diff options
Diffstat (limited to 'src/core')
-rw-r--r-- | src/core/NEON/kernels/NESoftmaxLayerKernel.cpp | 562 |
1 files changed, 129 insertions, 433 deletions
diff --git a/src/core/NEON/kernels/NESoftmaxLayerKernel.cpp b/src/core/NEON/kernels/NESoftmaxLayerKernel.cpp index 1003ebd2e3..a3ecce3a1e 100644 --- a/src/core/NEON/kernels/NESoftmaxLayerKernel.cpp +++ b/src/core/NEON/kernels/NESoftmaxLayerKernel.cpp @@ -30,6 +30,7 @@ #include "arm_compute/core/ITensor.h" #include "arm_compute/core/NEON/NEFixedPoint.h" #include "arm_compute/core/NEON/NEMath.h" +#include "arm_compute/core/NEON/wrapper/wrapper.h" #include "arm_compute/core/TensorInfo.h" #include "arm_compute/core/Utils.h" #include "arm_compute/core/Validate.h" @@ -43,309 +44,6 @@ namespace arm_compute { -template <typename T, int N> -struct vec_n_type; - -#define DECLARE_NEON_VEC_TYPE(T, N, V) \ - template <> \ - struct vec_n_type<T, N> \ - { \ - using type = V; \ - }; - -DECLARE_NEON_VEC_TYPE(uint8_t, 16, uint8x16_t) -DECLARE_NEON_VEC_TYPE(uint8_t, 8, uint8x8_t) - -DECLARE_NEON_VEC_TYPE(int8_t, 16, int8x16_t) -DECLARE_NEON_VEC_TYPE(int8_t, 8, int8x8_t) - -DECLARE_NEON_VEC_TYPE(uint16_t, 8, uint16x8_t) -DECLARE_NEON_VEC_TYPE(uint16_t, 4, uint16x4_t) - -DECLARE_NEON_VEC_TYPE(int16_t, 8, int16x8_t) -DECLARE_NEON_VEC_TYPE(int16_t, 4, int16x4_t) - -DECLARE_NEON_VEC_TYPE(int32_t, 4, int32x4_t) -DECLARE_NEON_VEC_TYPE(int32_t, 2, int32x2_t) - -DECLARE_NEON_VEC_TYPE(uint32_t, 4, uint32x4_t) -DECLARE_NEON_VEC_TYPE(uint32_t, 2, uint32x2_t) - -#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC -DECLARE_NEON_VEC_TYPE(float16_t, 8, float16x8_t) -DECLARE_NEON_VEC_TYPE(float16_t, 4, float16x4_t) -#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ - -DECLARE_NEON_VEC_TYPE(float, 4, float32x4_t) -DECLARE_NEON_VEC_TYPE(float, 2, float32x2_t) - -template <typename T, int N> -using vec_n_t = typename vec_n_type<T, N>::type; - -template <typename T, int N> -using vec_n_byte_t = vec_n_t < T, N / sizeof(T) >; - -template <typename T> -using vec_16_byte_t = vec_n_byte_t<T, 16>; - -template <typename T> -using vec_8_byte_t = vec_n_byte_t<T, 8>; - -template <typename T> -using const_ptr_t = const T *; - -template <typename T> -using ptr_t = T *; - -#define FORWARD_DECLARE_VGET_LANE_FOR_TYPE(TYPE) \ - template <int lane> \ - TYPE vget_lane(vec_8_byte_t<TYPE> vec); \ - template <int lane> \ - TYPE vget_lane(vec_16_byte_t<TYPE> vec); - -FORWARD_DECLARE_VGET_LANE_FOR_TYPE(uint8_t) -FORWARD_DECLARE_VGET_LANE_FOR_TYPE(int8_t) -FORWARD_DECLARE_VGET_LANE_FOR_TYPE(uint16_t) -FORWARD_DECLARE_VGET_LANE_FOR_TYPE(int16_t) -FORWARD_DECLARE_VGET_LANE_FOR_TYPE(uint32_t) -FORWARD_DECLARE_VGET_LANE_FOR_TYPE(int32_t) -#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC -FORWARD_DECLARE_VGET_LANE_FOR_TYPE(float16_t) -#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ -FORWARD_DECLARE_VGET_LANE_FOR_TYPE(float) -template <int lane> -float vget_lane(float32x4x4_t vec); - -template <typename V> -using elem_type_t = decltype(vget_lane<0>(std::declval<V>())); - -template <typename V> -constexpr size_t vec_size_of(const V &vec) -{ - return sizeof(vec) / sizeof(elem_type_t<V>); -} - -template <typename V> -V vdup_n(elem_type_t<V> val); -template <typename V> -V vld(const_ptr_t<elem_type_t<V>> ptr); - -#define DECLARE_NEON_FUNCTIONS_FOR_TYPE(TYPE, TAG) \ - template <> \ - inline vec_8_byte_t<TYPE> vdup_n<vec_8_byte_t<TYPE>>(TYPE val) \ - { \ - return vdup_n_##TAG(val); \ - } \ - template <> \ - inline vec_16_byte_t<TYPE> vdup_n<vec_16_byte_t<TYPE>>(TYPE val) \ - { \ - return vdupq_n_##TAG(val); \ - } \ - template <> \ - inline vec_8_byte_t<TYPE> vld<vec_8_byte_t<TYPE>>(const_ptr_t<TYPE> ptr) \ - { \ - return vld1_##TAG(ptr); \ - } \ - template <> \ - inline vec_16_byte_t<TYPE> vld<vec_16_byte_t<TYPE>>(const_ptr_t<TYPE> ptr) \ - { \ - return vld1q_##TAG(ptr); \ - } \ - inline void vst(ptr_t<TYPE> ptr, vec_8_byte_t<TYPE> vec) \ - { \ - vst1_##TAG(ptr, vec); \ - } \ - inline void vst(ptr_t<TYPE> ptr, vec_16_byte_t<TYPE> vec) \ - { \ - vst1q_##TAG(ptr, vec); \ - } \ - inline vec_16_byte_t<TYPE> vmax(vec_16_byte_t<TYPE> a, vec_16_byte_t<TYPE> b) \ - { \ - return vmaxq_##TAG(a, b); \ - } \ - inline vec_8_byte_t<TYPE> vpmax(vec_8_byte_t<TYPE> a, vec_8_byte_t<TYPE> b) \ - { \ - return vpmax_##TAG(a, b); \ - } \ - inline vec_8_byte_t<TYPE> vget_low(vec_16_byte_t<TYPE> vec) \ - { \ - return vget_low_##TAG(vec); \ - } \ - inline vec_8_byte_t<TYPE> vget_high(vec_16_byte_t<TYPE> vec) \ - { \ - return vget_high_##TAG(vec); \ - } \ - template <int lane> \ - inline TYPE vget_lane(vec_8_byte_t<TYPE> vec) \ - { \ - static_assert(lane >= 0, "lane is out of bounds"); \ - static_assert(lane < vec_size_of(vec), "lane is out of bounds"); \ - return vget_lane_##TAG(vec, lane); \ - } \ - template <int lane> \ - inline TYPE vget_lane(vec_16_byte_t<TYPE> vec) \ - { \ - static_assert(lane >= 0, "lane is out of bounds"); \ - static_assert(lane < vec_size_of(vec), "lane is out of bounds"); \ - return vgetq_lane_##TAG(vec, lane); \ - } - -template <typename T> -T sqadd(T a, T b); -template <typename T> -T sqsub(T a, T b); -template <typename T> -T sqmul(T a, T b); - -#define DECLARE_NEON_FUNCTIONS_FOR_FLOAT(TYPE, TAG) \ - inline vec_8_byte_t<TYPE> vadd(vec_8_byte_t<TYPE> a, vec_8_byte_t<TYPE> b) \ - { \ - return vadd_##TAG(a, b); \ - } \ - inline vec_16_byte_t<TYPE> vadd(vec_16_byte_t<TYPE> a, vec_16_byte_t<TYPE> b) \ - { \ - return vaddq_##TAG(a, b); \ - } \ - inline vec_16_byte_t<TYPE> vsub(vec_16_byte_t<TYPE> a, vec_16_byte_t<TYPE> b) \ - { \ - return vsubq_##TAG(a, b); \ - } \ - inline vec_16_byte_t<TYPE> vmul_n(vec_16_byte_t<TYPE> vec, TYPE val) \ - { \ - return vmulq_n_##TAG(vec, val); \ - } - -DECLARE_NEON_FUNCTIONS_FOR_TYPE(uint8_t, u8) -DECLARE_NEON_FUNCTIONS_FOR_TYPE(int8_t, s8) -DECLARE_NEON_FUNCTIONS_FOR_TYPE(uint16_t, u16) -DECLARE_NEON_FUNCTIONS_FOR_TYPE(int16_t, s16) -DECLARE_NEON_FUNCTIONS_FOR_TYPE(uint32_t, u32) -DECLARE_NEON_FUNCTIONS_FOR_TYPE(int32_t, s32) -#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC -DECLARE_NEON_FUNCTIONS_FOR_TYPE(float16_t, f16) -#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ -DECLARE_NEON_FUNCTIONS_FOR_TYPE(float, f32) - -#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC -DECLARE_NEON_FUNCTIONS_FOR_FLOAT(float16_t, f16) -#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ -DECLARE_NEON_FUNCTIONS_FOR_FLOAT(float, f32) - -template <typename VO, typename VI> -VO vcvt(VI vec); - -template <> -float32x4x4_t vcvt<float32x4x4_t>(uint8x16_t vec) -{ - const auto low = vmovl_u8(vget_low(vec)); - const auto high = vmovl_u8(vget_high(vec)); - float32x4x4_t res = { { - vcvtq_f32_u32(vmovl_u16(vget_low(low))), - vcvtq_f32_u32(vmovl_u16(vget_high(low))), - vcvtq_f32_u32(vmovl_u16(vget_low(high))), - vcvtq_f32_u32(vmovl_u16(vget_high(high))) - } - }; - return res; -} - -template <> -uint8x16_t vcvt<uint8x16_t>(float32x4x4_t vec) -{ - uint16x8x2_t resU16 = { { - vcombine_u16(vqmovn_u32(vcvtq_u32_f32(vec.val[0])), - vqmovn_u32(vcvtq_u32_f32(vec.val[1]))), - vcombine_u16(vqmovn_u32(vcvtq_u32_f32(vec.val[2])), - vqmovn_u32(vcvtq_u32_f32(vec.val[3]))) - } - }; - - uint8x16_t res = vcombine_u8(vqmovn_u16(resU16.val[0]), vqmovn_u16(resU16.val[1])); - return res; -} - -float32x4x4_t vexp(float32x4x4_t vec) -{ - float32x4x4_t res = { { - vexpq_f32(vec.val[0]), - vexpq_f32(vec.val[1]), - vexpq_f32(vec.val[2]), - vexpq_f32(vec.val[3]) - } - }; - return res; -} - -float32x4_t vexp(const float32x4_t &vec) -{ - return vexpq_f32(vec); -} - -#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC -// TODO (COMPMID-1535) : Revisit FP16 approximations -float16x8_t vexp(const float16x8_t &vec) -{ - float16x4x2_t res = - { - { - vcvt_f16_f32(vexpq_f32(vcvt_f32_f16(vget_low_f16(vec)))), - vcvt_f16_f32(vexpq_f32(vcvt_f32_f16(vget_high_f16(vec)))) - } - }; - return vcombine_f16(res.val[0], res.val[1]); -} -#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ - -template <> -float32x4x4_t vdup_n<float32x4x4_t>(float val) -{ - float32x4x4_t res = { { - vdupq_n_f32(val), - vdupq_n_f32(val), - vdupq_n_f32(val), - vdupq_n_f32(val) - } - }; - return res; -} - -float32x4x4_t vmul_n(float32x4x4_t vec, float val) -{ - float32x4x4_t res = { { - vmulq_n_f32(vec.val[0], val), - vmulq_n_f32(vec.val[1], val), - vmulq_n_f32(vec.val[2], val), - vmulq_n_f32(vec.val[3], val) - } - }; - return res; -} - -float32x4x4_t vadd(float32x4x4_t a, float32x4x4_t b) -{ - float32x4x4_t res = { { - vaddq_f32(a.val[0], b.val[0]), - vaddq_f32(a.val[1], b.val[1]), - vaddq_f32(a.val[2], b.val[2]), - vaddq_f32(a.val[3], b.val[3]) - } - }; - return res; -} - -float32x4x4_t vsub_n(float32x4x4_t a, float val) -{ - auto scalar_vector = vdup_n<float32x4x4_t>(val); - float32x4x4_t res = { { - vsubq_f32(a.val[0], scalar_vector.val[0]), - vsubq_f32(a.val[1], scalar_vector.val[1]), - vsubq_f32(a.val[2], scalar_vector.val[2]), - vsubq_f32(a.val[3], scalar_vector.val[3]) - } - }; - return res; -} - namespace { Status validate_arguments_logits_1d_max(const ITensorInfo &input, const ITensorInfo &output) @@ -390,30 +88,20 @@ std::pair<Status, Window> validate_and_configure_window_logits_1d_max(ITensorInf return std::make_pair(err, win); } -template <typename V> -auto reduce_max(V vec) -> elem_type_t<V> -{ - constexpr int N = vec_size_of(vec); - - auto carry_max = vpmax(vget_high(vec), vget_low(vec)); - - for(int k = N / 2; k > 1; k /= 2) - { - carry_max = vpmax(carry_max, carry_max); - } - - return vget_lane<0>(carry_max); -} - template <typename T> void logits_1d_max(const ITensor &in, ITensor &out, const Window &window) { const auto start_x = in.info()->valid_region().anchor.x(); const size_t input_width = in.info()->valid_region().shape.x(); + /** NEON vector tag type. */ + using ExactTagType = typename wrapper::traits::neon_bitvector_tag_t<T, wrapper::traits::BitWidth::W128>; + Iterator input(&in, window); Iterator output(&out, window); + constexpr int window_step_x = 16 / sizeof(T); + const int sum_stages = log2(window_step_x / 2); execute_window_loop(window, [&](const Coordinates &) { // Get pointers @@ -421,16 +109,22 @@ void logits_1d_max(const ITensor &in, ITensor &out, const Window &window) const auto out_ptr = reinterpret_cast<T *>(output.ptr()); // Init max value - auto vec_max = vdup_n<vec_16_byte_t<T>>(support::cpp11::lowest<T>()); + auto vec_max = wrapper::vdup_n(support::cpp11::lowest<T>(), ExactTagType{}); // Loop over input row - for(const T *it = in_ptr; it < (in_ptr + input_width); it += vec_size_of(vec_max)) + for(const T *it = in_ptr; it < (in_ptr + input_width); it += window_step_x) { - const auto current_value = vld<vec_16_byte_t<T>>(it); - vec_max = vmax(vec_max, current_value); + const auto current_value = wrapper::vloadq(it); + vec_max = wrapper::vmax(vec_max, current_value); } - const T max_val = reduce_max(vec_max); + auto carry_max = wrapper::vpmax(wrapper::vgethigh(vec_max), wrapper::vgetlow(vec_max)); + + for(int i = 0; i < sum_stages; ++i) + { + carry_max = wrapper::vpmax(carry_max, carry_max); + } + const T max_val = wrapper::vgetlane(carry_max, 0); *out_ptr = max_val; }, input, output); @@ -575,45 +269,19 @@ std::pair<Status, Window> validate_and_configure_window_logits_softmax(ITensorIn return std::make_pair(err, win); } -template <typename T, int N, int S, int E> -struct reduce_add_impl -{ - template <typename F> - static T reduce(F add_fn, vec_n_t<T, N> vec) - { - constexpr int H = (S + E + 1) / 2; - const auto reduced_high = reduce_add_impl < T, N, S, H - 1 >::reduce(add_fn, vec); - const auto reduced_low = reduce_add_impl<T, N, H, E>::reduce(add_fn, vec); - return add_fn(reduced_high, reduced_low); - } -}; -template <typename T, int N, int I> -struct reduce_add_impl<T, N, I, I> -{ - template <typename F> - static T reduce(F /*add_fn*/, vec_n_t<T, N> vec) - { - return vget_lane<I>(vec); - } -}; -template <typename V, typename F> -elem_type_t<V> reduce_add(F add_fn, V vec) -{ - constexpr int N = vec_size_of(vec); - return reduce_add_impl < elem_type_t<V>, N, 0, N - 1 >::reduce(add_fn, vec); -} - template <bool is_log> void logits_1d_softmax_qasymm8(const ITensor &in, const ITensor &max, void *const tmp, ITensor &out, const float beta, const Window &window) { const int start_x = in.info()->valid_region().anchor.x(); const int input_width = in.info()->valid_region().shape.x(); - const float scale_beta = -beta * in.info()->quantization_info().uniform().scale; + const float scale_beta = -beta * in.info()->quantization_info().uniform().scale; + const auto scale_beta_vec = vdupq_n_f32(scale_beta); - Iterator in_it(&in, window); - Iterator max_it(&max, window); - Iterator out_it(&out, window); + Iterator in_it(&in, window); + Iterator max_it(&max, window); + Iterator out_it(&out, window); + constexpr int vec_size = 16; execute_window_loop(window, [&](const Coordinates &) { @@ -629,57 +297,73 @@ void logits_1d_softmax_qasymm8(const ITensor &in, const ITensor &max, void *cons { /* Get max value */ const auto max_val = *reinterpret_cast<const qasymm8_t *>(max_it.ptr()); - const auto vec_max = vdup_n<vec_16_byte_t<qasymm8_t>>(max_val); + const auto vec_max = vdupq_n_u8(max_val); /* Init sum to zero */ - auto vec_sum = vdup_n<float32x4x4_t>(0.f); + float32x4x4_t vec_sum = + { + vdupq_n_f32(0.f), + vdupq_n_f32(0.f), + vdupq_n_f32(0.f), + vdupq_n_f32(0.f), + }; /* Loop over row and compute exponentials and sum */ - int i = 0; - constexpr int vec_size = vec_size_of(vec_max); - - for(; i <= (input_width - vec_size); i += vec_size) + int x = 0; + for(; x <= (input_width - vec_size); x += vec_size) { - auto vec_elements = vld<vec_16_byte_t<qasymm8_t>>(in_ptr + i); - vec_elements = vsubq_u8(vec_max, vec_elements); - - auto vec_elements_flt = vcvt<float32x4x4_t>(vec_elements); + auto vec_elements = wrapper::vloadq(in_ptr + x); + vec_elements = vsubq_u8(vec_max, vec_elements); + auto vec_elements_flt = convert_uint8x16_to_float32x4x4(vec_elements); if(is_log) { - vec_elements_flt = vmul_n(vec_elements_flt, scale_beta); - vec_sum = vadd(vec_sum, vexp(vec_elements_flt)); + vec_elements_flt.val[0] = vmulq_f32(vec_elements_flt.val[0], scale_beta_vec); + vec_elements_flt.val[1] = vmulq_f32(vec_elements_flt.val[1], scale_beta_vec); + vec_elements_flt.val[2] = vmulq_f32(vec_elements_flt.val[2], scale_beta_vec); + vec_elements_flt.val[3] = vmulq_f32(vec_elements_flt.val[3], scale_beta_vec); + vec_sum.val[0] = vaddq_f32(vec_sum.val[0], vexpq_f32(vec_elements_flt.val[0])); + vec_sum.val[1] = vaddq_f32(vec_sum.val[1], vexpq_f32(vec_elements_flt.val[1])); + vec_sum.val[2] = vaddq_f32(vec_sum.val[2], vexpq_f32(vec_elements_flt.val[2])); + vec_sum.val[3] = vaddq_f32(vec_sum.val[3], vexpq_f32(vec_elements_flt.val[3])); } else { - vec_elements_flt = vexp(vmul_n(vec_elements_flt, scale_beta)); - vec_sum = vadd(vec_sum, vec_elements_flt); + vec_elements_flt.val[0] = vexpq_f32(vmulq_f32(vec_elements_flt.val[0], scale_beta_vec)); + vec_elements_flt.val[1] = vexpq_f32(vmulq_f32(vec_elements_flt.val[1], scale_beta_vec)); + vec_elements_flt.val[2] = vexpq_f32(vmulq_f32(vec_elements_flt.val[2], scale_beta_vec)); + vec_elements_flt.val[3] = vexpq_f32(vmulq_f32(vec_elements_flt.val[3], scale_beta_vec)); + vec_sum.val[0] = vaddq_f32(vec_sum.val[0], vec_elements_flt.val[0]); + vec_sum.val[1] = vaddq_f32(vec_sum.val[1], vec_elements_flt.val[1]); + vec_sum.val[2] = vaddq_f32(vec_sum.val[2], vec_elements_flt.val[2]); + vec_sum.val[3] = vaddq_f32(vec_sum.val[3], vec_elements_flt.val[3]); } - vst4q_f32(tmp_ptr + i, vec_elements_flt); + + vst4q_f32(tmp_ptr + x, vec_elements_flt); } /* Reduce sum */ - const auto sum_16_byte = vaddq_f32(vaddq_f32(vec_sum.val[0], vec_sum.val[1]), - vaddq_f32(vec_sum.val[2], vec_sum.val[3])); - const auto sum_8_byte = vadd_f32(vget_low(sum_16_byte), vget_high(sum_16_byte)); - sum = reduce_add(std::plus<float>(), sum_8_byte); + const auto sum_16_byte = vaddq_f32(vaddq_f32(vec_sum.val[0], vec_sum.val[1]), vaddq_f32(vec_sum.val[2], vec_sum.val[3])); + auto sum_res = vpadd_f32(vget_high_f32(sum_16_byte), vget_low_f32(sum_16_byte)); + sum_res = vpadd_f32(sum_res, sum_res); + sum = wrapper::vgetlane(sum_res, 0); /* Run remaining elements */ - for(; i < input_width; ++i) + for(; x < input_width; ++x) { float element{}; if(is_log) { - element = (max_val - in_ptr[i]) * scale_beta; + element = (max_val - in_ptr[x]) * scale_beta; sum += std::exp(element); } else { - element = std::exp((max_val - in_ptr[i]) * scale_beta); + element = std::exp((max_val - in_ptr[x]) * scale_beta); sum += element; } - tmp_ptr[i] = element; + tmp_ptr[x] = element; } if(!is_log) @@ -691,35 +375,45 @@ void logits_1d_softmax_qasymm8(const ITensor &in, const ITensor &max, void *cons /* Normalize exponentials */ { /* Loop over row and compute softmax */ - int i = 0; + int x = 0; + for(; x <= (input_width - vec_size); x += vec_size) { - constexpr int vec_size = 16; - - for(; i <= (input_width - vec_size); i += vec_size) + float32x4x4_t vec_in = vld4q_f32(tmp_ptr + x); + uint8x16_t normalized_value{}; + if(is_log) { - float32x4x4_t vec_in = vld4q_f32(tmp_ptr + i); - vec_16_byte_t<qasymm8_t> normalized_value{}; - if(is_log) + const float32x4x4_t sub = { - normalized_value = vcvt<vec_16_byte_t<qasymm8_t>>(vsub_n(vec_in, sum)); - } - else + vsubq_f32(vec_in.val[0], vdupq_n_f32(sum)), + vsubq_f32(vec_in.val[1], vdupq_n_f32(sum)), + vsubq_f32(vec_in.val[2], vdupq_n_f32(sum)), + vsubq_f32(vec_in.val[3], vdupq_n_f32(sum)), + }; + convert_float32x4x4_to_unit8x16(sub, normalized_value); + } + else + { + const float32x4x4_t mul = { - normalized_value = vcvt<vec_16_byte_t<qasymm8_t>>(vmul_n(vec_in, sum_inversed)); - } - vst(out_ptr + i, normalized_value); + vmulq_f32(vec_in.val[0], vdupq_n_f32(sum_inversed)), + vmulq_f32(vec_in.val[1], vdupq_n_f32(sum_inversed)), + vmulq_f32(vec_in.val[2], vdupq_n_f32(sum_inversed)), + vmulq_f32(vec_in.val[3], vdupq_n_f32(sum_inversed)), + }; + convert_float32x4x4_to_unit8x16(mul, normalized_value); } + vst1q_u8(out_ptr + x, normalized_value); } /* Run remaining elements */ - for(; i < input_width; ++i) + for(; x < input_width; ++x) { if(is_log) { - out_ptr[i] = utils::cast::saturate_cast<qasymm8_t>(tmp_ptr[i] - sum); + out_ptr[x] = utils::cast::saturate_cast<qasymm8_t>(tmp_ptr[x] - sum); } else { - out_ptr[i] = utils::cast::saturate_cast<qasymm8_t>(tmp_ptr[i] * sum_inversed); + out_ptr[x] = utils::cast::saturate_cast<qasymm8_t>(tmp_ptr[x] * sum_inversed); } } } @@ -738,6 +432,12 @@ void logits_1d_softmax_float(const ITensor &in, const ITensor &max, void *const Iterator max_it(&max, window); Iterator out_it(&out, window); + /** NEON vector tag type. */ + using ExactTagType = typename wrapper::traits::neon_bitvector_tag_t<T, wrapper::traits::BitWidth::W128>; + + constexpr int vec_size = 16 / sizeof(T); + const int sum_stages = log2(vec_size / 2); + execute_window_loop(window, [&](const Coordinates &) { /* Get pointers */ @@ -752,53 +452,54 @@ void logits_1d_softmax_float(const ITensor &in, const ITensor &max, void *const { /* Get max value */ const auto max_val = *reinterpret_cast<const T *>(max_it.ptr()); - const auto vec_max = vdup_n<vec_16_byte_t<T>>(max_val); + const auto vec_max = wrapper::vdup_n(max_val, ExactTagType{}); /* Init sum to zero */ - auto vec_sum = vdup_n<vec_16_byte_t<T>>(0); + auto vec_sum = wrapper::vdup_n(static_cast<T>(0), ExactTagType{}); /* Loop over row and compute exponentials and sum */ - int i = 0; - constexpr int vec_size = vec_size_of(vec_sum); - - for(; i <= (input_width - vec_size); i += vec_size) + int x = 0; + for(; x <= (input_width - vec_size); x += vec_size) { - auto vec_elements = vld<vec_16_byte_t<T>>(in_ptr + i); - vec_elements = vsub(vec_elements, vec_max); + auto vec_elements = wrapper::vloadq(in_ptr + x); + vec_elements = wrapper::vsub(vec_elements, vec_max); if(is_log) { - vec_elements = vmul_n(vec_elements, static_cast<T>(beta)); - vec_sum = vadd(vec_sum, vexp(vec_elements)); + vec_elements = wrapper::vmul(vec_elements, wrapper::vdup_n(static_cast<T>(beta), ExactTagType{})); + vec_sum = wrapper::vadd(vec_sum, wrapper::vexpq(vec_elements)); } else { - vec_elements = vexp(vmul_n(vec_elements, static_cast<T>(beta))); - vec_sum = vadd(vec_sum, vec_elements); + vec_elements = wrapper::vexpq(wrapper::vmul(vec_elements, wrapper::vdup_n(static_cast<T>(beta), ExactTagType{}))); + vec_sum = wrapper::vadd(vec_sum, vec_elements); } - vst(tmp_ptr + i, vec_elements); + wrapper::vstore(tmp_ptr + x, vec_elements); } /* Reduce sum */ - const auto sum_8_byte = vadd(vget_high(vec_sum), vget_low(vec_sum)); - sum = reduce_add([](T a, T b) -> T { return a + b; }, sum_8_byte); + auto sum_res = wrapper::vpadd(wrapper::vgethigh(vec_sum), wrapper::vgetlow(vec_sum)); + for(int i = 0; i < sum_stages; ++i) + { + sum_res = wrapper::vpadd(sum_res, sum_res); + } + sum = wrapper::vgetlane(sum_res, 0); /* Run remaining elements */ - - for(; i < input_width; ++i) + for(; x < input_width; ++x) { T element{}; if(is_log) { - element = (in_ptr[i] - max_val) * beta; + element = (in_ptr[x] - max_val) * beta; sum += std::exp(element); } else { - element = std::exp((in_ptr[i] - max_val) * beta); + element = std::exp((in_ptr[x] - max_val) * beta); sum += element; } - tmp_ptr[i] = element; + tmp_ptr[x] = element; } if(!is_log) @@ -810,36 +511,31 @@ void logits_1d_softmax_float(const ITensor &in, const ITensor &max, void *const /* Normalize exponentials */ { /* Loop over row and compute softmax */ - int i = 0; - + int x = 0; + for(; x <= (input_width - vec_size); x += vec_size) { - constexpr int vec_size = vec_size_of(vec_16_byte_t<T> {}); - - for(; i <= (input_width - vec_size); i += vec_size) + auto vec_in = wrapper::vloadq(tmp_ptr + x); + auto normalized_value = wrapper::vdup_n(static_cast<T>(0), ExactTagType{}); + if(is_log) { - auto vec_in = vld<vec_16_byte_t<T>>(tmp_ptr + i); - vec_16_byte_t<T> normalized_value{}; - if(is_log) - { - normalized_value = vsub(vec_in, vdup_n<vec_16_byte_t<T>>(sum)); - } - else - { - normalized_value = vmul_n(vec_in, sum_inversed); - } - vst(out_ptr + i, normalized_value); + normalized_value = wrapper::vsub(vec_in, wrapper::vdup_n(static_cast<T>(sum), ExactTagType{})); + } + else + { + normalized_value = wrapper::vmul(vec_in, wrapper::vdup_n(static_cast<T>(sum_inversed), ExactTagType{})); } + wrapper::vstore(out_ptr + x, normalized_value); } /* Run remaining elements */ - for(; i < input_width; ++i) + for(; x < input_width; ++x) { if(is_log) { - out_ptr[i] = tmp_ptr[i] - sum; + out_ptr[x] = tmp_ptr[x] - sum; } else { - out_ptr[i] = tmp_ptr[i] * sum_inversed; + out_ptr[x] = tmp_ptr[x] * sum_inversed; } } } |