diff options
Diffstat (limited to 'src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp')
| -rw-r--r-- | src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp | 1382 |
1 files changed, 854 insertions, 528 deletions
diff --git a/src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp b/src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp index ca59e66293..b23a20d019 100644 --- a/src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp +++ b/src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp @@ -43,8 +43,6 @@ const float scale255_constant = 1.f / 255.f; const float32x4_t scale255_constant_f32q = vdupq_n_f32(scale255_constant); const float32x4_t positive_round_f32q = vdupq_n_f32(0.5f); -constexpr unsigned int num_elems_processed_per_iteration = 16; - inline Status validate_arguments(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy) { ARM_COMPUTE_UNUSED(overflow_policy); @@ -100,60 +98,6 @@ inline Status validate_arguments(const ITensorInfo *input1, const ITensorInfo *i return Status{}; } -inline std::pair<Status, Window> validate_and_configure_window(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output) -{ - const std::pair<TensorShape, ValidRegion> broadcast_pair = ITensorInfo::broadcast_shape_and_valid_region(*input1, *input2); - const ValidRegion &valid_region = broadcast_pair.second; - - // Auto initialize output if not initialized - { - ARM_COMPUTE_UNUSED(set_shape_if_empty(*output, input1->tensor_shape())); - - if(input1->data_type() == DataType::S16 || input2->data_type() == DataType::S16) - { - set_format_if_unknown(*output, Format::S16); - } - else if(input1->data_type() == DataType::F32 || input2->data_type() == DataType::F32) - { - set_format_if_unknown(*output, Format::F32); - } - else if(input1->data_type() == DataType::F16 || input2->data_type() == DataType::F16) - { - set_format_if_unknown(*output, Format::F16); - } - else if(input1->data_type() == DataType::QASYMM8 || input2->data_type() == DataType::QASYMM8) - { - set_data_type_if_unknown(*output, DataType::QASYMM8); - } - else if(input1->data_type() == DataType::QASYMM8_SIGNED || input2->data_type() == DataType::QASYMM8_SIGNED) - { - set_data_type_if_unknown(*output, DataType::QASYMM8_SIGNED); - } - else if(input1->data_type() == DataType::QSYMM16 || input2->data_type() == DataType::QSYMM16) - { - set_data_type_if_unknown(*output, DataType::QSYMM16); - } - } - - // Configure kernel window - Window win = calculate_max_window(valid_region, Steps(num_elems_processed_per_iteration)); - Window win_input1 = win.broadcast_if_dimension_le_one(*input1); - Window win_input2 = win.broadcast_if_dimension_le_one(*input2); - - AccessWindowHorizontal input1_access(input1, 0, num_elems_processed_per_iteration); - AccessWindowHorizontal input2_access(input2, 0, num_elems_processed_per_iteration); - AccessWindowHorizontal output_access(output, 0, num_elems_processed_per_iteration); - - bool window_changed = update_window_and_padding(win_input1, input1_access) - || update_window_and_padding(win_input2, input2_access) - || update_window_and_padding(win, output_access); - - output_access.set_valid_region(win, valid_region); - - Status err = (window_changed) ? ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Insufficient Padding!") : Status{}; - return std::make_pair(err, win); -} - /* Scales a given vector by 1/255. * * @note This does not work for all cases. e.g. for float of 0.49999999999999994 and large floats. @@ -178,224 +122,390 @@ inline uint16x8_t scale255_U16_U16(uint16x8_t in) return vreinterpretq_u16_s16(vcombine_s16(vmovn_s32(tmp_s2), vmovn_s32(tmp_s1))); } -inline void mul_saturate_QASYMM8_QASYMM8_QASYMM8_n_opt(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, - float32x4_t input1_vscale, int32x4_t input1_voffset, float32x4_t input2_vscale, int32x4_t input2_voffset, float32x4_t output_voffset, float32x4_t vinvscale) +template <typename T> +inline typename std::enable_if<std::is_same<T, int8_t>::value, int8x16_t>::type +vquantize(float32x4x4_t val, const UniformQuantizationInfo &info) { - const auto input1 = static_cast<const qasymm8_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const qasymm8_t *__restrict>(input2_ptr); - const auto output = static_cast<qasymm8_t *__restrict>(output_ptr); - - const qasymm8x16_t input1_q = vld1q_u8(input1); - const qasymm8x16_t input2_q = vld1q_u8(input2); - - // Dequantitize inputs - float32x4x4_t in1_f32x4x4; - float32x4x4_t in2_f32x4x4; - in1_f32x4x4.val[0] = vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(vmovl_u8(vget_low_u8(input1_q))))), input1_voffset)), input1_vscale); - in1_f32x4x4.val[1] = vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(vmovl_u8(vget_low_u8(input1_q))))), input1_voffset)), input1_vscale); - in1_f32x4x4.val[2] = vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(vmovl_u8(vget_high_u8(input1_q))))), input1_voffset)), input1_vscale); - in1_f32x4x4.val[3] = vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(vmovl_u8(vget_high_u8(input1_q))))), input1_voffset)), input1_vscale); - - in2_f32x4x4.val[0] = vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(vmovl_u8(vget_low_u8(input2_q))))), input2_voffset)), input2_vscale); - in2_f32x4x4.val[1] = vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(vmovl_u8(vget_low_u8(input2_q))))), input2_voffset)), input2_vscale); - in2_f32x4x4.val[2] = vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(vmovl_u8(vget_high_u8(input2_q))))), input2_voffset)), input2_vscale); - in2_f32x4x4.val[3] = vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(vmovl_u8(vget_high_u8(input2_q))))), input2_voffset)), input2_vscale); - - float32x4x4_t out_f32x4x4; - out_f32x4x4.val[0] = vmulq_f32(in1_f32x4x4.val[0], in2_f32x4x4.val[0]); - out_f32x4x4.val[1] = vmulq_f32(in1_f32x4x4.val[1], in2_f32x4x4.val[1]); - out_f32x4x4.val[2] = vmulq_f32(in1_f32x4x4.val[2], in2_f32x4x4.val[2]); - out_f32x4x4.val[3] = vmulq_f32(in1_f32x4x4.val[3], in2_f32x4x4.val[3]); - - int32x4x4_t rf; -#ifdef __aarch64__ - rf.val[0] = vcvtnq_s32_f32(vmlaq_f32(output_voffset, out_f32x4x4.val[0], vinvscale)); - rf.val[1] = vcvtnq_s32_f32(vmlaq_f32(output_voffset, out_f32x4x4.val[1], vinvscale)); - rf.val[2] = vcvtnq_s32_f32(vmlaq_f32(output_voffset, out_f32x4x4.val[2], vinvscale)); - rf.val[3] = vcvtnq_s32_f32(vmlaq_f32(output_voffset, out_f32x4x4.val[3], vinvscale)); -#else //__aarch64__ - rf.val[0] = vcvtq_s32_f32(vmlaq_f32(output_voffset, out_f32x4x4.val[0], vinvscale)); - rf.val[1] = vcvtq_s32_f32(vmlaq_f32(output_voffset, out_f32x4x4.val[1], vinvscale)); - rf.val[2] = vcvtq_s32_f32(vmlaq_f32(output_voffset, out_f32x4x4.val[2], vinvscale)); - rf.val[3] = vcvtq_s32_f32(vmlaq_f32(output_voffset, out_f32x4x4.val[3], vinvscale)); -#endif //__aarch64__ - const uint8x8_t pa = vqmovun_s16(vcombine_s16(vqmovn_s32(rf.val[0]), vqmovn_s32(rf.val[1]))); - const uint8x8_t pb = vqmovun_s16(vcombine_s16(vqmovn_s32(rf.val[2]), vqmovn_s32(rf.val[3]))); - - vst1q_u8(output, vcombine_u8(pa, pb)); + return vquantize_signed(val, info); } -inline void mul_saturate_QASYMM8_SIGNED_QASYMM8_SIGNED_QASYMM8_SIGNED_n( - const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, - float scale, const UniformQuantizationInfo &input1_qua_info, const UniformQuantizationInfo &input2_qua_info, - const UniformQuantizationInfo &output_qua_info) +template <typename T> +inline typename std::enable_if<std::is_same<T, uint8_t>::value, uint8x16_t>::type +vquantize(float32x4x4_t val, const UniformQuantizationInfo &info) +{ + return vquantize(val, info); +} +template <typename T> +inline typename std::enable_if<std::is_same<T, int8_t>::value, int8_t>::type +quantize(float val, const UniformQuantizationInfo &info) { - const auto input1 = static_cast<const qasymm8_signed_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const qasymm8_signed_t *__restrict>(input2_ptr); - const auto output = static_cast<qasymm8_signed_t *__restrict>(output_ptr); - const qasymm8x16_signed_t input1_q = vld1q_s8(input1); - const qasymm8x16_signed_t input2_q = vld1q_s8(input2); - // Dequantitize inputs - const float32x4x4_t in1_f32x4x4 = vdequantize(input1_q, input1_qua_info); - const float32x4x4_t in2_f32x4x4 = vdequantize(input2_q, input2_qua_info); - const UniformQuantizationInfo tmp_qua_info = { output_qua_info.scale / scale, output_qua_info.offset }; - const float32x4x4_t out_f32x4x4 = - { - vmulq_f32(in1_f32x4x4.val[0], in2_f32x4x4.val[0]), - vmulq_f32(in1_f32x4x4.val[1], in2_f32x4x4.val[1]), - vmulq_f32(in1_f32x4x4.val[2], in2_f32x4x4.val[2]), - vmulq_f32(in1_f32x4x4.val[3], in2_f32x4x4.val[3]), - }; - const int8x16_t result = vquantize_signed(out_f32x4x4, tmp_qua_info); - vst1q_s8(output, result); + int32_t tmp = static_cast<int32_t>(val / info.scale) + info.offset; + + T tmp_qua = static_cast<T>(tmp > SCHAR_MAX) ? SCHAR_MAX : ((tmp < SCHAR_MIN) ? SCHAR_MIN : tmp); + return tmp_qua; +} + +template <typename T> +inline typename std::enable_if<std::is_same<T, uint8_t>::value, uint8_t>::type +quantize(float val, const UniformQuantizationInfo &info) +{ + int32_t tmp = static_cast<int32_t>(val / info.scale) + info.offset; + + T tmp_qua = static_cast<T>((tmp > UCHAR_MAX) ? UCHAR_MAX : tmp); + return tmp_qua; +} + +template <typename T> +inline float dequantize(const T *input, const UniformQuantizationInfo &info) +{ + return static_cast<float>((*input) - info.offset) * info.scale; } -void mul_saturate_QSYMM16_QSYMM16_QSYMM16_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, float scale, - const UniformQuantizationInfo &input1_qua_info, const UniformQuantizationInfo &input2_qua_info, const UniformQuantizationInfo &output_qua_info) +template <typename T> +void mul_saturate_quantized_8(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, float scale) { - const auto input1 = static_cast<const qsymm16_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const qsymm16_t *__restrict>(input2_ptr); - const auto output = static_cast<qsymm16_t *__restrict>(output_ptr); + const UniformQuantizationInfo input1_qua_info = in1->info()->quantization_info().uniform(); + const UniformQuantizationInfo input2_qua_info = in2->info()->quantization_info().uniform(); + const UniformQuantizationInfo output_qua_info = out->info()->quantization_info().uniform(); + + // Create input windows + Window win = window; + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); - const qsymm16x8x2_t input1_q = + // Clear X Dimension on execution window as we handle manually + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); + + const int window_step_x = 16 / sizeof(T); + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); + + const UniformQuantizationInfo tmp_qua_info = { output_qua_info.scale / scale, output_qua_info.offset }; + + execute_window_loop(win, [&](const Coordinates &) { + const auto input1_ptr = reinterpret_cast<const T *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const T *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<T *>(output.ptr()); + + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) { - vld1q_s16(input1), - vld1q_s16(input1 + 8), + const auto input1_q = wrapper::vloadq(input1_ptr + x); + const auto input2_q = wrapper::vloadq(input2_ptr + x); + + // Dequantize inputs + const float32x4x4_t in1_f32x4x4 = vdequantize(input1_q, input1_qua_info); + const float32x4x4_t in2_f32x4x4 = vdequantize(input2_q, input2_qua_info); + + const float32x4x4_t out_f32x4x4 = + { + vmulq_f32(in1_f32x4x4.val[0], in2_f32x4x4.val[0]), + vmulq_f32(in1_f32x4x4.val[1], in2_f32x4x4.val[1]), + vmulq_f32(in1_f32x4x4.val[2], in2_f32x4x4.val[2]), + vmulq_f32(in1_f32x4x4.val[3], in2_f32x4x4.val[3]), + }; + + // Quantize output + const auto result = vquantize<T>(out_f32x4x4, tmp_qua_info); + wrapper::vstore(output_ptr + x, result); } - }; - const qsymm16x8x2_t input2_q = - { + + // Compute left-over elements + for(; x < window_end_x; ++x) { - vld1q_s16(input2), - vld1q_s16(input2 + 8), + // Dequantize inputs + float tmp_in1 = dequantize(input1_ptr + x, input1_qua_info); + float tmp_in2 = dequantize(input2_ptr + x, input2_qua_info); + float tmp_f = tmp_in1 * tmp_in2; + + // Quantize output + const auto tmp_qua = quantize<T>(tmp_f, tmp_qua_info); + *(output_ptr + x) = tmp_qua; } - }; + }, + input1, input2, output); +} - // Dequantitize inputs - const float32x4x4_t in1_f32x4x4 = vdequantize(input1_q, input1_qua_info); - const float32x4x4_t in2_f32x4x4 = vdequantize(input2_q, input2_qua_info); +void mul_saturate_QSYMM16_QSYMM16_QSYMM16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, float scale) +{ + const UniformQuantizationInfo input1_qua_info = in1->info()->quantization_info().uniform(); + const UniformQuantizationInfo input2_qua_info = in2->info()->quantization_info().uniform(); + const UniformQuantizationInfo output_qua_info = out->info()->quantization_info().uniform(); - const UniformQuantizationInfo tmp_qua_info = { output_qua_info.scale / scale, output_qua_info.offset }; + // Create input windows + Window win = window; + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); - const float32x4x4_t out_f32x4x4 = - { - vmulq_f32(in1_f32x4x4.val[0], in2_f32x4x4.val[0]), - vmulq_f32(in1_f32x4x4.val[1], in2_f32x4x4.val[1]), - vmulq_f32(in1_f32x4x4.val[2], in2_f32x4x4.val[2]), - vmulq_f32(in1_f32x4x4.val[3], in2_f32x4x4.val[3]), - }; + // Clear X Dimension on execution window as we handle manually + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); - const qsymm16x8x2_t result = vquantize_qsymm16(out_f32x4x4, tmp_qua_info); - vst1q_s16(output, result.val[0]); - vst1q_s16(output + 8, result.val[1]); -} + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); -void mul_QSYMM16_QSYMM16_S32_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, int scale) -{ - ARM_COMPUTE_UNUSED(scale); - const auto input1 = static_cast<const qsymm16_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const qsymm16_t *__restrict>(input2_ptr); - const auto output = static_cast<int32_t *__restrict>(output_ptr); + const int window_step_x = 16; + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); - const qsymm16x8x2_t input1_q = - { - { - vld1q_s16(input1), - vld1q_s16(input1 + 8), - } - }; - const qsymm16x8x2_t input2_q = + const UniformQuantizationInfo tmp_qua_info = { output_qua_info.scale / scale, output_qua_info.offset }; + + execute_window_loop(win, [&](const Coordinates &) { + const auto input1_ptr = reinterpret_cast<const qsymm16_t *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const qsymm16_t *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<qsymm16_t *>(output.ptr()); + + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) { - vld1q_s16(input2), - vld1q_s16(input2 + 8), + const qsymm16x8x2_t input1_q = + { + { + vld1q_s16(input1_ptr + x), + vld1q_s16(input1_ptr + x + 8), + } + }; + const qsymm16x8x2_t input2_q = + { + { + vld1q_s16(input2_ptr + x), + vld1q_s16(input2_ptr + x + 8), + } + }; + + // Dequantize inputs + const float32x4x4_t in1_f32x4x4 = vdequantize(input1_q, input1_qua_info); + const float32x4x4_t in2_f32x4x4 = vdequantize(input2_q, input2_qua_info); + + const float32x4x4_t out_f32x4x4 = + { + vmulq_f32(in1_f32x4x4.val[0], in2_f32x4x4.val[0]), + vmulq_f32(in1_f32x4x4.val[1], in2_f32x4x4.val[1]), + vmulq_f32(in1_f32x4x4.val[2], in2_f32x4x4.val[2]), + vmulq_f32(in1_f32x4x4.val[3], in2_f32x4x4.val[3]), + }; + + const qsymm16x8x2_t result = vquantize_qsymm16(out_f32x4x4, tmp_qua_info); + vst1q_s16(output_ptr + x, result.val[0]); + vst1q_s16(output_ptr + x + 8, result.val[1]); } - }; - const int32x4x4_t in1_s32 = - { + // Compute left-over elements + for(; x < window_end_x; ++x) { - vmovl_s16(vget_low_s16(input1_q.val[0])), - vmovl_s16(vget_high_s16(input1_q.val[0])), - vmovl_s16(vget_low_s16(input1_q.val[1])), - vmovl_s16(vget_high_s16(input1_q.val[1])), + // Dequantize inputs + float tmp_in1 = static_cast<float>(*(input1_ptr + x)) * input1_qua_info.scale; + float tmp_in2 = static_cast<float>(*(input2_ptr + x)) * input2_qua_info.scale; + float tmp_f = tmp_in1 * tmp_in2; + + // Quantize output, lrintf() has same rounding mode as vcombine_s16 + int32_t tmp = lrintf(tmp_f / tmp_qua_info.scale); + qsymm16_t tmp_qua = static_cast<qsymm16_t>(tmp > SHRT_MAX) ? SHRT_MAX : ((tmp < SHRT_MIN) ? SHRT_MIN : tmp); + *(output_ptr + x) = tmp_qua; } - }; - const int32x4x4_t in2_s32 = + }, + input1, input2, output); +} + +void mul_QSYMM16_QSYMM16_S32(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, int scale) +{ + ARM_COMPUTE_UNUSED(scale); + + // Create input windows + Window win = window; + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); + + // Clear X Dimension on execution window as we handle manually + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); + + const int window_step_x = 16; + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); + + execute_window_loop(win, [&](const Coordinates &) { + const auto input1_ptr = reinterpret_cast<const qsymm16_t *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const qsymm16_t *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<int32_t *>(output.ptr()); + + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) { - vmovl_s16(vget_low_s16(input2_q.val[0])), - vmovl_s16(vget_high_s16(input2_q.val[0])), - vmovl_s16(vget_low_s16(input2_q.val[1])), - vmovl_s16(vget_high_s16(input2_q.val[1])), + const qsymm16x8x2_t input1_q = + { + { + vld1q_s16(input1_ptr + x), + vld1q_s16(input1_ptr + x + 8), + } + }; + const qsymm16x8x2_t input2_q = + { + { + vld1q_s16(input2_ptr + x), + vld1q_s16(input2_ptr + x + 8), + } + }; + + const int32x4x4_t in1_s32 = + { + { + vmovl_s16(vget_low_s16(input1_q.val[0])), + vmovl_s16(vget_high_s16(input1_q.val[0])), + vmovl_s16(vget_low_s16(input1_q.val[1])), + vmovl_s16(vget_high_s16(input1_q.val[1])), + } + }; + const int32x4x4_t in2_s32 = + { + { + vmovl_s16(vget_low_s16(input2_q.val[0])), + vmovl_s16(vget_high_s16(input2_q.val[0])), + vmovl_s16(vget_low_s16(input2_q.val[1])), + vmovl_s16(vget_high_s16(input2_q.val[1])), + } + }; + + const int32x4x4_t result = + { + { + vmulq_s32(in1_s32.val[0], in2_s32.val[0]), + vmulq_s32(in1_s32.val[1], in2_s32.val[1]), + vmulq_s32(in1_s32.val[2], in2_s32.val[2]), + vmulq_s32(in1_s32.val[3], in2_s32.val[3]), + } + }; + + vst1q_s32(output_ptr + x, result.val[0]); + vst1q_s32(output_ptr + x + 4, result.val[1]); + vst1q_s32(output_ptr + x + 8, result.val[2]); + vst1q_s32(output_ptr + x + 12, result.val[3]); } - }; - const int32x4x4_t result = - { + // Compute left-over elements + for(; x < window_end_x; ++x) { - vmulq_s32(in1_s32.val[0], in2_s32.val[0]), - vmulq_s32(in1_s32.val[1], in2_s32.val[1]), - vmulq_s32(in1_s32.val[2], in2_s32.val[2]), - vmulq_s32(in1_s32.val[3], in2_s32.val[3]), + int32_t tmp = static_cast<int32_t>(*(input1_ptr + x)) * static_cast<int32_t>(*(input2_ptr + x)); + *(output_ptr + x) = tmp; } - }; - - vst1q_s32(output, result.val[0]); - vst1q_s32(output + 4, result.val[1]); - vst1q_s32(output + 8, result.val[2]); - vst1q_s32(output + 12, result.val[3]); + }, + input1, input2, output); } template <bool is_scale255, bool is_sat> -void mul_U8_U8_U8_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, int n) +void mul_U8_U8_U8(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, int n) { - const auto input1 = static_cast<const uint8_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const uint8_t *__restrict>(input2_ptr); - const auto output = static_cast<uint8_t *__restrict>(output_ptr); + // Create input windows + Window win = window; + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); - const uint8x16_t ta1 = vld1q_u8(input1); - const uint8x16_t ta2 = vld1q_u8(input2); + // Clear X Dimension on execution window as we handle manually + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); - uint16x8_t tmp1_high = vmovl_u8(vget_high_u8(ta1)); - const uint16x8_t tmp2_high = vmovl_u8(vget_high_u8(ta2)); - uint16x8_t tmp1_low = vmovl_u8(vget_low_u8(ta1)); - const uint16x8_t tmp2_low = vmovl_u8(vget_low_u8(ta2)); + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); - tmp1_high = vmulq_u16(tmp1_high, tmp2_high); - tmp1_low = vmulq_u16(tmp1_low, tmp2_low); + const int window_step_x = 16 / sizeof(uint8_t); + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); - if(is_scale255) - { - tmp1_high = scale255_U16_U16(tmp1_high); - tmp1_low = scale255_U16_U16(tmp1_low); - } - else + execute_window_loop(win, [&](const Coordinates &) { - const int16x8_t vn = vdupq_n_s16(-n); + const auto input1_ptr = reinterpret_cast<const uint8_t *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const uint8_t *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<uint8_t *>(output.ptr()); - if(is_sat) + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) { - tmp1_high = vqshlq_u16(tmp1_high, vn); - tmp1_low = vqshlq_u16(tmp1_low, vn); + const uint8x16_t ta1 = wrapper::vloadq(input1_ptr + x); + const uint8x16_t ta2 = wrapper::vloadq(input2_ptr + x); + + uint16x8_t tmp1_high = vmovl_u8(vget_high_u8(ta1)); + const uint16x8_t tmp2_high = vmovl_u8(vget_high_u8(ta2)); + uint16x8_t tmp1_low = vmovl_u8(vget_low_u8(ta1)); + const uint16x8_t tmp2_low = vmovl_u8(vget_low_u8(ta2)); + + tmp1_high = vmulq_u16(tmp1_high, tmp2_high); + tmp1_low = vmulq_u16(tmp1_low, tmp2_low); + + if(is_scale255) + { + tmp1_high = scale255_U16_U16(tmp1_high); + tmp1_low = scale255_U16_U16(tmp1_low); + } + else + { + const int16x8_t vn = vdupq_n_s16(-n); + + if(is_sat) + { + tmp1_high = vqshlq_u16(tmp1_high, vn); + tmp1_low = vqshlq_u16(tmp1_low, vn); + } + else + { + tmp1_high = vshlq_u16(tmp1_high, vn); + tmp1_low = vshlq_u16(tmp1_low, vn); + } + } + if(is_sat) + { + vst1q_u8(output_ptr, vcombine_u8(vqmovn_u16(tmp1_low), vqmovn_u16(tmp1_high))); + } + else + { + vst1q_u8(output_ptr, vcombine_u8(vmovn_u16(tmp1_low), vmovn_u16(tmp1_high))); + } } - else + + // Compute left-over elements + for(; x < window_end_x; ++x) { - tmp1_high = vshlq_u16(tmp1_high, vn); - tmp1_low = vshlq_u16(tmp1_low, vn); - } - } + uint16_t tmp = static_cast<uint16_t>(*(input1_ptr + x)) * static_cast<uint16_t>(*(input2_ptr + x)); - if(is_sat) - { - vst1q_u8(output, vcombine_u8(vqmovn_u16(tmp1_low), vqmovn_u16(tmp1_high))); - } - else - { - vst1q_u8(output, vcombine_u8(vmovn_u16(tmp1_low), vmovn_u16(tmp1_high))); - } + if(is_scale255) + { + float tmp_f = static_cast<float>(tmp) * scale255_constant; + tmp = static_cast<uint16_t>(tmp_f + 0.5f); + } + else + { + tmp >>= n; + } + if(is_sat && tmp > 255) + { + tmp = 255; + } + *(output_ptr + x) = static_cast<uint8_t>(tmp); + } + }, + input1, input2, output); } template <bool is_scale255, bool is_sat> @@ -468,51 +578,189 @@ inline int16x8x2_t mul_S16_S16_S16_n_k(const int16x8x2_t &input1, const int16x8x } template <bool is_scale255, bool is_sat> -void mul_S16_S16_S16_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, int n) +void mul_S16_S16_S16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, int n) { - const auto input1 = static_cast<const int16_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const int16_t *__restrict>(input2_ptr); - const auto output = static_cast<int16_t *__restrict>(output_ptr); + // Create input windows + Window win = window; + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); - const int16x8x2_t ta1 = + // Clear X Dimension on execution window as we handle manually + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); + + const int window_step_x = 16; + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); + + execute_window_loop(win, [&](const Coordinates &) { + const auto input1_ptr = reinterpret_cast<const int16_t *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const int16_t *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<int16_t *>(output.ptr()); + + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) { - vld1q_s16(input1), - vld1q_s16(input1 + 8), + const int16x8x2_t ta1 = + { + { + vld1q_s16(input1_ptr + x), + vld1q_s16(input1_ptr + x + 8), + } + }; + const int16x8x2_t ta2 = + { + { + vld1q_s16(input2_ptr + x), + vld1q_s16(input2_ptr + x + 8), + } + }; + const int16x8x2_t result = mul_S16_S16_S16_n_k<is_scale255, is_sat>(ta1, ta2, n); + + vst1q_s16(output_ptr + x, result.val[0]); + vst1q_s16(output_ptr + x + 8, result.val[1]); } - }; - const int16x8x2_t ta2 = - { + + // Compute left-over elements + for(; x < window_end_x; ++x) { - vld1q_s16(input2), - vld1q_s16(input2 + 8), - } - }; - const int16x8x2_t result = mul_S16_S16_S16_n_k<is_scale255, is_sat>(ta1, ta2, n); + int32_t tmp = static_cast<int32_t>(*(input1_ptr + x)) * static_cast<int32_t>(*(input2_ptr + x)); - vst1q_s16(output, result.val[0]); - vst1q_s16(output + 8, result.val[1]); + if(is_scale255) + { + float tmp_f = static_cast<float>(tmp) * scale255_constant; + + tmp = static_cast<int32_t>(tmp_f + 0.5f); + } + else + { + if(tmp >= 0) + { + tmp >>= n; + } + else + { + uint32_t mask = (1u << n) - 1; + tmp = (tmp + static_cast<int32_t>(mask)) >> n; + } + } + if(is_sat) + { + tmp = (tmp > SHRT_MAX) ? SHRT_MAX : ((tmp < SHRT_MIN) ? SHRT_MIN : tmp); + } + *(output_ptr + x) = static_cast<int16_t>(tmp); + } + }, + input1, input2, output); } -void mul_F32_F32_F32_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, float scale) +void mul_F32_F32_F32(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, float scale) { - const auto input1 = static_cast<const float *__restrict>(input1_ptr); - const auto input2 = static_cast<const float *__restrict>(input2_ptr); - const auto output = static_cast<float *__restrict>(output_ptr); + // Create input windows + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); + + // Clear X Dimension on execution window as we handle manually + Window win = window; + win.set(Window::DimX, Window::Dimension(0, 1, 1)); - const float32x4x4_t ta1 = vld4q_f32(input1); - const float32x4x4_t ta2 = vld4q_f32(input2); - const float32x4_t scale_vec = vdupq_n_f32(scale); - const float32x4x4_t result = + constexpr int window_step_x = 16 / sizeof(float); + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); + const bool is_broadcast_across_x = (input1_win.x().step() == 0) || (input2_win.x().step() == 0); + + Iterator input1(in1, window.broadcast_if_dimension_le_one(in1->info()->tensor_shape())); + Iterator input2(in2, window.broadcast_if_dimension_le_one(in2->info()->tensor_shape())); + Iterator output(out, window); + + using ExactTagType = typename wrapper::traits::neon_vector<float, window_step_x>::tag_type; + + if(is_broadcast_across_x) { + const bool is_broadcast_input_2 = input2_win.x().step() == 0; + Window broadcast_win = is_broadcast_input_2 ? input2_win : input1_win; + Window non_broadcast_win = !is_broadcast_input_2 ? input2_win : input1_win; + const ITensor *broadcast_tensor = is_broadcast_input_2 ? in2 : in1; + const ITensor *non_broadcast_tensor = !is_broadcast_input_2 ? in2 : in1; + + // Clear X Dimension on execution window as we handle manually + non_broadcast_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator broadcast_input(broadcast_tensor, broadcast_win); + Iterator non_broadcast_input(non_broadcast_tensor, non_broadcast_win); + Iterator output(out, win); + + execute_window_loop(win, [&](const Coordinates &) { - vmulq_f32(vmulq_f32(ta1.val[0], ta2.val[0]), scale_vec), - vmulq_f32(vmulq_f32(ta1.val[1], ta2.val[1]), scale_vec), - vmulq_f32(vmulq_f32(ta1.val[2], ta2.val[2]), scale_vec), - vmulq_f32(vmulq_f32(ta1.val[3], ta2.val[3]), scale_vec) - } - }; - vst4q_f32(output, result); + const auto non_broadcast_input_ptr = reinterpret_cast<const float *>(non_broadcast_input.ptr()); + const auto output_ptr = reinterpret_cast<float *>(output.ptr()); + + const float broadcast_value = *reinterpret_cast<const float *>(broadcast_input.ptr()); + const auto broadcast_value_vec = wrapper::vdup_n(broadcast_value, ExactTagType{}); + const auto scale_vec = wrapper::vdup_n(scale, ExactTagType{}); + + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + const auto non_broadcast_v = wrapper::vloadq(non_broadcast_input_ptr + x); + auto res = wrapper::vmul(wrapper::vmul(broadcast_value_vec, non_broadcast_v), scale_vec); + wrapper::vstore(output_ptr + x, res); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const auto non_broadcast_v = *(non_broadcast_input_ptr + x); + *(output_ptr + x) = broadcast_value * non_broadcast_v * scale; + } + }, + broadcast_input, non_broadcast_input, output); + } + else + { + // Clear X Dimension on execution window as we handle manually + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); + + execute_window_loop(win, [&](const Coordinates &) + { + const auto input1_ptr = reinterpret_cast<const float *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const float *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<float *>(output.ptr()); + + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + const auto ta1 = wrapper::vloadq(input1_ptr + x); + const auto ta2 = wrapper::vloadq(input2_ptr + x); + const auto scale_vec = wrapper::vdup_n(scale, ExactTagType{}); + const auto res = wrapper::vmul(wrapper::vmul(ta1, ta2), scale_vec); + wrapper::vstore(output_ptr + x, res); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const auto ta1 = *(input1_ptr + x); + const auto ta2 = *(input2_ptr + x); + *(output_ptr + x) = ta1 * ta2 * scale; + } + }, + input1, input2, output); + } } void c_mul_F32_F32_F32_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr) @@ -544,138 +792,275 @@ void c_mul_F32_F32_F32_n(const void *__restrict input1_ptr, const void *__restri wrapper::vstore(output, res); } -void mul_F16_F16_F16_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, float scale) -{ #ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC - const auto input1 = static_cast<const float16_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const float16_t *__restrict>(input2_ptr); - const auto output = static_cast<float16_t *__restrict>(output_ptr); - const float16x8x2_t ta1 = - { - { - vld1q_f16(input1), - vld1q_f16(input1 + 8), - } - }; - const float16x8x2_t ta2 = +void mul_F16_F16_F16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, float scale) +{ + // Create input windows + Window win = window; + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); + + // Clear X Dimension on execution window as we handle manually + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); + + const int window_step_x = 16; + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); + + execute_window_loop(win, [&](const Coordinates &) { + const auto input1_ptr = reinterpret_cast<const float16_t *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const float16_t *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<float16_t *>(output.ptr()); + + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) { - vld1q_f16(input2), - vld1q_f16(input2 + 8), + const float16x8x2_t ta1 = + { + { + vld1q_f16(input1_ptr + x), + vld1q_f16(input1_ptr + x + 8), + } + }; + const float16x8x2_t ta2 = + { + { + vld1q_f16(input2_ptr + x), + vld1q_f16(input2_ptr + x + 8), + } + }; + const float16x8_t scale_vec = vdupq_n_f16(scale); + const float16x8x2_t result = + { + { + vmulq_f16(vmulq_f16(ta1.val[0], ta2.val[0]), scale_vec), + vmulq_f16(vmulq_f16(ta1.val[1], ta2.val[1]), scale_vec), + } + }; + vst1q_f16(output_ptr + x, result.val[0]); + vst1q_f16(output_ptr + x + 8, result.val[1]); } - }; - const float16x8_t scale_vec = vdupq_n_f16(scale); - const float16x8x2_t result = - { + + // Compute left-over elements + for(; x < window_end_x; ++x) { - vmulq_f16(vmulq_f16(ta1.val[0], ta2.val[0]), scale_vec), - vmulq_f16(vmulq_f16(ta1.val[1], ta2.val[1]), scale_vec), + const auto ta1 = *(input1_ptr + x); + const auto ta2 = *(input2_ptr + x); + *(output_ptr + x) = ta1 * ta2 * scale; } - }; - vst1q_f16(output, result.val[0]); - vst1q_f16(output + 8, result.val[1]); -#else /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ - ARM_COMPUTE_UNUSED(input1_ptr); - ARM_COMPUTE_UNUSED(input2_ptr); - ARM_COMPUTE_UNUSED(output_ptr); - ARM_COMPUTE_UNUSED(scale); - ARM_COMPUTE_ERROR("Not supported. Recompile the library with arch=arm64-v8.2-a."); -#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ + }, + input1, input2, output); } +#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ template <bool is_scale255, bool is_sat> -void mul_U8_U8_S16_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, int n) +void mul_U8_U8_S16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, int n) { - const auto input1 = static_cast<const uint8_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const uint8_t *__restrict>(input2_ptr); - const auto output = static_cast<int16_t *__restrict>(output_ptr); + // Create input windows + Window win = window; + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); - const uint8x16_t bv = vld1q_u8(input2); - const uint8x16_t av = vld1q_u8(input1); + // Clear X Dimension on execution window as we handle manually + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); - uint16x8_t tmp_low = vmovl_u8(vget_low_u8(av)); - uint16x8_t tmp_high = vmovl_u8(vget_high_u8(av)); - tmp_low = vmulq_u16(tmp_low, vmovl_u8(vget_low_u8(bv))); - tmp_high = vmulq_u16(tmp_high, vmovl_u8(vget_high_u8(bv))); + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); - if(is_scale255) - { - tmp_low = scale255_U16_U16(tmp_low); - tmp_high = scale255_U16_U16(tmp_high); - } - else + const int window_step_x = 16 / sizeof(uint8_t); + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); + + execute_window_loop(win, [&](const Coordinates &) { - const int16x8_t vn = vdupq_n_s16(-n); + const auto input1_ptr = reinterpret_cast<const uint8_t *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const uint8_t *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<int16_t *>(output.ptr()); - if(is_sat) + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) { - tmp_low = vqshlq_u16(tmp_low, vn); - tmp_high = vqshlq_u16(tmp_high, vn); + const uint8x16_t bv = wrapper::vloadq(input2_ptr + x); + const uint8x16_t av = wrapper::vloadq(input1_ptr + x); + + uint16x8_t tmp_low = vmovl_u8(vget_low_u8(av)); + uint16x8_t tmp_high = vmovl_u8(vget_high_u8(av)); + tmp_low = vmulq_u16(tmp_low, vmovl_u8(vget_low_u8(bv))); + tmp_high = vmulq_u16(tmp_high, vmovl_u8(vget_high_u8(bv))); + + if(is_scale255) + { + tmp_low = scale255_U16_U16(tmp_low); + tmp_high = scale255_U16_U16(tmp_high); + } + else + { + const int16x8_t vn = vdupq_n_s16(-n); + + if(is_sat) + { + tmp_low = vqshlq_u16(tmp_low, vn); + tmp_high = vqshlq_u16(tmp_high, vn); + } + else + { + tmp_low = vshlq_u16(tmp_low, vn); + tmp_high = vshlq_u16(tmp_high, vn); + } + } + + if(is_sat) + { + static const uint16x8_t max = vdupq_n_u16(SHRT_MAX); + + tmp_low = vminq_u16(tmp_low, max); + tmp_high = vminq_u16(tmp_high, max); + } + + vst1q_s16(output_ptr + x, vreinterpretq_s16_u16(tmp_low)); + vst1q_s16(output_ptr + x + 8, vreinterpretq_s16_u16(tmp_high)); } - else + + // Compute left-over elements + for(; x < window_end_x; ++x) { - tmp_low = vshlq_u16(tmp_low, vn); - tmp_high = vshlq_u16(tmp_high, vn); - } - } + int32_t tmp = static_cast<int32_t>(*(input1_ptr + x)) * static_cast<int32_t>(*(input2_ptr + x)); - if(is_sat) - { - static const uint16x8_t max = vdupq_n_u16(SHRT_MAX); + if(is_scale255) + { + float tmp_f = static_cast<float>(tmp) * scale255_constant; + tmp = static_cast<int32_t>(tmp_f + 0.5f); + } + else + { + tmp >>= n; + } - tmp_low = vminq_u16(tmp_low, max); - tmp_high = vminq_u16(tmp_high, max); - } + if(is_sat) + { + tmp = (tmp > SHRT_MAX) ? SHRT_MAX : tmp; + } - vst1q_s16(output, vreinterpretq_s16_u16(tmp_low)); - vst1q_s16(output + 8, vreinterpretq_s16_u16(tmp_high)); + *(output_ptr + x) = static_cast<int16_t>(tmp); + } + }, + input1, input2, output); } template <bool is_scale255, bool is_sat> -void mul_S16_U8_S16_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, int n) +void mul_S16_U8_S16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, int n) { - const auto input1 = static_cast<const int16_t *__restrict>(input1_ptr); - const auto input2 = static_cast<const uint8_t *__restrict>(input2_ptr); - const auto output = static_cast<int16_t *__restrict>(output_ptr); + // Create input windows + Window win = window; + Window input1_win = window.broadcast_if_dimension_le_one(in1->info()->tensor_shape()); + Window input2_win = window.broadcast_if_dimension_le_one(in2->info()->tensor_shape()); - const int16x8x2_t ta1 = - { - { - vld1q_s16(input1), - vld1q_s16(input1 + 8), - } - }; - const uint8x8x2_t ta2u = + // Clear X Dimension on execution window as we handle manually + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input1_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + input2_win.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator input1(in1, input1_win); + Iterator input2(in2, input2_win); + Iterator output(out, win); + + const int window_step_x = 16; + const auto window_start_x = static_cast<int>(window.x().start()); + const auto window_end_x = static_cast<int>(window.x().end()); + + execute_window_loop(win, [&](const Coordinates &) { + const auto input1_ptr = reinterpret_cast<const int16_t *>(input1.ptr()); + const auto input2_ptr = reinterpret_cast<const uint8_t *>(input2.ptr()); + const auto output_ptr = reinterpret_cast<int16_t *>(output.ptr()); + + // Compute window_step_x elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) { - vld1_u8(input2), - vld1_u8(input2 + 8), + const int16x8x2_t ta1 = + { + { + vld1q_s16(input1_ptr + x), + vld1q_s16(input1_ptr + x + 8), + } + }; + const uint8x8x2_t ta2u = + { + { + vld1_u8(input2_ptr + x), + vld1_u8(input2_ptr + x + 8), + } + }; + const int16x8x2_t ta2 = + { + { + vreinterpretq_s16_u16(vmovl_u8(ta2u.val[0])), + vreinterpretq_s16_u16(vmovl_u8(ta2u.val[1])) + } + }; + + const int16x8x2_t result = mul_S16_S16_S16_n_k<is_scale255, is_sat>(ta1, ta2, n); + + vst1q_s16(output_ptr + x, result.val[0]); + vst1q_s16(output_ptr + x + 8, result.val[1]); } - }; - const int16x8x2_t ta2 = - { + + // Compute left-over elements + for(; x < window_end_x; ++x) { - vreinterpretq_s16_u16(vmovl_u8(ta2u.val[0])), - vreinterpretq_s16_u16(vmovl_u8(ta2u.val[1])) - } - }; + int32_t tmp = static_cast<int32_t>(*(input1_ptr + x)) * static_cast<int32_t>(*(input2_ptr + x)); - const int16x8x2_t result = mul_S16_S16_S16_n_k<is_scale255, is_sat>(ta1, ta2, n); + if(is_scale255) + { + float tmp_f = static_cast<float>(tmp) * scale255_constant; - vst1q_s16(output, result.val[0]); - vst1q_s16(output + 8, result.val[1]); + tmp = static_cast<int32_t>(tmp_f + 0.5f); + } + else + { + if(tmp >= 0) + { + tmp >>= n; + } + else + { + uint32_t mask = (1u << n) - 1; + tmp = (tmp + static_cast<int32_t>(mask)) >> n; + } + } + if(is_sat) + { + tmp = (tmp > SHRT_MAX) ? SHRT_MAX : ((tmp < SHRT_MIN) ? SHRT_MIN : tmp); + } + *(output_ptr + x) = static_cast<int16_t>(tmp); + } + }, + input1, input2, output); } template <bool is_scale255, bool is_sat> -void mul_U8_S16_S16_n(const void *__restrict input1_ptr, const void *__restrict input2_ptr, void *__restrict output_ptr, int n) +void mul_U8_S16_S16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window, int n) { // Simply swap the two input buffers - mul_S16_U8_S16_n<is_scale255, is_sat>(input2_ptr, input1_ptr, output_ptr, n); + mul_S16_U8_S16<is_scale255, is_sat>(in2, in1, out, window, n); } } // namespace NEPixelWiseMultiplicationKernel::NEPixelWiseMultiplicationKernel() - : _func_float(nullptr), _func_int(nullptr), _func_quantized(nullptr), _input1(nullptr), _input2(nullptr), _output(nullptr), _scale{ 0 }, _scale_exponent{ 0 }, _run_optimized_qasymm8(false) + : _func_float(nullptr), _func_int(nullptr), _func_quantized(nullptr), _input1(nullptr), _input2(nullptr), _output(nullptr), _scale{ 0 }, _scale_exponent{ 0 } { } @@ -686,19 +1071,21 @@ void NEPixelWiseMultiplicationKernel::configure(const ITensor *input1, const ITe ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input1->info(), input2->info(), output->info(), scale, overflow_policy, rounding_policy)); - // Configure kernel window - auto win_config = validate_and_configure_window(input1->info(), input2->info(), output->info()); - ARM_COMPUTE_ERROR_THROW_ON(win_config.first); + const std::pair<TensorShape, ValidRegion> broadcast_pair = ITensorInfo::broadcast_shape_and_valid_region(*input1->info(), *input2->info()); + const TensorShape &out_shape = broadcast_pair.first; + const ValidRegion &valid_region = broadcast_pair.second; + + // Auto initialize output if not initialized + set_shape_if_empty(*output->info(), out_shape); - _input1 = input1; - _input2 = input2; - _output = output; - _scale = scale; - _scale_exponent = 0; - _func_quantized = nullptr; - _func_int = nullptr; - _func_float = nullptr; - _run_optimized_qasymm8 = false; + _input1 = input1; + _input2 = input2; + _output = output; + _scale = scale; + _scale_exponent = 0; + _func_quantized = nullptr; + _func_int = nullptr; + _func_float = nullptr; bool is_scale_255 = false; // Check and validate scaling factor @@ -722,93 +1109,109 @@ void NEPixelWiseMultiplicationKernel::configure(const ITensor *input1, const ITe const DataType dt_output = output->info()->data_type(); const bool is_sat = (overflow_policy == ConvertPolicy::SATURATE); - if(dt_input1 == DataType::QASYMM8 && dt_input2 == DataType::QASYMM8) + switch(dt_input1) { - _run_optimized_qasymm8 = true; - } - else if(dt_input1 == DataType::QASYMM8_SIGNED && dt_input2 == DataType::QASYMM8_SIGNED) - { - _func_quantized = &mul_saturate_QASYMM8_SIGNED_QASYMM8_SIGNED_QASYMM8_SIGNED_n; - } - else if(dt_input1 == DataType::QSYMM16 && dt_input2 == DataType::QSYMM16 && dt_output == DataType::QSYMM16) - { - _func_quantized = &mul_saturate_QSYMM16_QSYMM16_QSYMM16_n; - } - else if(dt_input1 == DataType::QSYMM16 && dt_input2 == DataType::QSYMM16 && dt_output == DataType::S32) - { - _func_int = &mul_QSYMM16_QSYMM16_S32_n; - } - else if(DataType::U8 == dt_input1 && DataType::U8 == dt_input2 && DataType::U8 == dt_output) - { - if(is_scale_255) - { - _func_int = is_sat ? &mul_U8_U8_U8_n<true, true> : &mul_U8_U8_U8_n<true, false>; - } - else - { - _func_int = is_sat ? &mul_U8_U8_U8_n<false, true> : &mul_U8_U8_U8_n<false, false>; - } - } - else if(DataType::S16 == dt_input1 && DataType::S16 == dt_input2 && DataType::S16 == dt_output) - { - if(is_scale_255) - { - _func_int = is_sat ? &mul_S16_S16_S16_n<true, true> : &mul_S16_S16_S16_n<true, false>; - } - else - { - _func_int = is_sat ? &mul_S16_S16_S16_n<false, true> : &mul_S16_S16_S16_n<false, false>; - } - } - else if(DataType::S16 == dt_input1 && DataType::U8 == dt_input2 && DataType::S16 == dt_output) - { - if(is_scale_255) - { - _func_int = is_sat ? &mul_S16_U8_S16_n<true, true> : &mul_S16_U8_S16_n<true, false>; - } - else - { - _func_int = is_sat ? &mul_S16_U8_S16_n<false, true> : &mul_S16_U8_S16_n<false, false>; - } - } - else if(DataType::U8 == dt_input1 && DataType::S16 == dt_input2 && DataType::S16 == dt_output) - { - if(is_scale_255) - { - _func_int = is_sat ? &mul_U8_S16_S16_n<true, true> : &mul_U8_S16_S16_n<true, false>; - } - else - { - _func_int = is_sat ? &mul_U8_S16_S16_n<false, true> : &mul_U8_S16_S16_n<false, false>; - } - } - else if(DataType::U8 == dt_input1 && DataType::U8 == dt_input2 && DataType::S16 == dt_output) - { - if(is_scale_255) - { - _func_int = is_sat ? &mul_U8_U8_S16_n<true, true> : &mul_U8_U8_S16_n<true, false>; - } - else - { - _func_int = is_sat ? &mul_U8_U8_S16_n<false, true> : &mul_U8_U8_S16_n<false, false>; - } - } - else if(DataType::F16 == dt_input1 && DataType::F16 == dt_input2 && DataType::F16 == dt_output) - { - _func_float = &mul_F16_F16_F16_n; - _func_int = nullptr; - } - else if(DataType::F32 == dt_input1 && DataType::F32 == dt_input2 && DataType::F32 == dt_output) - { - _func_float = &mul_F32_F32_F32_n; - _func_int = nullptr; - } - else - { - ARM_COMPUTE_ERROR("You called with the wrong img formats"); + case DataType::QASYMM8: + if(dt_input2 == DataType::QASYMM8 && dt_output == DataType::QASYMM8) + { + _func_quantized = &mul_saturate_quantized_8<uint8_t>; + } + break; + case DataType::QASYMM8_SIGNED: + if(dt_input2 == DataType::QASYMM8_SIGNED) + { + _func_quantized = &mul_saturate_quantized_8<int8_t>; + ; + } + break; + case DataType::QSYMM16: + if(dt_input2 == DataType::QSYMM16 && dt_output == DataType::QSYMM16) + { + _func_quantized = &mul_saturate_QSYMM16_QSYMM16_QSYMM16; + } + else if(dt_input2 == DataType::QSYMM16 && dt_output == DataType::S32) + { + _func_int = &mul_QSYMM16_QSYMM16_S32; + } + break; + case DataType::S16: + if(DataType::U8 == dt_input2 && DataType::S16 == dt_output) + { + if(is_scale_255) + { + _func_int = is_sat ? &mul_S16_U8_S16<true, true> : &mul_S16_U8_S16<true, false>; + } + else + { + _func_int = is_sat ? &mul_S16_U8_S16<false, true> : &mul_S16_U8_S16<false, false>; + } + } + if(DataType::S16 == dt_input2 && DataType::S16 == dt_output) + { + if(is_scale_255) + { + _func_int = is_sat ? &mul_S16_S16_S16<true, true> : &mul_S16_S16_S16<true, false>; + } + else + { + _func_int = is_sat ? &mul_S16_S16_S16<false, true> : &mul_S16_S16_S16<false, false>; + } + } + break; + case DataType::U8: + if(DataType::U8 == dt_input2 && DataType::U8 == dt_output) + { + if(is_scale_255) + { + _func_int = is_sat ? &mul_U8_U8_U8<true, true> : &mul_U8_U8_U8<true, false>; + } + else + { + _func_int = is_sat ? &mul_U8_U8_U8<false, true> : &mul_U8_U8_U8<false, false>; + } + } + else if(DataType::U8 == dt_input2 && DataType::S16 == dt_output) + { + if(is_scale_255) + { + _func_int = is_sat ? &mul_U8_U8_S16<true, true> : &mul_U8_U8_S16<true, false>; + } + else + { + _func_int = is_sat ? &mul_U8_U8_S16<false, true> : &mul_U8_U8_S16<false, false>; + } + } + else if(DataType::S16 == dt_input2 && DataType::S16 == dt_output) + { + if(is_scale_255) + { + _func_int = is_sat ? &mul_U8_S16_S16<true, true> : &mul_U8_S16_S16<true, false>; + } + else + { + _func_int = is_sat ? &mul_U8_S16_S16<false, true> : &mul_U8_S16_S16<false, false>; + } + } + break; +#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC + case DataType::F16: + _func_float = &mul_F16_F16_F16; + break; +#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ + case DataType::F32: + _func_float = &mul_F32_F32_F32; + break; + default: + ARM_COMPUTE_ERROR("You called with the wrong img formats"); } - INEKernel::configure(win_config.second); + // Configure kernel window + Coordinates coord; + coord.set_num_dimensions(output->info()->num_dimensions()); + output->info()->set_valid_region(valid_region); + Window win = calculate_max_window(valid_region, Steps()); + + INEKernel::configure(win); } Status NEPixelWiseMultiplicationKernel::validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float scale, ConvertPolicy overflow_policy, @@ -816,7 +1219,6 @@ Status NEPixelWiseMultiplicationKernel::validate(const ITensorInfo *input1, cons { ARM_COMPUTE_ERROR_ON_NULLPTR(input1, input2, output); ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input1, input2, output, scale, overflow_policy, rounding_policy)); - ARM_COMPUTE_RETURN_ON_ERROR(validate_and_configure_window(input1->clone().get(), input2->clone().get(), output->clone().get()).first); return Status{}; } @@ -827,97 +1229,21 @@ void NEPixelWiseMultiplicationKernel::run(const Window &window, const ThreadInfo ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window); - const TensorShape &in_shape1 = _input1->info()->tensor_shape(); - const TensorShape &in_shape2 = _input2->info()->tensor_shape(); - const TensorShape &out_shape = _output->info()->tensor_shape(); - - bool can_collapse = true; - if(std::min(in_shape1.total_size(), in_shape2.total_size()) > 1) - { - can_collapse = (std::min(in_shape1.num_dimensions(), in_shape2.num_dimensions()) > Window::DimZ); - for(size_t d = Window::DimZ; can_collapse && (d < out_shape.num_dimensions()); ++d) - { - can_collapse = (in_shape1[d] == in_shape2[d]); - } - } - - bool has_collapsed = false; - Window collapsed = can_collapse ? window.collapse_if_possible(INEKernel::window(), Window::DimZ, &has_collapsed) : window; - - const TensorShape &in_shape1_collapsed = has_collapsed ? in_shape1.collapsed_from(Window::DimZ) : in_shape1; - const TensorShape &in_shape2_collapsed = has_collapsed ? in_shape2.collapsed_from(Window::DimZ) : in_shape2; - - Window slice = collapsed.first_slice_window_3D(); - Window slice_input1 = slice.broadcast_if_dimension_le_one(in_shape1_collapsed); - Window slice_input2 = slice.broadcast_if_dimension_le_one(in_shape2_collapsed); - - Iterator input1(_input1, slice_input1); - Iterator input2(_input2, slice_input2); - Iterator output(_output, slice); - - if((_run_optimized_qasymm8) || (_func_quantized != nullptr)) + if(_func_quantized != nullptr) { - if(_run_optimized_qasymm8) - { - const int32x4_t input1_voffset = vdupq_n_s32(_input1->info()->quantization_info().uniform().offset); - const float32x4_t input1_vscale = vdupq_n_f32(_input1->info()->quantization_info().uniform().scale); - const int32x4_t input2_voffset = vdupq_n_s32(_input2->info()->quantization_info().uniform().offset); - const float32x4_t input2_vscale = vdupq_n_f32(_input2->info()->quantization_info().uniform().scale); - const float32x4_t output_voffset = vdupq_n_f32(static_cast<float>(_output->info()->quantization_info().uniform().offset)); - const float output_scale = _output->info()->quantization_info().uniform().scale; - const float32x4_t vinvscale = vdupq_n_f32(1.f / (output_scale / _scale)); - - execute_window_loop(collapsed, [&](const Coordinates &) - { - mul_saturate_QASYMM8_QASYMM8_QASYMM8_n_opt(input1.ptr(), input2.ptr(), output.ptr(), - input1_vscale, input1_voffset, input2_vscale, input2_voffset, output_voffset, vinvscale); - ARM_COMPUTE_UNUSED(collapsed.slide_window_slice_3D(slice_input1)); - ARM_COMPUTE_UNUSED(collapsed.slide_window_slice_3D(slice_input2)); - }, - input1, input2, output); - } - else - { - execute_window_loop(collapsed, [&](const Coordinates &) - { - (*_func_quantized)(input1.ptr(), input2.ptr(), output.ptr(), _scale, - _input1->info()->quantization_info().uniform(), _input2->info()->quantization_info().uniform(), _output->info()->quantization_info().uniform()); - ARM_COMPUTE_UNUSED(collapsed.slide_window_slice_3D(slice_input1)); - ARM_COMPUTE_UNUSED(collapsed.slide_window_slice_3D(slice_input2)); - }, - input1, input2, output); - } + (*_func_quantized)(_input1, _input2, _output, window, _scale); } else if(_func_int != nullptr) { - execute_window_loop(collapsed, [&](const Coordinates &) - { - (*_func_int)(input1.ptr(), input2.ptr(), output.ptr(), _scale_exponent); - ARM_COMPUTE_UNUSED(collapsed.slide_window_slice_3D(slice_input1)); - ARM_COMPUTE_UNUSED(collapsed.slide_window_slice_3D(slice_input2)); - }, - input1, input2, output); + (*_func_int)(_input1, _input2, _output, window, _scale_exponent); } else { ARM_COMPUTE_ERROR_ON(_func_float == nullptr); - execute_window_loop(collapsed, [&](const Coordinates &) - { - (*_func_float)(input1.ptr(), input2.ptr(), output.ptr(), _scale); - ARM_COMPUTE_UNUSED(collapsed.slide_window_slice_3D(slice_input1)); - ARM_COMPUTE_UNUSED(collapsed.slide_window_slice_3D(slice_input2)); - }, - input1, input2, output); + (*_func_float)(_input1, _input2, _output, window, _scale); } } -BorderSize NEPixelWiseMultiplicationKernel::border_size() const -{ - const unsigned int replicateSize = _output->info()->dimension(0) - std::min(_input1->info()->dimension(0), _input2->info()->dimension(0)); - const unsigned int border = std::min<unsigned int>(num_elems_processed_per_iteration - 1U, replicateSize); - return BorderSize{ 0, border, 0, 0 }; -} - namespace { constexpr unsigned int num_elems_processed_per_iteration_complex = 2; |