From c7f9b893b8edc5660542821e2d0508460bc40225 Mon Sep 17 00:00:00 2001 From: Gian Marco Date: Thu, 30 Nov 2017 14:31:13 +0000 Subject: COMPMID-722 - Support for vector-matrix in GEMMLowp (NEON) This patch includes COMPMID-716 as well - Added vector-matrix case in NEGEMMLowpMatrixMultiplyKernel - Added benchmarks for NEON and OpenCL Change-Id: I715cd25e8668a4d6c8127e9a298a865e7713267f Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/111468 Tested-by: BSG Visual Compute Jenkins server to access repositories on http://mpd-gerrit.cambridge.arm.com Reviewed-by: Georgios Pinitas --- .../kernels/NEGEMMLowpMatrixMultiplyKernel.cpp | 701 ++++++++++++++++++--- 1 file changed, 613 insertions(+), 88 deletions(-) (limited to 'src/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.cpp') diff --git a/src/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.cpp b/src/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.cpp index 208a60ce27..a68a01f6a6 100644 --- a/src/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.cpp +++ b/src/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.cpp @@ -42,81 +42,439 @@ using namespace arm_compute; namespace arm_compute { -class Coordinates; -} // namespace arm_compute - namespace { -Error validate_arguments(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output) +void inline vector_matrix_multiply_u8(Iterator &ina, Iterator &inb, Iterator &out, int width_a, int width_b, size_t stride_b, const Window &window) { - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input0, 1, DataType::QASYMM8, DataType::S8, DataType::U8); - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input0, input1); - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(output, 1, DataType::S32); + execute_window_loop(window, [&](const Coordinates & id) + { + if(id.x() > width_b) + { + return; + } - TensorShape in0_shape = input0->tensor_shape(); - TensorShape in1_shape = input1->tensor_shape(); - TensorShape out_shape = output->tensor_shape(); + // Note: Since the input are all positives, we can use uint32_t + // Accumulators for the block 0 + uint32x4x4_t c0 = + { + { + vdupq_n_u32(0), + vdupq_n_u32(0), + vdupq_n_u32(0), + vdupq_n_u32(0) + } + }; - in0_shape.collapse(2); - in1_shape.collapse(2); - out_shape.collapse(2); + auto vec_a = reinterpret_cast(ina.ptr()); + auto matrix_b = reinterpret_cast(inb.ptr()); + auto vec_a_end_addr = vec_a + width_a; - ARM_COMPUTE_RETURN_ERROR_ON_MSG(in0_shape[2] != out_shape[2], "Output tensor must have the same number of batches of input0 tensor"); - ARM_COMPUTE_RETURN_ERROR_ON_MSG(in1_shape[2] != 1 && in0_shape[2] != in1_shape[2], "Input1 tensor must have the same number of batches of input0 or the number of batches must be set to 1"); + // This for loop performs 8 accumulations + for(; vec_a <= (vec_a_end_addr - 8);) + { + const uint8x8_t a00_u8 = vld1_u8(vec_a); + const uint8x16_t b00_u8 = vld1q_u8(matrix_b + 0 * stride_b); + const uint8x16_t b10_u8 = vld1q_u8(matrix_b + 1 * stride_b); + const uint8x16_t b20_u8 = vld1q_u8(matrix_b + 2 * stride_b); + const uint8x16_t b30_u8 = vld1q_u8(matrix_b + 3 * stride_b); + const uint8x16_t b40_u8 = vld1q_u8(matrix_b + 4 * stride_b); + const uint8x16_t b50_u8 = vld1q_u8(matrix_b + 5 * stride_b); + const uint8x16_t b60_u8 = vld1q_u8(matrix_b + 6 * stride_b); + const uint8x16_t b70_u8 = vld1q_u8(matrix_b + 7 * stride_b); + + // Convert a00_u8 to uint16_t and get the lower part + const uint16x4x2_t a00_u16 = + { + { + vget_low_u16(vmovl_u8(a00_u8)), + vget_high_u16(vmovl_u8(a00_u8)) + } + }; - return Error{}; -} + const uint16x4x4_t b00_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b00_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b00_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b00_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b00_u8))) + } + }; -std::pair validate_and_configure_window(ITensorInfo *input0, ITensorInfo *input1, ITensorInfo *output) -{ - constexpr unsigned int num_elems_processed_per_iteration_x = 16; - constexpr unsigned int num_elems_processed_per_iteration_y = 4; + const uint16x4x4_t b10_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b10_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b10_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b10_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b10_u8))) + } + }; - Window win = calculate_max_window(*output, Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y)); + const uint16x4x4_t b20_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b20_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b20_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b20_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b20_u8))) + } + }; - AccessWindowStatic in0_access(input0, 0, 0, ceil_to_multiple(input0->dimension(0), 8), input0->dimension(1)); - AccessWindowHorizontal in1_access(input1, 0, num_elems_processed_per_iteration_x); - AccessWindowRectangle output_access(output, 0, 0, num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y); + const uint16x4x4_t b30_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b30_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b30_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b30_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b30_u8))) + } + }; - bool window_changed = update_window_and_padding(win, in0_access, in1_access, output_access); + const uint16x4x4_t b40_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b40_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b40_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b40_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b40_u8))) + } + }; - output_access.set_valid_region(win, ValidRegion(Coordinates(0, 0), output->tensor_shape())); + const uint16x4x4_t b50_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b50_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b50_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b50_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b50_u8))) + } + }; - Error err = (window_changed) ? ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Insufficient Padding!") : Error{}; - return std::make_pair(err, win); -} -} // namespace + const uint16x4x4_t b60_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b60_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b60_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b60_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b60_u8))) + } + }; -NEGEMMLowpMatrixMultiplyKernel::NEGEMMLowpMatrixMultiplyKernel() - : _input0(nullptr), _input1(nullptr), _output(nullptr), _slide_matrix_b(true) -{ -} + const uint16x4x4_t b70_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b70_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b70_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b70_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b70_u8))) + } + }; -void NEGEMMLowpMatrixMultiplyKernel::configure(const ITensor *input0, const ITensor *input1, ITensor *output) -{ - ARM_COMPUTE_ERROR_ON_NULLPTR(input0, input1, output); - ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input0->info(), input1->info(), output->info())); + // Accumulate 0: + c0.val[0] = vmlal_lane_u16(c0.val[0], b00_u16.val[0], a00_u16.val[0], 0); + c0.val[1] = vmlal_lane_u16(c0.val[1], b00_u16.val[1], a00_u16.val[0], 0); + c0.val[2] = vmlal_lane_u16(c0.val[2], b00_u16.val[2], a00_u16.val[0], 0); + c0.val[3] = vmlal_lane_u16(c0.val[3], b00_u16.val[3], a00_u16.val[0], 0); + + // Accumulate 1: + c0.val[0] = vmlal_lane_u16(c0.val[0], b10_u16.val[0], a00_u16.val[0], 1); + c0.val[1] = vmlal_lane_u16(c0.val[1], b10_u16.val[1], a00_u16.val[0], 1); + c0.val[2] = vmlal_lane_u16(c0.val[2], b10_u16.val[2], a00_u16.val[0], 1); + c0.val[3] = vmlal_lane_u16(c0.val[3], b10_u16.val[3], a00_u16.val[0], 1); + + // Accumulate 2: + c0.val[0] = vmlal_lane_u16(c0.val[0], b20_u16.val[0], a00_u16.val[0], 2); + c0.val[1] = vmlal_lane_u16(c0.val[1], b20_u16.val[1], a00_u16.val[0], 2); + c0.val[2] = vmlal_lane_u16(c0.val[2], b20_u16.val[2], a00_u16.val[0], 2); + c0.val[3] = vmlal_lane_u16(c0.val[3], b20_u16.val[3], a00_u16.val[0], 2); + + // Accumulate 3: + c0.val[0] = vmlal_lane_u16(c0.val[0], b30_u16.val[0], a00_u16.val[0], 3); + c0.val[1] = vmlal_lane_u16(c0.val[1], b30_u16.val[1], a00_u16.val[0], 3); + c0.val[2] = vmlal_lane_u16(c0.val[2], b30_u16.val[2], a00_u16.val[0], 3); + c0.val[3] = vmlal_lane_u16(c0.val[3], b30_u16.val[3], a00_u16.val[0], 3); + + // Accumulate 4: + c0.val[0] = vmlal_lane_u16(c0.val[0], b40_u16.val[0], a00_u16.val[1], 0); + c0.val[1] = vmlal_lane_u16(c0.val[1], b40_u16.val[1], a00_u16.val[1], 0); + c0.val[2] = vmlal_lane_u16(c0.val[2], b40_u16.val[2], a00_u16.val[1], 0); + c0.val[3] = vmlal_lane_u16(c0.val[3], b40_u16.val[3], a00_u16.val[1], 0); + + // Accumulate 5: + c0.val[0] = vmlal_lane_u16(c0.val[0], b50_u16.val[0], a00_u16.val[1], 1); + c0.val[1] = vmlal_lane_u16(c0.val[1], b50_u16.val[1], a00_u16.val[1], 1); + c0.val[2] = vmlal_lane_u16(c0.val[2], b50_u16.val[2], a00_u16.val[1], 1); + c0.val[3] = vmlal_lane_u16(c0.val[3], b50_u16.val[3], a00_u16.val[1], 1); + + // Accumulate 6: + c0.val[0] = vmlal_lane_u16(c0.val[0], b60_u16.val[0], a00_u16.val[1], 2); + c0.val[1] = vmlal_lane_u16(c0.val[1], b60_u16.val[1], a00_u16.val[1], 2); + c0.val[2] = vmlal_lane_u16(c0.val[2], b60_u16.val[2], a00_u16.val[1], 2); + c0.val[3] = vmlal_lane_u16(c0.val[3], b60_u16.val[3], a00_u16.val[1], 2); + + // Accumulate 7: + c0.val[0] = vmlal_lane_u16(c0.val[0], b70_u16.val[0], a00_u16.val[1], 3); + c0.val[1] = vmlal_lane_u16(c0.val[1], b70_u16.val[1], a00_u16.val[1], 3); + c0.val[2] = vmlal_lane_u16(c0.val[2], b70_u16.val[2], a00_u16.val[1], 3); + c0.val[3] = vmlal_lane_u16(c0.val[3], b70_u16.val[3], a00_u16.val[1], 3); + + vec_a += 8; + matrix_b += 8 * stride_b; + } - TensorShape in1_shape = input1->info()->tensor_shape(); - in1_shape.collapse(2); + // This for loop performs the left-over accumulations + for(; vec_a < vec_a_end_addr;) + { + const uint8x8_t a00_u8 = vld1_dup_u8(vec_a); + const uint8x16_t b00_u8 = vld1q_u8(matrix_b); - _input0 = input0; - _input1 = input1; - _output = output; - _slide_matrix_b = in1_shape[2] != 1; + const uint16x4x4_t b00_u16 = + { + { + vget_low_u16(vmovl_u8(vget_low_u8(b00_u8))), + vget_high_u16(vmovl_u8(vget_low_u8(b00_u8))), + vget_low_u16(vmovl_u8(vget_high_u8(b00_u8))), + vget_high_u16(vmovl_u8(vget_high_u8(b00_u8))) + } + }; - // Configure kernel window - auto win_config = validate_and_configure_window(input0->info(), input1->info(), output->info()); - ARM_COMPUTE_ERROR_THROW_ON(win_config.first); - INEKernel::configure(win_config.second); + // Convert a00_u8 to uint16_t and get the lower part + const uint16x4_t a00_u16 = vget_low_u16(vmovl_u8(a00_u8)); + + // Accumulate 0: + c0.val[0] = vmlal_lane_u16(c0.val[0], b00_u16.val[0], a00_u16, 0); + c0.val[1] = vmlal_lane_u16(c0.val[1], b00_u16.val[1], a00_u16, 0); + c0.val[2] = vmlal_lane_u16(c0.val[2], b00_u16.val[2], a00_u16, 0); + c0.val[3] = vmlal_lane_u16(c0.val[3], b00_u16.val[3], a00_u16, 0); + + vec_a += 1; + matrix_b += stride_b; + } + + auto vec_out = reinterpret_cast(out.ptr()); + vst1q_s32(vec_out + 0, vreinterpretq_s32_u32(c0.val[0])); + vst1q_s32(vec_out + 4, vreinterpretq_s32_u32(c0.val[1])); + vst1q_s32(vec_out + 8, vreinterpretq_s32_u32(c0.val[2])); + vst1q_s32(vec_out + 12, vreinterpretq_s32_u32(c0.val[3])); + }, + ina, inb, out); } -Error NEGEMMLowpMatrixMultiplyKernel::validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output) +void inline vector_matrix_multiply_s8(Iterator &ina, Iterator &inb, Iterator &out, int width_a, int width_b, size_t stride_b, const Window &window) { - ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input0, input1, output)); - ARM_COMPUTE_RETURN_ON_ERROR(validate_and_configure_window(input0->clone().get(), input1->clone().get(), output->clone().get()).first); + execute_window_loop(window, [&](const Coordinates & id) + { + if(id.x() > width_b) + { + return; + } - return Error{}; + // Accumulators for the block 0 + int32x4x4_t c0 = + { + { + vdupq_n_s32(0), + vdupq_n_s32(0), + vdupq_n_s32(0), + vdupq_n_s32(0) + } + }; + + auto vec_a = reinterpret_cast(ina.ptr()); + auto matrix_b = reinterpret_cast(inb.ptr()); + auto vec_a_end_addr = vec_a + width_a; + + // This for loop performs 8 accumulations + for(; vec_a <= (vec_a_end_addr - 8);) + { + const int8x8_t a00_s8 = vld1_s8(vec_a); + const int8x16_t b00_s8 = vld1q_s8(matrix_b + 0 * stride_b); + const int8x16_t b10_s8 = vld1q_s8(matrix_b + 1 * stride_b); + const int8x16_t b20_s8 = vld1q_s8(matrix_b + 2 * stride_b); + const int8x16_t b30_s8 = vld1q_s8(matrix_b + 3 * stride_b); + const int8x16_t b40_s8 = vld1q_s8(matrix_b + 4 * stride_b); + const int8x16_t b50_s8 = vld1q_s8(matrix_b + 5 * stride_b); + const int8x16_t b60_s8 = vld1q_s8(matrix_b + 6 * stride_b); + const int8x16_t b70_s8 = vld1q_s8(matrix_b + 7 * stride_b); + + // Convert a00_s8 to int16_t and get the lower part + const int16x4x2_t a00_s16 = + { + { + vget_low_s16(vmovl_s8(a00_s8)), + vget_high_s16(vmovl_s8(a00_s8)) + } + }; + + const int16x4x4_t b00_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b00_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b00_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b00_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b00_s8))) + } + }; + + const int16x4x4_t b10_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b10_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b10_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b10_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b10_s8))) + } + }; + + const int16x4x4_t b20_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b20_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b20_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b20_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b20_s8))) + } + }; + + const int16x4x4_t b30_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b30_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b30_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b30_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b30_s8))) + } + }; + + const int16x4x4_t b40_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b40_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b40_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b40_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b40_s8))) + } + }; + + const int16x4x4_t b50_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b50_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b50_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b50_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b50_s8))) + } + }; + + const int16x4x4_t b60_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b60_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b60_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b60_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b60_s8))) + } + }; + + const int16x4x4_t b70_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b70_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b70_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b70_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b70_s8))) + } + }; + + // Accumulate 0: + c0.val[0] = vmlal_lane_s16(c0.val[0], b00_s16.val[0], a00_s16.val[0], 0); + c0.val[1] = vmlal_lane_s16(c0.val[1], b00_s16.val[1], a00_s16.val[0], 0); + c0.val[2] = vmlal_lane_s16(c0.val[2], b00_s16.val[2], a00_s16.val[0], 0); + c0.val[3] = vmlal_lane_s16(c0.val[3], b00_s16.val[3], a00_s16.val[0], 0); + + // Accumulate 1: + c0.val[0] = vmlal_lane_s16(c0.val[0], b10_s16.val[0], a00_s16.val[0], 1); + c0.val[1] = vmlal_lane_s16(c0.val[1], b10_s16.val[1], a00_s16.val[0], 1); + c0.val[2] = vmlal_lane_s16(c0.val[2], b10_s16.val[2], a00_s16.val[0], 1); + c0.val[3] = vmlal_lane_s16(c0.val[3], b10_s16.val[3], a00_s16.val[0], 1); + + // Accumulate 2: + c0.val[0] = vmlal_lane_s16(c0.val[0], b20_s16.val[0], a00_s16.val[0], 2); + c0.val[1] = vmlal_lane_s16(c0.val[1], b20_s16.val[1], a00_s16.val[0], 2); + c0.val[2] = vmlal_lane_s16(c0.val[2], b20_s16.val[2], a00_s16.val[0], 2); + c0.val[3] = vmlal_lane_s16(c0.val[3], b20_s16.val[3], a00_s16.val[0], 2); + + // Accumulate 3: + c0.val[0] = vmlal_lane_s16(c0.val[0], b30_s16.val[0], a00_s16.val[0], 3); + c0.val[1] = vmlal_lane_s16(c0.val[1], b30_s16.val[1], a00_s16.val[0], 3); + c0.val[2] = vmlal_lane_s16(c0.val[2], b30_s16.val[2], a00_s16.val[0], 3); + c0.val[3] = vmlal_lane_s16(c0.val[3], b30_s16.val[3], a00_s16.val[0], 3); + + // Accumulate 4: + c0.val[0] = vmlal_lane_s16(c0.val[0], b40_s16.val[0], a00_s16.val[1], 0); + c0.val[1] = vmlal_lane_s16(c0.val[1], b40_s16.val[1], a00_s16.val[1], 0); + c0.val[2] = vmlal_lane_s16(c0.val[2], b40_s16.val[2], a00_s16.val[1], 0); + c0.val[3] = vmlal_lane_s16(c0.val[3], b40_s16.val[3], a00_s16.val[1], 0); + + // Accumulate 5: + c0.val[0] = vmlal_lane_s16(c0.val[0], b50_s16.val[0], a00_s16.val[1], 1); + c0.val[1] = vmlal_lane_s16(c0.val[1], b50_s16.val[1], a00_s16.val[1], 1); + c0.val[2] = vmlal_lane_s16(c0.val[2], b50_s16.val[2], a00_s16.val[1], 1); + c0.val[3] = vmlal_lane_s16(c0.val[3], b50_s16.val[3], a00_s16.val[1], 1); + + // Accumulate 6: + c0.val[0] = vmlal_lane_s16(c0.val[0], b60_s16.val[0], a00_s16.val[1], 2); + c0.val[1] = vmlal_lane_s16(c0.val[1], b60_s16.val[1], a00_s16.val[1], 2); + c0.val[2] = vmlal_lane_s16(c0.val[2], b60_s16.val[2], a00_s16.val[1], 2); + c0.val[3] = vmlal_lane_s16(c0.val[3], b60_s16.val[3], a00_s16.val[1], 2); + + // Accumulate 7: + c0.val[0] = vmlal_lane_s16(c0.val[0], b70_s16.val[0], a00_s16.val[1], 3); + c0.val[1] = vmlal_lane_s16(c0.val[1], b70_s16.val[1], a00_s16.val[1], 3); + c0.val[2] = vmlal_lane_s16(c0.val[2], b70_s16.val[2], a00_s16.val[1], 3); + c0.val[3] = vmlal_lane_s16(c0.val[3], b70_s16.val[3], a00_s16.val[1], 3); + + vec_a += 8; + matrix_b += 8 * stride_b; + } + + // This for loop performs the left-over accumulations + for(; vec_a < vec_a_end_addr;) + { + const int8x8_t a00_s8 = vld1_dup_s8(vec_a); + const int8x16_t b00_s8 = vld1q_s8(matrix_b); + + const int16x4x4_t b00_s16 = + { + { + vget_low_s16(vmovl_s8(vget_low_s8(b00_s8))), + vget_high_s16(vmovl_s8(vget_low_s8(b00_s8))), + vget_low_s16(vmovl_s8(vget_high_s8(b00_s8))), + vget_high_s16(vmovl_s8(vget_high_s8(b00_s8))) + } + }; + + // Convert a00_s8 to uint16_t and get the lower part + const int16x4_t a00_s16 = vget_low_s16(vmovl_s8(a00_s8)); + + // Accumulate 0: + c0.val[0] = vmlal_lane_s16(c0.val[0], b00_s16.val[0], a00_s16, 0); + c0.val[1] = vmlal_lane_s16(c0.val[1], b00_s16.val[1], a00_s16, 0); + c0.val[2] = vmlal_lane_s16(c0.val[2], b00_s16.val[2], a00_s16, 0); + c0.val[3] = vmlal_lane_s16(c0.val[3], b00_s16.val[3], a00_s16, 0); + + vec_a += 1; + matrix_b += stride_b; + } + + auto vec_out = reinterpret_cast(out.ptr()); + vst1q_s32(vec_out + 0, c0.val[0]); + vst1q_s32(vec_out + 4, c0.val[1]); + vst1q_s32(vec_out + 8, c0.val[2]); + vst1q_s32(vec_out + 12, c0.val[3]); + }, + ina, inb, out); } void inline matrix_multiply_u8(Iterator &ina, Iterator &inb, Iterator &out, int width_b, size_t out_stride, const Window &window) @@ -176,7 +534,7 @@ void inline matrix_multiply_u8(Iterator &ina, Iterator &inb, Iterator &out, int const uint8x8_t a00_u8 = vld1_u8(mtx_a0); const uint8x16_t b00_u8 = vld1q_u8(mtx_b0); - // Convert a00_s8 to uint16_t and get the lower part + // Convert a00_u8 to uint16_t and get the lower part const uint16x4_t a00_u16 = vget_low_u16(vmovl_u8(a00_u8)); // Convert b00_s8 to uint16_t @@ -355,55 +713,222 @@ void inline matrix_multiply_s8(Iterator &ina, Iterator &inb, Iterator &out, int }, ina, inb, out); } +} // namespace -void NEGEMMLowpMatrixMultiplyKernel::run(const Window &window, const ThreadInfo &info) +class Coordinates; +} // namespace arm_compute + +namespace { - ARM_COMPUTE_UNUSED(info); - ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); - ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window); +Error validate_arguments(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output) +{ + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input0, 1, DataType::QASYMM8, DataType::S8, DataType::U8); + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input0, input1); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(output, 1, DataType::S32); + + TensorShape in0_shape = input0->tensor_shape(); + TensorShape in1_shape = input1->tensor_shape(); + TensorShape out_shape = output->tensor_shape(); + + // Check vector-by-matrix case + if(out_shape[1] == 1) + { + ARM_COMPUTE_RETURN_ERROR_ON_MSG(in0_shape[0] != in1_shape[1], "The number of input0's columns must be equal to input1's rows"); + } + else + { + in0_shape.collapse(2); + in1_shape.collapse(2); + out_shape.collapse(2); - const size_t in_b_stride = _input1->info()->strides_in_bytes()[1]; - const size_t out_stride = _output->info()->strides_in_bytes()[1] / _output->info()->element_size(); + ARM_COMPUTE_RETURN_ERROR_ON_MSG(in0_shape[2] != out_shape[2], "Output tensor must have the same number of batches of input0 tensor"); + ARM_COMPUTE_RETURN_ERROR_ON_MSG(in1_shape[2] != 1 && in0_shape[2] != in1_shape[2], "Input1 tensor must have the same number of batches of input0 or the number of batches must be set to 1"); + } + + return Error{}; +} + +std::pair validate_and_configure_window(ITensorInfo *input0, ITensorInfo *input1, ITensorInfo *output) +{ + constexpr unsigned int num_elems_processed_per_iteration_x = 16; + constexpr unsigned int num_elems_processed_per_iteration_y = 4; + + Window win; + bool window_changed = false; - // Set step_x and step_y for matrix A. Scale by a factor of 4 the Y range as the input interleaved matrix A has 4 times less the rows of the output matrix - Window win_a(window); - win_a.set(Window::DimX, Window::Dimension(0, 0, 0)); - win_a.set(Window::DimY, Window::Dimension(window.y().start() / 4, window.y().end() / 4, 1)); + // Check if the output tensor is a vector. If so,the kernel runs the vector-matrix multiplication + if((output->dimension(1) == 1)) + { + // Configure kernel window + win = calculate_max_window(*output, Steps(num_elems_processed_per_iteration_x)); + + // We cannot read out-of-bound elements from matrix A as we use the left-over for loop + AccessWindowStatic in0_access(input0, 0, 0, input0->tensor_shape().x(), 1); + AccessWindowHorizontal in1_access(input1, 0, num_elems_processed_per_iteration_x); + AccessWindowHorizontal output_access(output, 0, num_elems_processed_per_iteration_x); + + window_changed = update_window_and_padding(win, in0_access, in1_access, output_access); - // Set step_x and step_y for matrix B. Scale by a factor of 16 the X range as the input transposed matrix A has 16 times less the columns of the output matrix - Window win_b; - // Don't slice matrix B along the z dimension if matrix B has just 2 dimensions and matrix A more than 2 - // This scenario can happen when the the matrix multiplication is used to perform a convolution operation - if(_slide_matrix_b) + Coordinates coord; + coord.set_num_dimensions(output->num_dimensions()); + output_access.set_valid_region(win, ValidRegion(coord, output->tensor_shape())); + } + else { - win_b = window; + win = calculate_max_window(*output, Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y)); + + AccessWindowStatic in0_access(input0, 0, 0, ceil_to_multiple(input0->dimension(0), 8), input0->dimension(1)); + AccessWindowHorizontal in1_access(input1, 0, num_elems_processed_per_iteration_x); + AccessWindowRectangle output_access(output, 0, 0, num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y); + + window_changed = update_window_and_padding(win, in0_access, in1_access, output_access); + + output_access.set_valid_region(win, ValidRegion(Coordinates(0, 0), output->tensor_shape())); } - win_b.set(Window::DimX, Window::Dimension(window.x().start() / 16, window.x().end() / 16, in_b_stride)); - win_b.set(Window::DimY, Window::Dimension(0, 0, 0)); - // The step x and step y for the output matrix has been already set using in configure() - Iterator ina(_input0, win_a); - Iterator inb(_input1, win_b); - Iterator out(_output, window); + Error err = (window_changed) ? ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Insufficient Padding!") : Error{}; + return std::make_pair(err, win); +} +} // namespace + +NEGEMMLowpMatrixMultiplyKernel::NEGEMMLowpMatrixMultiplyKernel() + : _input0(nullptr), _input1(nullptr), _output(nullptr), _slide_matrix_b(true) +{ +} + +void NEGEMMLowpMatrixMultiplyKernel::configure(const ITensor *input0, const ITensor *input1, ITensor *output) +{ + ARM_COMPUTE_ERROR_ON_NULLPTR(input0, input1, output); + ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input0->info(), input1->info(), output->info())); + + TensorShape in1_shape = input1->info()->tensor_shape(); + in1_shape.collapse(2); + + _input0 = input0; + _input1 = input1; + _output = output; + _slide_matrix_b = in1_shape[2] != 1; - const int width_b = _input1->info()->dimension(0); - switch(_input0->info()->data_type()) + // Configure kernel window + auto win_config = validate_and_configure_window(input0->info(), input1->info(), output->info()); + ARM_COMPUTE_ERROR_THROW_ON(win_config.first); + INEKernel::configure(win_config.second); +} + +Error NEGEMMLowpMatrixMultiplyKernel::validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output) +{ + ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input0, input1, output)); + ARM_COMPUTE_RETURN_ON_ERROR(validate_and_configure_window(input0->clone().get(), input1->clone().get(), output->clone().get()).first); + + return Error{}; +} + +void NEGEMMLowpMatrixMultiplyKernel::run(const Window &window, const ThreadInfo &info) +{ + ARM_COMPUTE_UNUSED(info); + ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); + ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window); + + // Check if the output tensor is a vector. If so,the kernel runs the vector-matrix multiplication path + if((_output->info()->dimension(1) == 1)) { - case DataType::S8: + const auto width_matrix_a = static_cast(_input0->info()->dimension(0)); + const auto width_matrix_b = static_cast(_input1->info()->dimension(0)); + const auto in_b_stride = static_cast(_input1->info()->strides_in_bytes()[1] / data_size_from_type(_input1->info()->data_type())); + + // The implementation computes 16 elements per iteration + const int window_start_x = 16 * info.thread_id; + const int window_step_x = 16 * info.num_threads; + // Make sure (window_end_x - window_start_x) is a multiple of window_step_x + const int window_end_x = ceil_to_multiple(width_matrix_b - window_start_x, window_step_x) + window_start_x; + + Window win_out(window); + win_out.set(Window::DimX, Window::Dimension(window_start_x, window_end_x, window_step_x)); + win_out.set(Window::DimY, Window::Dimension(0, 1, 1)); + + Window win_a(window); + win_a.set(Window::DimX, Window::Dimension(0, 0, 0)); + win_a.set(Window::DimY, Window::Dimension(0, 0, 0)); + + Window win_b; + // Don't slice matrix B along the z dimension if matrix B has just 2 dimensions and matrix A more than 2 + // This scenario can happen when the the matrix multiplication is used to perform a convolution operation + if(_input1->info()->num_dimensions() >= 3) { - matrix_multiply_s8(ina, inb, out, width_b, out_stride, window); - break; + win_b = window; } - case DataType::U8: - case DataType::QASYMM8: + win_b.set(Window::DimX, Window::Dimension(window_start_x, window_end_x, window_step_x)); + win_b.set(Window::DimY, Window::Dimension(0, 1, 1)); + + Iterator ina(_input0, win_a); + Iterator inb(_input1, win_b); + Iterator out(_output, win_out); + + switch(_input0->info()->data_type()) { - matrix_multiply_u8(ina, inb, out, width_b, out_stride, window); - break; + case DataType::S8: + { + vector_matrix_multiply_s8(ina, inb, out, width_matrix_a, width_matrix_b, in_b_stride, window); + break; + } + case DataType::U8: + case DataType::QASYMM8: + { + vector_matrix_multiply_u8(ina, inb, out, width_matrix_a, width_matrix_b, in_b_stride, window); + break; + } + default: + { + ARM_COMPUTE_ERROR("Not supported"); + break; + } } - default: + } + else + { + const size_t in_b_stride = _input1->info()->strides_in_bytes()[1]; + const size_t out_stride = _output->info()->strides_in_bytes()[1] / _output->info()->element_size(); + + // Set step_x and step_y for matrix A. Scale by a factor of 4 the Y range as the input interleaved matrix A has 4 times less the rows of the output matrix + Window win_a(window); + win_a.set(Window::DimX, Window::Dimension(0, 0, 0)); + win_a.set(Window::DimY, Window::Dimension(window.y().start() / 4, window.y().end() / 4, 1)); + + // Set step_x and step_y for matrix B. Scale by a factor of 16 the X range as the input transposed matrix A has 16 times less the columns of the output matrix + Window win_b; + // Don't slice matrix B along the z dimension if matrix B has just 2 dimensions and matrix A more than 2 + // This scenario can happen when the the matrix multiplication is used to perform a convolution operation + if(_slide_matrix_b) { - ARM_COMPUTE_ERROR("Not supported"); - break; + win_b = window; + } + win_b.set(Window::DimX, Window::Dimension(window.x().start() / 16, window.x().end() / 16, in_b_stride)); + win_b.set(Window::DimY, Window::Dimension(0, 0, 0)); + + // The step x and step y for the output matrix has been already set using in configure() + Iterator ina(_input0, win_a); + Iterator inb(_input1, win_b); + Iterator out(_output, window); + + const int width_b = _input1->info()->dimension(0); + switch(_input0->info()->data_type()) + { + case DataType::S8: + { + matrix_multiply_s8(ina, inb, out, width_b, out_stride, window); + break; + } + case DataType::U8: + case DataType::QASYMM8: + { + matrix_multiply_u8(ina, inb, out, width_b, out_stride, window); + break; + } + default: + { + ARM_COMPUTE_ERROR("Not supported"); + break; + } } } } -- cgit v1.2.1