile("vmrs %0, fpscr" : "=r"(fpsr)); #endif #else #if ! (defined(JUCE_INTEL) || defined(JUCE_ARM)) jassertfalse; // No support for getting the floating point status register for your platform #endif #endif return fpsr; } void JUCE_CALLTYPE FloatVectorOperations::setFpStatusRegister (intptr_t fpsr) noexcept { #if JUCE_INTEL && JUCE_USE_SSE_INTRINSICS // the volatile keyword here is needed to workaround a bug in AppleClang 13.0 // which aggressively optimises away the variable otherwise volatile auto fpsr_w = static_cast (fpsr); _mm_setcsr (fpsr_w); #elif defined(__arm64__) || defined(__aarch64__) || JUCE_USE_ARM_NEON #if defined(__arm64__) || defined(__aarch64__) asm volatile("msr fpcr, %0" : : "ri"(fpsr)); #elif JUCE_USE_ARM_NEON asm volatile("vmsr fpscr, %0" : : "ri"(fpsr)); #endif #else #if ! (defined(JUCE_INTEL) || defined(JUCE_ARM)) jassertfalse; // No support for getting the floating point status register for your platform #endif ignoreUnused (fpsr); #endif } void JUCE_CALLTYPE FloatVectorOperations::enableFlushToZeroMode (bool shouldEnable) noexcept { #if JUCE_USE_SSE_INTRINSICS || (JUCE_USE_ARM_NEON || defined(__arm64__) || defined(__aarch64__)) #if JUCE_USE_SSE_INTRINSICS intptr_t mask = _MM_FLUSH_ZERO_MASK; #else /*JUCE_USE_ARM_NEON*/ intptr_t mask = (1 << 24 /* FZ */); #endif setFpStatusRegister ((getFpStatusRegister() & (~mask)) | (shouldEnable ? mask : 0)); #else #if ! (defined(JUCE_INTEL) || defined(JUCE_ARM)) jassertfalse; // No support for flush to zero mode on your platform #endif ignoreUnused (shouldEnable); #endif } void JUCE_CALLTYPE FloatVectorOperations::disableDenormalisedNumberSupport (bool shouldDisable) noexcept { #if JUCE_USE_SSE_INTRINSICS || (JUCE_USE_ARM_NEON || defined(__arm64__) || defined(__aarch64__)) #if JUCE_USE_SSE_INTRINSICS intptr_t mask = 0x8040; #else /*JUCE_USE_ARM_NEON*/ intptr_t mask = (1 << 24 /* FZ */); #endif setFpStatusRegister ((getFpStatusRegister() & (~mask)) | (shouldDisable ? mask : 0)); #else ignoreUnused (shouldDisable); #if ! (defined(JUCE_INTEL) || defined(JUCE_ARM)) jassertfalse; // No support for disable denormals mode on your platform #endif #endif } bool JUCE_CALLTYPE FloatVectorOperations::areDenormalsDisabled() noexcept { #if JUCE_USE_SSE_INTRINSICS || (JUCE_USE_ARM_NEON || defined(__arm64__) || defined(__aarch64__)) #if JUCE_USE_SSE_INTRINSICS intptr_t mask = 0x8040; #else /*JUCE_USE_ARM_NEON*/ intptr_t mask = (1 << 24 /* FZ */); #endif return ((getFpStatusRegister() & mask) == mask); #else return false; #endif } ScopedNoDenormals::ScopedNoDenormals() noexcept { #if JUCE_USE_SSE_INTRINSICS || (JUCE_USE_ARM_NEON || defined(__arm64__) || defined(__aarch64__)) #if JUCE_USE_SSE_INTRINSICS intptr_t mask = 0x8040; #else /*JUCE_USE_ARM_NEON*/ intptr_t mask = (1 << 24 /* FZ */); #endif fpsr = FloatVectorOperations::getFpStatusRegister(); FloatVectorOperations::setFpStatusRegister (fpsr | mask); #endif } ScopedNoDenormals::~ScopedNoDenormals() noexcept { #if JUCE_USE_SSE_INTRINSICS || (JUCE_USE_ARM_NEON || defined(__arm64__) || defined(__aarch64__)) FloatVectorOperations::setFpStatusRegister (fpsr); #endif } //============================================================================== //============================================================================== #if JUCE_UNIT_TESTS class FloatVectorOperationsTests : public UnitTest { public: FloatVectorOperationsTests() : UnitTest ("FloatVectorOperations", UnitTestCategories::audio) {} template struct TestRunner { static void runTest (UnitTest& u, Random random) { const int range = random.nextBool() ? 500 : 10; const int num = random.nextInt (range) + 1; HeapBlock buffer1 (num + 16), buffer2 (num + 16); HeapBlock buffer3 (num + 16, true); #if JUCE_ARM ValueType* const data1 = buffer1; ValueType* const data2 = buffer2; int* const int1 = buffer3; #else // These tests deliberately operate on misaligned memory and will be flagged up by // checks for undefined behavior! ValueType* const data1 = addBytesToPointer (buffer1.get(), random.nextInt (16)); ValueType* const data2 = addBytesToPointer (buffer2.get(), random.nextInt (16)); int* const int1 = addBytesToPointer (buffer3.get(), random.nextInt (16)); #endif fillRandomly (random, data1, num); fillRandomly (random, data2, num); Range minMax1 (FloatVectorOperations::findMinAndMax (data1, num)); Range minMax2 (Range::findMinAndMax (data1, num)); u.expect (minMax1 == minMax2); u.expect (valuesMatch (FloatVectorOperations::findMinimum (data1, num), juce::findMinimum (data1, num))); u.expect (valuesMatch (FloatVectorOperations::findMaximum (data1, num), juce::findMaximum (data1, num))); u.expect (valuesMatch (FloatVectorOperations::findMinimum (data2, num), juce::findMinimum (data2, num))); u.expect (valuesMatch (FloatVectorOperations::findMaximum (data2, num), juce::findMaximum (data2, num))); FloatVectorOperations::clear (data1, num); u.expect (areAllValuesEqual (data1, num, 0)); FloatVectorOperations::fill (data1, (ValueType) 2, num); u.expect (areAllValuesEqual (data1, num, (ValueType) 2)); FloatVectorOperations::add (data1, (ValueType) 2, num); u.expect (areAllValuesEqual (data1, num, (ValueType) 4)); FloatVectorOperations::copy (data2, data1, num); u.expect (areAllValuesEqual (data2, num, (ValueType) 4)); FloatVectorOperations::add (data2, data1, num); u.expect (areAllValuesEqual (data2, num, (ValueType) 8)); FloatVectorOperations::copyWithMultiply (data2, data1, (ValueType) 4, num); u.expect (areAllValuesEqual (data2, num, (ValueType) 16)); FloatVectorOperations::addWithMultiply (data2, data1, (ValueType) 4, num); u.expect (areAllValuesEqual (data2, num, (ValueType) 32)); FloatVectorOperations::multiply (data1, (ValueType) 2, num); u.expect (areAllValuesEqual (data1, num, (ValueType) 8)); FloatVectorOperations::multiply (data1, data2, num); u.expect (areAllValuesEqual (data1, num, (ValueType) 256)); FloatVectorOperations::negate (data2, data1, num); u.expect (areAllValuesEqual (data2, num, (ValueType) -256)); FloatVectorOperations::subtract (data1, data2, num); u.expect (areAllValuesEqual (data1, num, (ValueType) 512)); FloatVectorOperations::abs (data1, data2, num); u.expect (areAllValuesEqual (data1, num, (ValueType) 256)); FloatVectorOperations::abs (data2, data1, num); u.expect (areAllValuesEqual (data2, num, (ValueType) 256)); fillRandomly (random, int1, num); doConversionTest (u, data1, data2, int1, num); FloatVectorOperations::fill (data1, (ValueType) 2, num); FloatVectorOperations::fill (data2, (ValueType) 3, num); FloatVectorOperations::addWithMultiply (data1, data1, data2, num); u.expect (areAllValuesEqual (data1, num, (ValueType) 8)); } static void doConversionTest (UnitTest& u, float* data1, float* data2, int* const int1, int num) { FloatVectorOperations::convertFixedToFloat (data1, int1, 2.0f, num); convertFixed (data2, int1, 2.0f, num); u.expect (buffersMatch (data1, data2, num)); } static void doConversionTest (UnitTest&, double*, double*, int*, int) {} static void fillRandomly (Random& random, ValueType* d, int num) { while (--num >= 0) *d++ = (ValueType) (random.nextDouble() * 1000.0); } static void fillRandomly (Random& random, int* d, int num) { while (--num >= 0) *d++ = random.nextInt(); } static void convertFixed (float* d, const int* s, ValueType multiplier, int num) { while (--num >= 0) *d++ = (float) *s++ * multiplier; } static bool areAllValuesEqual (const ValueType* d, int num, ValueType target) { while (--num >= 0) if (*d++ != target) return false; return true; } static bool buffersMatch (const ValueType* d1, const ValueType* d2, int num) { while (--num >= 0) if (! valuesMatch (*d1++, *d2++)) return false; return true; } static bool valuesMatch (ValueType v1, ValueType v2) { return std::abs (v1 - v2) < std::numeric_limits::epsilon(); } }; void runTest() override { beginTest ("FloatVectorOperations"); for (int i = 1000; --i >= 0;) { TestRunner::runTest (*this, getRandom()); TestRunner::runTest (*this, getRandom()); } } }; static FloatVectorOperationsTests vectorOpTests; #endif } // namespace juce