COMPMID-481: Add AArch64 GEMM
Change-Id: I34f94f99cb05f0eabafee13c5e623ee779b72360
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83741
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
diff --git a/src/runtime/NEON/functions/NEConvolutionLayer.cpp b/src/runtime/NEON/functions/NEConvolutionLayer.cpp
index 0466a4a..44bf2de 100644
--- a/src/runtime/NEON/functions/NEConvolutionLayer.cpp
+++ b/src/runtime/NEON/functions/NEConvolutionLayer.cpp
@@ -23,17 +23,25 @@
*/
#include "arm_compute/runtime/NEON/functions/NEConvolutionLayer.h"
+#include "arm_compute/core/NEON/kernels/arm64/NEGEMMAArch64Kernel.h"
#include "arm_compute/core/PixelValue.h"
#include "arm_compute/core/Size2D.h"
#include "arm_compute/core/Utils.h"
#include "arm_compute/core/Validate.h"
#include "arm_compute/runtime/NEON/NEScheduler.h"
+#include "support/ToolchainSupport.h"
+
+namespace arm_compute
+{
+#include "arm_compute/core/NEON/kernels/assembly/gemm_interleaved.hpp"
+#include "arm_compute/core/NEON/kernels/assembly/kernels/a64_sgemm_12x8.hpp"
+} // namespace arm_compute
#include <cmath>
#include <tuple>
-using namespace arm_compute;
-
+namespace arm_compute
+{
NEConvolutionLayerReshapeWeights::NEConvolutionLayerReshapeWeights(std::shared_ptr<IMemoryManager> memory_manager)
: _memory_group(std::move(memory_manager)), _weights_reshape_kernel(), _weights_transposed_kernel(), _weights_reshaped(), _transpose1xW(false)
{
@@ -69,8 +77,10 @@
_weights_reshaped.allocator()->init(info_wr);
_memory_group.manage(&_weights_reshaped);
+
_weights_reshape_kernel.configure(weights, biases, &_weights_reshaped);
_weights_transposed_kernel.configure(&_weights_reshaped, output);
+
_weights_reshaped.allocator()->allocate();
}
else
@@ -84,6 +94,7 @@
_memory_group.acquire();
NEScheduler::get().schedule(&_weights_reshape_kernel, 3);
+
if(_transpose1xW)
{
NEScheduler::get().schedule(&_weights_transposed_kernel, Window::DimY);
@@ -93,8 +104,8 @@
}
NEConvolutionLayer::NEConvolutionLayer(std::shared_ptr<IMemoryManager> memory_manager)
- : _memory_group(std::move(memory_manager)), _input_im2col_kernel(), _input_interleave_kernel(), _reshape_weights(), _mm_kernel(), _output_col2im_kernel(), _input_im2col_reshaped(),
- _input_interleaved_reshaped(), _weights_reshaped(), _gemm_output(), _has_bias(false), _is_fully_connected_convolution(false), _are_weights_reshaped(false)
+ : _memory_group(std::move(memory_manager)), _input_im2col_kernel(), _input_interleave_kernel(), _reshape_weights(), _mm_kernel(), _mm_optimised_kernel(nullptr), _output_col2im_kernel(),
+ _input_im2col_reshaped(), _input_interleaved_reshaped(), _weights_reshaped(), _gemm_output(), _workspace(), _has_bias(false), _is_fully_connected_convolution(false), _are_weights_reshaped(false)
{
}
@@ -137,45 +148,72 @@
std::tie(conv_w, conv_h) = scaled_dimensions(input->info()->dimension(0), input->info()->dimension(1), kernel_width, kernel_height,
conv_info);
- // Check if its a "fully connected" convolution
+ // Check if its a "fully connected" convolution, i.e. the output size is 1x1xnum_kernels
_is_fully_connected_convolution = ((conv_w == 1) && (conv_h == 1));
+#if defined(__aarch64__)
+ if(NEScheduler::get().cpu_info().CPU >= CPUTarget::ARMV8 && dt == DataType::F32)
+ {
+ _mm_optimised_kernel = support::cpp14::make_unique<NEGEMMAArch64Kernel>();
+ }
+#endif /* defined(__aarch64__) */
+
unsigned int mat_weights_cols = weights->info()->dimension(3);
unsigned int mat_weights_rows = weights->info()->dimension(0) * weights->info()->dimension(1) * weights->info()->dimension(2) + (_has_bias ? 1 : 0);
// Reshape weights if needed
- if(_are_weights_reshaped)
+ if(_mm_optimised_kernel != nullptr)
{
- mat_weights_cols = weights_info.num_kernels();
- const unsigned int quarter_reshaped_cols = weights->info()->dimension(0) / 4;
- mat_weights_rows = (_has_bias ? 1 + quarter_reshaped_cols : quarter_reshaped_cols);
- }
- else
- {
- if(_is_fully_connected_convolution)
+ if(_are_weights_reshaped)
{
- // Create tensor to store the reshaped weights
- TensorShape shape_wr(mat_weights_cols, mat_weights_rows);
- TensorInfo info_wr(shape_wr, 1, dt, fixed_point_position);
- _weights_reshaped.allocator()->init(info_wr);
- _reshape_weights.configure(weights, biases, &_weights_reshaped, false /* 1xW transpose */);
+ mat_weights_cols = weights_info.num_kernels();
+ mat_weights_rows = weights->info()->dimension(1);
}
else
{
- // Create tensor to store transposed weights
- const float transpose_width = 16.0f / input->info()->element_size();
- TensorShape shape_wt(mat_weights_rows * static_cast<unsigned int>(transpose_width), static_cast<unsigned int>(std::ceil(mat_weights_cols / transpose_width)));
- TensorInfo info_wt(shape_wt, 1, dt, fixed_point_position);
- _weights_reshaped.allocator()->init(info_wt);
- _reshape_weights.configure(weights, biases, &_weights_reshaped, true /* 1xW transpose */);
+ TensorShape reshaped_weights_shape{ mat_weights_cols, mat_weights_rows };
+
+ // Create tensor to store the reshaped weights
+ _weights_reshaped.allocator()->init(TensorInfo(reshaped_weights_shape, 1, dt, fixed_point_position));
+ _reshape_weights.configure(weights, biases, &_weights_reshaped, false /* 1xW transpose */);
+ weights = &_weights_reshaped;
}
- weights = &_weights_reshaped;
+ }
+ else
+ {
+ if(_are_weights_reshaped)
+ {
+ mat_weights_cols = weights_info.num_kernels();
+ mat_weights_rows = weights->info()->dimension(0) / 4 + (_has_bias ? 1 : 0);
+ }
+ else
+ {
+ TensorShape reshaped_weights_shape;
+
+ if(_is_fully_connected_convolution)
+ {
+ reshaped_weights_shape = TensorShape{ mat_weights_cols, mat_weights_rows };
+ }
+ else
+ {
+ // Create tensor to store transposed weights
+ const float transpose_width = 16.0f / input->info()->element_size();
+ reshaped_weights_shape = TensorShape{ mat_weights_rows *static_cast<unsigned int>(transpose_width),
+ static_cast<unsigned int>(std::ceil(mat_weights_cols / transpose_width)) };
+ }
+
+ // Create tensor to store the reshaped weights
+ _weights_reshaped.allocator()->init(TensorInfo(reshaped_weights_shape, 1, dt, fixed_point_position));
+ _reshape_weights.configure(weights, biases, &_weights_reshaped, !_is_fully_connected_convolution /* 1xW transpose */);
+ weights = &_weights_reshaped;
+ }
}
// Create tensor to store im2col reshaped inputs
const unsigned int mat_input_cols = mat_weights_rows;
const unsigned int mat_input_rows = conv_w * conv_h;
- TensorShape shape_im2col = input->info()->tensor_shape();
+
+ TensorShape shape_im2col(input->info()->tensor_shape());
shape_im2col.set(0, mat_input_cols);
shape_im2col.set(1, mat_input_rows);
shape_im2col.set(2, 1);
@@ -185,7 +223,7 @@
// Create tensor (interleave) to prepare input tensor for GEMM
if(!_is_fully_connected_convolution)
{
- TensorShape shape_interleaved = shape_im2col;
+ TensorShape shape_interleaved(shape_im2col);
shape_interleaved.set(0, shape_interleaved.x() * 4);
shape_interleaved.set(1, std::ceil(shape_interleaved.y() / 4.f));
_input_interleaved_reshaped.allocator()->init(TensorInfo(shape_interleaved, 1, dt, fixed_point_position));
@@ -193,7 +231,7 @@
}
// Create GEMM output tensor
- TensorShape shape_gemm = _input_im2col_reshaped.info()->tensor_shape();
+ TensorShape shape_gemm(_input_im2col_reshaped.info()->tensor_shape());
shape_gemm.set(0, mat_weights_cols);
shape_gemm.set(1, mat_input_rows);
_gemm_output.allocator()->init(TensorInfo(shape_gemm, 1, dt, fixed_point_position));
@@ -201,16 +239,49 @@
// Configure kernels
_input_im2col_kernel.configure(input, &_input_im2col_reshaped, Size2D(kernel_width, kernel_height), conv_info, _has_bias);
- if(_is_fully_connected_convolution)
+
+#if defined(__aarch64__)
+ if(_mm_optimised_kernel != nullptr)
{
- _mm_kernel.configure(&_input_im2col_reshaped, weights, &_gemm_output, 1.0f);
+ struct CPUInfo ci = NEScheduler::get().cpu_info();
+
+ const int M = _gemm_output.info()->tensor_shape().y();
+ const int N = _gemm_output.info()->tensor_shape().x();
+ const int K = _input_im2col_reshaped.info()->tensor_shape().x();
+
+ GemmInterleaved<sgemm_12x8, float, float> gemm(&ci, M, N, K, false, false);
+
+ constexpr size_t alignment = 4096;
+ _workspace.allocator()->init(TensorInfo(TensorShape{ (gemm.get_working_size() + alignment - 1) * NEScheduler::get().num_threads() }, 1, DataType::U8));
+ _memory_group.manage(&_workspace);
+
+ // Configure matrix multiplication kernel
+ if(_is_fully_connected_convolution)
+ {
+ _mm_optimised_kernel->configure(&_input_im2col_reshaped, weights, &_gemm_output, &_workspace, 1.f, 0.f, false, false);
+ }
+ else
+ {
+ _mm_optimised_kernel->configure(&_input_im2col_reshaped, weights, &_gemm_output, &_workspace);
+ }
+
+ _workspace.allocator()->allocate();
}
else
+#endif /* defined(__aarch64__) */
{
- _input_interleave_kernel.configure(&_input_im2col_reshaped, &_input_interleaved_reshaped);
- _mm_kernel.configure(&_input_interleaved_reshaped, weights, &_gemm_output, 1.0f);
- _input_interleaved_reshaped.allocator()->allocate();
+ if(_is_fully_connected_convolution)
+ {
+ _mm_kernel.configure(&_input_im2col_reshaped, weights, &_gemm_output, 1.0f);
+ }
+ else
+ {
+ _input_interleave_kernel.configure(&_input_im2col_reshaped, &_input_interleaved_reshaped);
+ _mm_kernel.configure(&_input_interleaved_reshaped, weights, &_gemm_output, 1.0f);
+ _input_interleaved_reshaped.allocator()->allocate();
+ }
}
+
_input_im2col_reshaped.allocator()->allocate();
_output_col2im_kernel.configure(&_gemm_output, output, std::make_pair(conv_w, conv_h));
_gemm_output.allocator()->allocate();
@@ -237,17 +308,26 @@
// Run input reshaping
NEScheduler::get().schedule(&_input_im2col_kernel, Window::DimY);
- if(!_is_fully_connected_convolution)
- {
- // Run interleave
- NEScheduler::get().schedule(&_input_interleave_kernel, Window::DimY);
- }
// Runs matrix multiply on reshaped matrices
- NEScheduler::get().schedule(&_mm_kernel, Window::DimY);
+ if(_mm_optimised_kernel != nullptr)
+ {
+ NEScheduler::get().schedule(_mm_optimised_kernel.get(), Window::DimY);
+ }
+ else
+ {
+ if(!_is_fully_connected_convolution)
+ {
+ // Run interleave
+ NEScheduler::get().schedule(&_input_interleave_kernel, Window::DimY);
+ }
+
+ NEScheduler::get().schedule(&_mm_kernel, Window::DimY);
+ }
// Reshape output matrix
NEScheduler::get().schedule(&_output_col2im_kernel, Window::DimY);
_memory_group.release();
}
+} // namespace arm_compute
diff --git a/src/runtime/NEON/functions/NEGEMM.cpp b/src/runtime/NEON/functions/NEGEMM.cpp
index 85b283c..1d6aa65 100644
--- a/src/runtime/NEON/functions/NEGEMM.cpp
+++ b/src/runtime/NEON/functions/NEGEMM.cpp
@@ -26,18 +26,27 @@
#include "arm_compute/core/Error.h"
#include "arm_compute/core/Helpers.h"
#include "arm_compute/core/ITensor.h"
+#include "arm_compute/core/NEON/kernels/arm64/NEGEMMAArch64Kernel.h"
#include "arm_compute/core/TensorInfo.h"
#include "arm_compute/core/Types.h"
#include "arm_compute/core/Validate.h"
#include "arm_compute/runtime/NEON/NEScheduler.h"
#include "arm_compute/runtime/TensorAllocator.h"
+#include "support/ToolchainSupport.h"
+
+namespace arm_compute
+{
+#include "arm_compute/core/NEON/kernels/assembly/gemm_interleaved.hpp"
+#include "arm_compute/core/NEON/kernels/assembly/kernels/a64_sgemm_12x8.hpp"
+} // namespace arm_compute
#include <cmath>
-using namespace arm_compute;
-
+namespace arm_compute
+{
NEGEMM::NEGEMM(std::shared_ptr<IMemoryManager> memory_manager)
- : _memory_group(std::move(memory_manager)), _interleave_kernel(), _transpose_kernel(), _mm_kernel(), _ma_kernel(), _tmp_a(), _tmp_b(), _run_vector_matrix_multiplication(false), _run_addition(false)
+ : _memory_group(std::move(memory_manager)), _interleave_kernel(), _transpose_kernel(), _mm_kernel(), _mm_optimised_kernel(nullptr), _ma_kernel(), _tmp_a(), _tmp_b(), _workspace(),
+ _run_vector_matrix_multiplication(false), _run_addition(false)
{
}
@@ -57,57 +66,94 @@
ARM_COMPUTE_ERROR_ON_MSG(c->info()->dimension(1) != d->info()->dimension(1), "The C matrix must have the same number of columns as the output matrix");
}
- // Check if the first input tensor is a vector. If so, all the kernels for reshaping the tensors can be skipped
- if((a->info()->dimension(1) == 1))
- {
- _run_vector_matrix_multiplication = true;
+ _run_vector_matrix_multiplication = a->info()->dimension(1) < 2;
+#if defined(__aarch64__)
+ if(NEScheduler::get().cpu_info().CPU >= CPUTarget::ARMV8 && a->info()->data_type() == DataType::F32 && (c == nullptr || beta == 0.f))
+ {
+ _mm_optimised_kernel = support::cpp14::make_unique<NEGEMMAArch64Kernel>();
+ }
+#endif /* defined(__aarch64__) */
+
+ // Check if the first input tensor is a vector.
+ // If so, all the kernels for reshaping the tensors can be skipped
+ if(_run_vector_matrix_multiplication)
+ {
// Configure the matrix multiply kernel
_mm_kernel.configure(a, b, d, alpha);
+
+ // Configure matrix addition kernel
+ if(beta != 0 && c != nullptr)
+ {
+ _ma_kernel.configure(c, d, beta);
+ _run_addition = true;
+ }
}
else
{
- _run_vector_matrix_multiplication = false;
+#if defined(__aarch64__)
+ if(_mm_optimised_kernel != nullptr)
+ {
+ struct CPUInfo ci = NEScheduler::get().cpu_info();
- TensorShape shape_tmp_a = a->info()->tensor_shape();
- TensorShape shape_tmp_b = b->info()->tensor_shape();
+ const int M = d->info()->tensor_shape().y();
+ const int N = d->info()->tensor_shape().x();
+ const int K = a->info()->tensor_shape().x();
- shape_tmp_a.set(0, a->info()->dimension(0) * 4);
- shape_tmp_a.set(1, std::ceil(a->info()->dimension(1) / 4.0f));
+ GemmInterleaved<sgemm_12x8, float, float> gemm(&ci, M, N, K, false, false);
- const unsigned int transpose_w = 16 / data_size_from_type(b->info()->data_type());
- shape_tmp_b.set(0, b->info()->dimension(1) * transpose_w);
- shape_tmp_b.set(1, std::ceil(b->info()->dimension(0) / static_cast<float>(transpose_w)));
+ constexpr size_t alignment = 4096;
+ _workspace.allocator()->init(TensorInfo(TensorShape{ (gemm.get_working_size() + alignment - 1) * NEScheduler::get().num_threads() }, 1, DataType::U8));
+ _memory_group.manage(&_workspace);
- TensorInfo info_a(shape_tmp_a, 1, a->info()->data_type(), a->info()->fixed_point_position());
- TensorInfo info_b(shape_tmp_b, 1, b->info()->data_type(), a->info()->fixed_point_position());
+ // Configure matrix multiplication kernel
+ _mm_optimised_kernel->configure(a, b, d, &_workspace, alpha, 0.f);
- _tmp_a.allocator()->init(info_a);
- _tmp_b.allocator()->init(info_b);
+ _workspace.allocator()->allocate();
+ }
+ else
+#endif /* defined(__aarch64__) */
+ {
+ TensorShape shape_tmp_a = a->info()->tensor_shape();
+ TensorShape shape_tmp_b = b->info()->tensor_shape();
- // Manage intermediate buffers
- _memory_group.manage(&_tmp_a);
- _memory_group.manage(&_tmp_b);
+ shape_tmp_a.set(0, a->info()->dimension(0) * 4);
+ shape_tmp_a.set(1, std::ceil(a->info()->dimension(1) / 4.0f));
- // Configure interleave kernel
- _interleave_kernel.configure(a, &_tmp_a);
+ const unsigned int transpose_w = 16 / data_size_from_type(b->info()->data_type());
+ shape_tmp_b.set(0, b->info()->dimension(1) * transpose_w);
+ shape_tmp_b.set(1, std::ceil(b->info()->dimension(0) / static_cast<float>(transpose_w)));
- // Configure transpose kernel
- _transpose_kernel.configure(b, &_tmp_b);
+ TensorInfo info_a(shape_tmp_a, 1, a->info()->data_type(), a->info()->fixed_point_position());
+ TensorInfo info_b(shape_tmp_b, 1, b->info()->data_type(), a->info()->fixed_point_position());
- // Configure matrix multiplication kernel
- _mm_kernel.configure(&_tmp_a, &_tmp_b, d, alpha);
+ _tmp_a.allocator()->init(info_a);
+ _tmp_b.allocator()->init(info_b);
- // Allocate once the all configure methods have been called
- _tmp_a.allocator()->allocate();
- _tmp_b.allocator()->allocate();
- }
+ // Manage intermediate buffers
+ _memory_group.manage(&_tmp_a);
+ _memory_group.manage(&_tmp_b);
- // Configure matrix addition kernel
- if(beta != 0 && c != nullptr)
- {
- _ma_kernel.configure(c, d, beta);
- _run_addition = true;
+ // Configure interleave kernel
+ _interleave_kernel.configure(a, &_tmp_a);
+
+ // Configure transpose kernel
+ _transpose_kernel.configure(b, &_tmp_b);
+
+ // Configure matrix multiplication kernel
+ _mm_kernel.configure(&_tmp_a, &_tmp_b, d, alpha);
+
+ // Allocate once the all configure methods have been called
+ _tmp_a.allocator()->allocate();
+ _tmp_b.allocator()->allocate();
+
+ // Configure matrix addition kernel
+ if(beta != 0 && c != nullptr)
+ {
+ _ma_kernel.configure(c, d, beta);
+ _run_addition = true;
+ }
+ }
}
}
@@ -115,23 +161,31 @@
{
_memory_group.acquire();
- if(!_run_vector_matrix_multiplication)
+ if(_mm_optimised_kernel != nullptr)
{
- // Run interleave kernel
- NEScheduler::get().schedule(&_interleave_kernel, Window::DimY);
-
- // Run transpose kernel
- NEScheduler::get().schedule(&_transpose_kernel, Window::DimY);
+ NEScheduler::get().schedule(_mm_optimised_kernel.get(), Window::DimY);
+ _memory_group.release();
}
-
- // Run matrix multiply kernel
- NEScheduler::get().schedule(&_mm_kernel, _run_vector_matrix_multiplication ? Window::DimX : Window::DimY);
-
- _memory_group.release();
-
- // Run matrix addition kernel
- if(_run_addition)
+ else
{
- NEScheduler::get().schedule(&_ma_kernel, Window::DimY);
+ if(!_run_vector_matrix_multiplication)
+ {
+ // Run interleave kernel
+ NEScheduler::get().schedule(&_interleave_kernel, Window::DimY);
+
+ // Run transpose kernel
+ NEScheduler::get().schedule(&_transpose_kernel, Window::DimY);
+ }
+
+ NEScheduler::get().schedule(&_mm_kernel, _run_vector_matrix_multiplication ? Window::DimX : Window::DimY);
+
+ _memory_group.release();
+
+ // Run matrix addition kernel
+ if(_run_addition)
+ {
+ NEScheduler::get().schedule(&_ma_kernel, Window::DimY);
+ }
}
}
+} // namespace arm_compute