From a1f7851e2f776610019db8725c2963c36b0c85eb Mon Sep 17 00:00:00 2001 From: ramelg01 Date: Wed, 29 Jun 2022 16:28:10 +0100 Subject: Integrate new winograd APIs from MLTech Resolves: COMPMID-5400 Signed-off-by: Ramy Elgammal Change-Id: Ib4428436dd7a6e40d8b2d8a2f8dac1b079154551 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7894 Reviewed-by: Pablo Marquez Tello Tested-by: Arm Jenkins Comments-Addressed: Arm Jenkins Benchmark: Arm Jenkins --- .../weight_transforms/a64_fp16_4x4_3x3.cpp | 242 +++++++++++++ .../weight_transforms/arm_fp32_2x2_3x3.cpp | 200 +++++++++++ .../weight_transforms/arm_fp32_2x2_5x5.cpp | 381 +++++++++++++++++++++ .../weight_transforms/arm_fp32_4x4_3x3.cpp | 236 +++++++++++++ .../weight_transforms/cpp_fp32_1x2_1x7.cpp | 71 ++++ .../weight_transforms/cpp_fp32_1x4_1x5.cpp | 77 +++++ .../weight_transforms/cpp_fp32_1x6_1x3.cpp | 71 ++++ 7 files changed, 1278 insertions(+) create mode 100644 src/core/NEON/kernels/convolution/winograd/weight_transforms/a64_fp16_4x4_3x3.cpp create mode 100644 src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_2x2_3x3.cpp create mode 100644 src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_2x2_5x5.cpp create mode 100644 src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_4x4_3x3.cpp create mode 100644 src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x2_1x7.cpp create mode 100644 src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x4_1x5.cpp create mode 100644 src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x6_1x3.cpp (limited to 'src/core/NEON/kernels/convolution/winograd/weight_transforms') diff --git a/src/core/NEON/kernels/convolution/winograd/weight_transforms/a64_fp16_4x4_3x3.cpp b/src/core/NEON/kernels/convolution/winograd/weight_transforms/a64_fp16_4x4_3x3.cpp new file mode 100644 index 0000000000..0d9a65890e --- /dev/null +++ b/src/core/NEON/kernels/convolution/winograd/weight_transforms/a64_fp16_4x4_3x3.cpp @@ -0,0 +1,242 @@ +/* + * Copyright (c) 2022 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#if defined(__aarch64__) && defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) + +#include +#include + +namespace arm_conv { +namespace winograd { +namespace weight_transform { + +void a64_fp16_4x4_3x3( + unsigned int n_channels, + const __fp16* inptr, // NOTE: Data in HWIO order + const size_t ld_weight_row, + const size_t ld_weight_col, + __fp16* outptr, + const size_t matrix_stride +) +{ +#ifdef __aarch64__ + for (; n_channels >= 8; n_channels -= 8) + { + // Matrices used and computed in this kernel + float16x8_t w[3][3], Ww[6][3], V[6][6]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = vld1q_f16(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + // Ww[0][j] = 6*w[0][j]; + Ww[0][j] = vmulq_n_f16(w[0][j], 6.0); + + // Ww[1][j] = -4*w[0][j] + -4*w[1][j] + -4*w[2][j]; + Ww[1][j] = vmulq_n_f16(vaddq_f16(vaddq_f16(w[0][j], w[1][j]), w[2][j]), -4.0); + + // Ww[2][j] = -4*w[0][j] + 4*w[1][j] + -4*w[2][j]; + Ww[2][j] = vmulq_n_f16(vsubq_f16(vsubq_f16(w[1][j], w[0][j]), w[2][j]), 4.0); + + // Ww[3][j] = 1*w[0][j] + 2*w[1][j] + 4*w[2][j]; + Ww[3][j] = vaddq_f16(vaddq_f16(w[0][j], vmulq_f16(w[1][j], vdupq_n_f16(2.0f))), vmulq_f16(w[2][j], vdupq_n_f16(4.0f))); + + // Ww[4][j] = 1*w[0][j] + -2*w[1][j] + 4*w[2][j]; + Ww[4][j] = vaddq_f16(vsubq_f16(w[0][j], vmulq_f16(w[1][j], vdupq_n_f16(2.0f))), vmulq_f16(w[2][j], vdupq_n_f16(4.0f))); + + // Ww[5][j] = 24*w[2][j]; + Ww[5][j] = vmulq_n_f16(w[2][j], 24.0f); + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + const float recip576 = 1.0f / 576.0f; + + // V[i][0] = 6*Ww[i][0]; + V[i][0] = vmulq_n_f16(vmulq_n_f16(Ww[i][0], 6.0), recip576); + + // V[i][1] = -4*Ww[i][0] + -4*Ww[i][1] + -4*Ww[i][2]; + V[i][1] = vmulq_n_f16(vmulq_n_f16(vaddq_f16(vaddq_f16(Ww[i][0], Ww[i][1]), Ww[i][2]), -4.0), recip576); + + // V[i][2] = -4*Ww[i][0] + 4*Ww[i][1] + -4*Ww[i][2]; + V[i][2] = vmulq_n_f16(vmulq_n_f16(vsubq_f16(vsubq_f16(Ww[i][1], Ww[i][0]), Ww[i][2]), 4.0), recip576); + + // V[i][3] = 1*Ww[i][0] + 2*Ww[i][1] + 4*Ww[i][2]; + V[i][3] = vmulq_n_f16(vaddq_f16(vaddq_f16(Ww[i][0], vmulq_f16(Ww[i][1], vdupq_n_f16(2.0f))), vmulq_f16(Ww[i][2], vdupq_n_f16(4.0f))), recip576); + + // V[i][4] = 1*Ww[i][0] + -2*Ww[i][1] + 4*Ww[i][2]; + V[i][4] = vmulq_n_f16(vaddq_f16(vsubq_f16(Ww[i][0], vmulq_f16(Ww[i][1], vdupq_n_f16(2.0f))), vmulq_f16(Ww[i][2], vdupq_n_f16(4.0f))), recip576); + + // V[i][5] = 24*Ww[i][2]; + V[i][5] = vmulq_n_f16(vmulq_n_f16(Ww[i][2], 24.0f), recip576); + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + vst1q_f16(outptr + m*matrix_stride, V[i][j]); + } + } + inptr += 8; + outptr += 8; + } +#endif // __aarch64__ +#ifdef __arm_any__ + for (; n_channels >= 4; n_channels -= 4) + { + // Matrices used and computed in this kernel + float16x4_t w[3][3], Ww[6][3], V[6][6]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = vld1_f16(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + // Ww[0][j] = 6*w[0][j]; + Ww[0][j] = vmul_n_f16(w[0][j], 6.0); + + // Ww[1][j] = -4*w[0][j] + -4*w[1][j] + -4*w[2][j]; + Ww[1][j] = vmul_n_f16(vadd_f16(vadd_f16(w[0][j], w[1][j]), w[2][j]), -4.0); + + // Ww[2][j] = -4*w[0][j] + 4*w[1][j] + -4*w[2][j]; + Ww[2][j] = vmul_n_f16(vsub_f16(vsub_f16(w[1][j], w[0][j]), w[2][j]), 4.0); + + // Ww[3][j] = 1*w[0][j] + 2*w[1][j] + 4*w[2][j]; + Ww[3][j] = vadd_f16(vadd_f16(w[0][j], vmul_f16(w[1][j], vdup_n_f16(2.0f))), vmul_f16(w[2][j], vdup_n_f16(4.0f))); + + // Ww[4][j] = 1*w[0][j] + -2*w[1][j] + 4*w[2][j]; + Ww[4][j] = vadd_f16(vsub_f16(w[0][j], vmul_f16(w[1][j], vdup_n_f16(2.0f))), vmul_f16(w[2][j], vdup_n_f16(4.0f))); + + // Ww[5][j] = 24*w[2][j]; + Ww[5][j] = vmul_n_f16(w[2][j], 24.0f); + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + const float recip576 = 1.0f / 576.0f; + + // V[i][0] = 6*Ww[i][0]; + V[i][0] = vmul_n_f16(vmul_n_f16(Ww[i][0], 6.0), recip576); + + // V[i][1] = -4*Ww[i][0] + -4*Ww[i][1] + -4*Ww[i][2]; + V[i][1] = vmul_n_f16(vmul_n_f16(vadd_f16(vadd_f16(Ww[i][0], Ww[i][1]), Ww[i][2]), -4.0), recip576); + + // V[i][2] = -4*Ww[i][0] + 4*Ww[i][1] + -4*Ww[i][2]; + V[i][2] = vmul_n_f16(vmul_n_f16(vsub_f16(vsub_f16(Ww[i][1], Ww[i][0]), Ww[i][2]), 4.0), recip576); + + // V[i][3] = 1*Ww[i][0] + 2*Ww[i][1] + 4*Ww[i][2]; + V[i][3] = vmul_n_f16(vadd_f16(vadd_f16(Ww[i][0], vmul_f16(Ww[i][1], vdup_n_f16(2.0f))), vmul_f16(Ww[i][2], vdup_n_f16(4.0f))), recip576); + + // V[i][4] = 1*Ww[i][0] + -2*Ww[i][1] + 4*Ww[i][2]; + V[i][4] = vmul_n_f16(vadd_f16(vsub_f16(Ww[i][0], vmul_f16(Ww[i][1], vdup_n_f16(2.0f))), vmul_f16(Ww[i][2], vdup_n_f16(4.0f))), recip576); + + // V[i][5] = 24*Ww[i][2]; + V[i][5] = vmul_n_f16(vmul_n_f16(Ww[i][2], 24.0f), recip576); + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + vst1_f16(outptr + m*matrix_stride, V[i][j]); + } + } + inptr += 4; + outptr += 4; + } +#endif // __arm_any__ + for (; n_channels; n_channels--) + { + // Matrices used and computed in this kernel + __fp16 w[3][3], Ww[6][3], V[6][6]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = *(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + Ww[0][j] = 6*w[0][j]; + Ww[1][j] = -4*w[0][j] + -4*w[1][j] + -4*w[2][j]; + Ww[2][j] = -4*w[0][j] + 4*w[1][j] + -4*w[2][j]; + Ww[3][j] = 1*w[0][j] + 2*w[1][j] + 4*w[2][j]; + Ww[4][j] = 1*w[0][j] + -2*w[1][j] + 4*w[2][j]; + Ww[5][j] = 24*w[2][j]; + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + V[i][0] = ( 6*Ww[i][0]) / 576.0; + V[i][1] = (-4*Ww[i][0] + -4*Ww[i][1] + -4*Ww[i][2]) / 576.0; + V[i][2] = (-4*Ww[i][0] + 4*Ww[i][1] + -4*Ww[i][2]) / 576.0; + V[i][3] = ( 1*Ww[i][0] + 2*Ww[i][1] + 4*Ww[i][2]) / 576.0; + V[i][4] = ( 1*Ww[i][0] + -2*Ww[i][1] + 4*Ww[i][2]) / 576.0; + V[i][5] = (24*Ww[i][2]) / 576.0; + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + *(outptr + m*matrix_stride) = V[i][j]; + } + } + + inptr++; + outptr++; + } +} + +} // namespace weight_transform +} // namespace winograd +} // namespace arm_conv + +#endif // defined(__aarch64__) && defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) diff --git a/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_2x2_3x3.cpp b/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_2x2_3x3.cpp new file mode 100644 index 0000000000..e55bcb632f --- /dev/null +++ b/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_2x2_3x3.cpp @@ -0,0 +1,200 @@ +/* + * Copyright (c) 2022 ARM Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include +#include + +namespace arm_conv { +namespace winograd { +namespace weight_transform { + +void arm_fp32_2x2_3x3( + unsigned int n_channels, + const float *inptr, size_t ld_weight_row, size_t ld_weight_col, + float *outptr, size_t matrix_stride +) +{ + constexpr auto inner_tile_i = 4u; + constexpr auto inner_tile_j = 4u; + +#ifdef __aarch64__ + // For each output channel + for (; n_channels >= 4u; n_channels -= 4) + { + // Matrices used and computed in this kernel + float32x4_t w[3][3], Ww[inner_tile_i][3], V[inner_tile_i][inner_tile_j]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = vld1q_f32(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + Ww[0][j] = w[0][j]; + + // Ww[1][j] = 0.5*(w[0][j] + w[1][j] + w[2][j]); + Ww[1][j] = vmulq_n_f32(vaddq_f32(vaddq_f32(w[0][j], w[1][j]), w[2][j]), 0.5f); + + // Ww[2][j] = 0.5*(w[0][j] - w[1][j] + w[2][j]); + Ww[2][j] = vmulq_n_f32(vaddq_f32(vsubq_f32(w[0][j], w[1][j]), w[2][j]), 0.5f); + + Ww[3][j] = w[2][j]; + } + + // Compute V = W w WT + for (auto i = 0u; i < inner_tile_i; i++) + { + V[i][0] = Ww[i][0]; + + // V[i][1] = 0.5*(Ww[i][0] + Ww[i][1] + Ww[i][2]); + V[i][1] = vmulq_n_f32(vaddq_f32(vaddq_f32(Ww[i][0], Ww[i][1]), Ww[i][2]), 0.5f); + + // V[i][2] = 0.5*(Ww[i][0] - Ww[i][1] + Ww[i][2]); + V[i][2] = vmulq_n_f32(vaddq_f32(vsubq_f32(Ww[i][0], Ww[i][1]), Ww[i][2]), 0.5f); + + V[i][3] = Ww[i][2]; + } + + // Store the transformed weights + for (auto i = 0u, m = 0u; i < inner_tile_i; i++) + { + for (auto j = 0u; j < inner_tile_j; j++, m++) + { + vst1q_f32(outptr + m*matrix_stride, V[i][j]); + } + } + + inptr += 4; + outptr += 4; + } +#endif // __aarch64__ + for (; n_channels >= 2u; n_channels -= 2) + { + // Matrices used and computed in this kernel + float32x2_t w[3][3], Ww[inner_tile_i][3], V[inner_tile_i][inner_tile_j]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = vld1_f32(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + Ww[0][j] = w[0][j]; + + // Ww[1][j] = 0.5*(w[0][j] + w[1][j] + w[2][j]); + Ww[1][j] = vmul_n_f32(vadd_f32(vadd_f32(w[0][j], w[1][j]), w[2][j]), 0.5f); + + // Ww[2][j] = 0.5*(w[0][j] - w[1][j] + w[2][j]); + Ww[2][j] = vmul_n_f32(vadd_f32(vsub_f32(w[0][j], w[1][j]), w[2][j]), 0.5f); + + Ww[3][j] = w[2][j]; + } + + // Compute V = W w WT + for (auto i = 0u; i < inner_tile_i; i++) + { + V[i][0] = Ww[i][0]; + + // V[i][1] = 0.5*(Ww[i][0] + Ww[i][1] + Ww[i][2]); + V[i][1] = vmul_n_f32(vadd_f32(vadd_f32(Ww[i][0], Ww[i][1]), Ww[i][2]), 0.5f); + + // V[i][2] = 0.5*(Ww[i][0] - Ww[i][1] + Ww[i][2]); + V[i][2] = vmul_n_f32(vadd_f32(vsub_f32(Ww[i][0], Ww[i][1]), Ww[i][2]), 0.5f); + + V[i][3] = Ww[i][2]; + } + + // Store the transformed weights + for (auto i = 0u, m = 0u; i < inner_tile_i; i++) + { + for (auto j = 0u; j < inner_tile_j; j++, m++) + { + vst1_f32(outptr + m*matrix_stride, V[i][j]); + } + } + + inptr += 2; + outptr += 2; + } + for (; n_channels; n_channels--) + { + // Matrices used and computed in this kernel + float w[3][3], Ww[inner_tile_i][3], V[inner_tile_i][inner_tile_j]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = *(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + Ww[0][j] = w[0][j]; + Ww[1][j] = 0.5*(w[0][j] + w[1][j] + w[2][j]); + Ww[2][j] = 0.5*(w[0][j] - w[1][j] + w[2][j]); + Ww[3][j] = w[2][j]; + } + + // Compute V = W w WT + for (auto i = 0u; i < inner_tile_i; i++) + { + V[i][0] = Ww[i][0]; + V[i][1] = 0.5*(Ww[i][0] + Ww[i][1] + Ww[i][2]); + V[i][2] = 0.5*(Ww[i][0] - Ww[i][1] + Ww[i][2]); + V[i][3] = Ww[i][2]; + } + + // Store the transformed weights + for (auto i = 0u, m = 0u; i < inner_tile_i; i++) + { + for (auto j = 0u; j < inner_tile_j; j++, m++) + { + *(outptr + m*matrix_stride) = V[i][j]; + } + } + + inptr++; + outptr++; + } +} + +} // namespace weight_transform +} // namespace winograd +} // namespace arm_conv diff --git a/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_2x2_5x5.cpp b/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_2x2_5x5.cpp new file mode 100644 index 0000000000..9cdf15a4af --- /dev/null +++ b/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_2x2_5x5.cpp @@ -0,0 +1,381 @@ +/* + * Copyright (c) 2022 ARM Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include +#include + +namespace arm_conv { +namespace winograd { +namespace weight_transform { + +void arm_fp32_2x2_5x5( + unsigned int n_channels, + const float *inptr, const size_t ld_weight_row, const size_t ld_weight_col, + float *outptr, const size_t matrix_stride +) +{ +#ifdef __aarch64__ + // For each output channel + for (; n_channels >= 4; n_channels -= 4) + { + // Matrices used and computed in this kernel + float32x4_t w[5][5], Ww[6][5], V[6][6]; + + // Read weights + for (int i = 0; i < 5; i++) + { + for (int j = 0; j < 5; j++) + { + w[i][j] = vld1q_f32(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 5; j++) + { + // Ww[0][j] = w[0][j]/4.0f; + Ww[0][j] = vmulq_n_f32(w[0][j], 1.0f/4.0f); + + // Ww[1][j] = -( w[0][j] + w[1][j] + w[2][j] + w[3][j] + w[4][j])/6.0f; + Ww[1][j] = vmulq_n_f32( + vaddq_f32( + vaddq_f32( + vaddq_f32(w[1][j], w[0][j]), + vaddq_f32(w[3][j], w[2][j]) + ), + w[4][j] + ), + -1.0f/6.0f + ); + + // Ww[2][j] = +(-w[0][j] + w[1][j] - w[2][j] + w[3][j] - w[4][j])/6.0f; + // Ww[2][j] = ((w[1][j] - w[0][j]) + (w[3][j] - w[2][j]) - w[4][j])/6.0f; + Ww[2][j] = vmulq_n_f32( + vsubq_f32( + vaddq_f32( + vsubq_f32(w[1][j], w[0][j]), + vsubq_f32(w[3][j], w[2][j]) + ), + w[4][j] + ), + 1.0f/6.0f + ); + + // Ww[3][j] = (w[0][j]/8.0f + w[1][j]/4.0f + w[2][j]/2.0f + w[3][j] + 2*w[4][j])/3.0f; + Ww[3][j] = vmulq_n_f32( + vmlaq_n_f32( + vaddq_f32( + vaddq_f32(vmulq_n_f32(w[0][j], 1.0f/8.0f), vmulq_n_f32(w[1][j], 1.0f/4.0f)), + vaddq_f32(vmulq_n_f32(w[2][j], 1.0f/2.0f), w[3][j]) + ), + w[4][j], 2.0f + ), + 1.0f/3.0f + ); + + // Ww[4][j] = (w[0][j]/8.0f - w[1][j]/4.0f + w[2][j]/2.0f - w[3][j] + 2*w[4][j])/3.0f; + Ww[4][j] = vmulq_n_f32( + vmlaq_n_f32( + vaddq_f32( + vsubq_f32(vmulq_n_f32(w[0][j], 1.0f/8.0f), vmulq_n_f32(w[1][j], 1.0f/4.0f)), + vsubq_f32(vmulq_n_f32(w[2][j], 1.0f/2.0f), w[3][j]) + ), + w[4][j], 2.0f + ), + 1.0f/3.0f + ); + + // Ww[5][j] = w[4][j]; + Ww[5][j] = w[4][j]; + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + // V[i][0] = Ww[i][0]/4.0f; + V[i][0] = vmulq_n_f32(Ww[i][0], 1.0f/4.0f); + + // V[i][1] = -( Ww[i][0] + Ww[i][1] + Ww[i][2] + Ww[i][3] + Ww[i][4])/6.0f; + V[i][1] = vmulq_n_f32( + vaddq_f32( + vaddq_f32( + vaddq_f32(Ww[i][1], Ww[i][0]), + vaddq_f32(Ww[i][3], Ww[i][2]) + ), + Ww[i][4] + ), + -1.0f/6.0f + ); + + // V[i][2] = +(-Ww[i][0] + Ww[i][1] - Ww[i][2] + Ww[i][3] - Ww[i][4])/6.0f; + // V[i][2] = ((Ww[i][1] - Ww[i][0]) + (Ww[i][3] - Ww[i][2]) - Ww[i][4])/6.0f; + V[i][2] = vmulq_n_f32( + vsubq_f32( + vaddq_f32( + vsubq_f32(Ww[i][1], Ww[i][0]), + vsubq_f32(Ww[i][3], Ww[i][2]) + ), + Ww[i][4] + ), + 1.0f/6.0f + ); + + // V[i][3] = (Ww[i][0]/8.0f + Ww[i][1]/4.0f + Ww[i][2]/2.0f + Ww[i][3] + 2*Ww[i][4])/3.0f; + V[i][3] = vmulq_n_f32( + vmlaq_n_f32( + vaddq_f32( + vaddq_f32(vmulq_n_f32(Ww[i][0], 1.0f/8.0f), vmulq_n_f32(Ww[i][1], 1.0f/4.0f)), + vaddq_f32(vmulq_n_f32(Ww[i][2], 1.0f/2.0f), Ww[i][3]) + ), + Ww[i][4], 2.0f + ), + 1.0f/3.0f + ); + + // V[i][4] = (Ww[i][0]/8.0f - Ww[i][1]/4.0f + Ww[i][2]/2.0f - Ww[i][3] + 2*Ww[i][4])/3.0f; + V[i][4] = vmulq_n_f32( + vmlaq_n_f32( + vaddq_f32( + vsubq_f32(vmulq_n_f32(Ww[i][0], 1.0f/8.0f), vmulq_n_f32(Ww[i][1], 1.0f/4.0f)), + vsubq_f32(vmulq_n_f32(Ww[i][2], 1.0f/2.0f), Ww[i][3]) + ), + Ww[i][4], 2.0f + ), + 1.0f/3.0f + ); + + // V[i][5] = Ww[i][4]; + V[i][5] = Ww[i][4]; + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + vst1q_f32(outptr + m*matrix_stride, V[i][j]); + } + } + + inptr += 4; + outptr += 4; + } +#endif // __aarch64__ + for (; n_channels >= 2; n_channels -= 2) + { + // Matrices used and computed in this kernel + float32x2_t w[5][5], Ww[6][5], V[6][6]; + + // Read weights + for (int i = 0; i < 5; i++) + { + for (int j = 0; j < 5; j++) + { + w[i][j] = vld1_f32(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 5; j++) + { + // Ww[0][j] = w[0][j]/4.0f; + Ww[0][j] = vmul_n_f32(w[0][j], 1.0f/4.0f); + + // Ww[1][j] = -( w[0][j] + w[1][j] + w[2][j] + w[3][j] + w[4][j])/6.0f; + Ww[1][j] = vmul_n_f32( + vadd_f32( + vadd_f32( + vadd_f32(w[1][j], w[0][j]), + vadd_f32(w[3][j], w[2][j]) + ), + w[4][j] + ), + -1.0f/6.0f + ); + + // Ww[2][j] = +(-w[0][j] + w[1][j] - w[2][j] + w[3][j] - w[4][j])/6.0f; + // Ww[2][j] = ((w[1][j] - w[0][j]) + (w[3][j] - w[2][j]) - w[4][j])/6.0f; + Ww[2][j] = vmul_n_f32( + vsub_f32( + vadd_f32( + vsub_f32(w[1][j], w[0][j]), + vsub_f32(w[3][j], w[2][j]) + ), + w[4][j] + ), + 1.0f/6.0f + ); + + // Ww[3][j] = (w[0][j]/8.0f + w[1][j]/4.0f + w[2][j]/2.0f + w[3][j] + 2*w[4][j])/3.0f; + Ww[3][j] = vmul_n_f32( + vmla_n_f32( + vadd_f32( + vadd_f32(vmul_n_f32(w[0][j], 1.0f/8.0f), vmul_n_f32(w[1][j], 1.0f/4.0f)), + vadd_f32(vmul_n_f32(w[2][j], 1.0f/2.0f), w[3][j]) + ), + w[4][j], 2.0f + ), + 1.0f/3.0f + ); + + // Ww[4][j] = (w[0][j]/8.0f - w[1][j]/4.0f + w[2][j]/2.0f - w[3][j] + 2*w[4][j])/3.0f; + Ww[4][j] = vmul_n_f32( + vmla_n_f32( + vadd_f32( + vsub_f32(vmul_n_f32(w[0][j], 1.0f/8.0f), vmul_n_f32(w[1][j], 1.0f/4.0f)), + vsub_f32(vmul_n_f32(w[2][j], 1.0f/2.0f), w[3][j]) + ), + w[4][j], 2.0f + ), + 1.0f/3.0f + ); + + // Ww[5][j] = w[4][j]; + Ww[5][j] = w[4][j]; + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + // V[i][0] = Ww[i][0]/4.0f; + V[i][0] = vmul_n_f32(Ww[i][0], 1.0f/4.0f); + + // V[i][1] = -( Ww[i][0] + Ww[i][1] + Ww[i][2] + Ww[i][3] + Ww[i][4])/6.0f; + V[i][1] = vmul_n_f32( + vadd_f32( + vadd_f32( + vadd_f32(Ww[i][1], Ww[i][0]), + vadd_f32(Ww[i][3], Ww[i][2]) + ), + Ww[i][4] + ), + -1.0f/6.0f + ); + + // V[i][2] = +(-Ww[i][0] + Ww[i][1] - Ww[i][2] + Ww[i][3] - Ww[i][4])/6.0f; + // V[i][2] = ((Ww[i][1] - Ww[i][0]) + (Ww[i][3] - Ww[i][2]) - Ww[i][4])/6.0f; + V[i][2] = vmul_n_f32( + vsub_f32( + vadd_f32( + vsub_f32(Ww[i][1], Ww[i][0]), + vsub_f32(Ww[i][3], Ww[i][2]) + ), + Ww[i][4] + ), + 1.0f/6.0f + ); + + // V[i][3] = (Ww[i][0]/8.0f + Ww[i][1]/4.0f + Ww[i][2]/2.0f + Ww[i][3] + 2*Ww[i][4])/3.0f; + V[i][3] = vmul_n_f32( + vmla_n_f32( + vadd_f32( + vadd_f32(vmul_n_f32(Ww[i][0], 1.0f/8.0f), vmul_n_f32(Ww[i][1], 1.0f/4.0f)), + vadd_f32(vmul_n_f32(Ww[i][2], 1.0f/2.0f), Ww[i][3]) + ), + Ww[i][4], 2.0f + ), + 1.0f/3.0f + ); + + // V[i][4] = (Ww[i][0]/8.0f - Ww[i][1]/4.0f + Ww[i][2]/2.0f - Ww[i][3] + 2*Ww[i][4])/3.0f; + V[i][4] = vmul_n_f32( + vmla_n_f32( + vadd_f32( + vsub_f32(vmul_n_f32(Ww[i][0], 1.0f/8.0f), vmul_n_f32(Ww[i][1], 1.0f/4.0f)), + vsub_f32(vmul_n_f32(Ww[i][2], 1.0f/2.0f), Ww[i][3]) + ), + Ww[i][4], 2.0f + ), + 1.0f/3.0f + ); + + // V[i][5] = Ww[i][4]; + V[i][5] = Ww[i][4]; + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + vst1_f32(outptr + m*matrix_stride, V[i][j]); + } + } + + inptr += 2; + outptr += 2; + } + for (; n_channels; n_channels--) + { + // Matrices used and computed in this kernel + float w[5][5], Ww[6][5], V[6][6]; + + // Read weights + for (int i = 0; i < 5; i++) + { + for (int j = 0; j < 5; j++) + { + w[i][j] = *(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 5; j++) + { + Ww[0][j] = w[0][j]/4.0f; + Ww[1][j] = -( w[0][j] + w[1][j] + w[2][j] + w[3][j] + w[4][j])/6.0f; + Ww[2][j] = +(-w[0][j] + w[1][j] - w[2][j] + w[3][j] - w[4][j])/6.0f; + Ww[3][j] = (w[0][j]/8.0f + w[1][j]/4.0f + w[2][j]/2.0f + w[3][j] + 2*w[4][j])/3.0f; + Ww[4][j] = (w[0][j]/8.0f - w[1][j]/4.0f + w[2][j]/2.0f - w[3][j] + 2*w[4][j])/3.0f; + Ww[5][j] = w[4][j]; + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + V[i][0] = Ww[i][0]/4.0f; + V[i][1] = -( Ww[i][0] + Ww[i][1] + Ww[i][2] + Ww[i][3] + Ww[i][4])/6.0f; + V[i][2] = +(-Ww[i][0] + Ww[i][1] - Ww[i][2] + Ww[i][3] - Ww[i][4])/6.0f; + V[i][3] = (Ww[i][0]/8.0f + Ww[i][1]/4.0f + Ww[i][2]/2.0f + Ww[i][3] + 2*Ww[i][4])/3.0f; + V[i][4] = (Ww[i][0]/8.0f - Ww[i][1]/4.0f + Ww[i][2]/2.0f - Ww[i][3] + 2*Ww[i][4])/3.0f; + V[i][5] = Ww[i][4]; + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + *(outptr + m*matrix_stride) = V[i][j]; + } + } + + inptr++; + outptr++; + } +} + +} // namespace weight_transform +} // namespace winograd +} // namespace arm_conv diff --git a/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_4x4_3x3.cpp b/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_4x4_3x3.cpp new file mode 100644 index 0000000000..53cfa3d1d4 --- /dev/null +++ b/src/core/NEON/kernels/convolution/winograd/weight_transforms/arm_fp32_4x4_3x3.cpp @@ -0,0 +1,236 @@ +/* + * Copyright (c) 2022 ARM Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include +#include + +namespace arm_conv { +namespace winograd { +namespace weight_transform { + +void arm_fp32_4x4_3x3( + unsigned int n_channels, + const float *inptr, const size_t ld_weight_row, const size_t ld_weight_col, + float *outptr, const size_t matrix_stride +) +{ +#ifdef __aarch64__ + for (; n_channels >= 4; n_channels -= 4) + { + // Matrices used and computed in this kernel + float32x4_t w[3][3], Ww[6][3], V[6][6]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = vld1q_f32(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + // Ww[0][j] = 6*w[0][j]; + Ww[0][j] = vmulq_n_f32(w[0][j], 6.0); + + // Ww[1][j] = -4*w[0][j] + -4*w[1][j] + -4*w[2][j]; + Ww[1][j] = vmulq_n_f32(vaddq_f32(vaddq_f32(w[0][j], w[1][j]), w[2][j]), -4.0); + + // Ww[2][j] = -4*w[0][j] + 4*w[1][j] + -4*w[2][j]; + Ww[2][j] = vmulq_n_f32(vsubq_f32(vsubq_f32(w[1][j], w[0][j]), w[2][j]), 4.0); + + // Ww[3][j] = 1*w[0][j] + 2*w[1][j] + 4*w[2][j]; + Ww[3][j] = vmlaq_n_f32(vmlaq_n_f32(w[0][j], w[1][j], 2.0f), w[2][j], 4.0f); + + // Ww[4][j] = 1*w[0][j] + -2*w[1][j] + 4*w[2][j]; + Ww[4][j] = vmlaq_n_f32(vmlsq_n_f32(w[0][j], w[1][j], 2.0f), w[2][j], 4.0f); + + // Ww[5][j] = 24*w[2][j]; + Ww[5][j] = vmulq_n_f32(w[2][j], 24.0f); + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + const float recip576 = 1.0f / 576.0f; + + // V[i][0] = 6*Ww[i][0]; + V[i][0] = vmulq_n_f32(vmulq_n_f32(Ww[i][0], 6.0), recip576); + + // V[i][1] = -4*Ww[i][0] + -4*Ww[i][1] + -4*Ww[i][2]; + V[i][1] = vmulq_n_f32(vmulq_n_f32(vaddq_f32(vaddq_f32(Ww[i][0], Ww[i][1]), Ww[i][2]), -4.0), recip576); + + // V[i][2] = -4*Ww[i][0] + 4*Ww[i][1] + -4*Ww[i][2]; + V[i][2] = vmulq_n_f32(vmulq_n_f32(vsubq_f32(vsubq_f32(Ww[i][1], Ww[i][0]), Ww[i][2]), 4.0), recip576); + + // V[i][3] = 1*Ww[i][0] + 2*Ww[i][1] + 4*Ww[i][2]; + V[i][3] = vmulq_n_f32(vmlaq_n_f32(vmlaq_n_f32(Ww[i][0], Ww[i][1], 2.0f), Ww[i][2], 4.0f), recip576); + + // V[i][4] = 1*Ww[i][0] + -2*Ww[i][1] + 4*Ww[i][2]; + V[i][4] = vmulq_n_f32(vmlaq_n_f32(vmlsq_n_f32(Ww[i][0], Ww[i][1], 2.0f), Ww[i][2], 4.0f), recip576); + + // V[i][5] = 24*Ww[i][2]; + V[i][5] = vmulq_n_f32(vmulq_n_f32(Ww[i][2], 24.0f), recip576); + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + vst1q_f32(outptr + m*matrix_stride, V[i][j]); + } + } + + inptr += 4; + outptr += 4; + } +#endif // __aarch64__ + for (; n_channels >= 2; n_channels -= 2) + { + // Matrices used and computed in this kernel + float32x2_t w[3][3], Ww[6][3], V[6][6]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = vld1_f32(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + // Ww[0][j] = 6*w[0][j]; + Ww[0][j] = vmul_n_f32(w[0][j], 6.0); + + // Ww[1][j] = -4*w[0][j] + -4*w[1][j] + -4*w[2][j]; + Ww[1][j] = vmul_n_f32(vadd_f32(vadd_f32(w[0][j], w[1][j]), w[2][j]), -4.0); + + // Ww[2][j] = -4*w[0][j] + 4*w[1][j] + -4*w[2][j]; + Ww[2][j] = vmul_n_f32(vsub_f32(vsub_f32(w[1][j], w[0][j]), w[2][j]), 4.0); + + // Ww[3][j] = 1*w[0][j] + 2*w[1][j] + 4*w[2][j]; + Ww[3][j] = vmla_n_f32(vmla_n_f32(w[0][j], w[1][j], 2.0f), w[2][j], 4.0f); + + // Ww[4][j] = 1*w[0][j] + -2*w[1][j] + 4*w[2][j]; + Ww[4][j] = vmla_n_f32(vmls_n_f32(w[0][j], w[1][j], 2.0f), w[2][j], 4.0f); + + // Ww[5][j] = 24*w[2][j]; + Ww[5][j] = vmul_n_f32(w[2][j], 24.0f); + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + const float recip576 = 1.0f / 576.0f; + + // V[i][0] = 6*Ww[i][0]; + V[i][0] = vmul_n_f32(vmul_n_f32(Ww[i][0], 6.0), recip576); + + // V[i][1] = -4*Ww[i][0] + -4*Ww[i][1] + -4*Ww[i][2]; + V[i][1] = vmul_n_f32(vmul_n_f32(vadd_f32(vadd_f32(Ww[i][0], Ww[i][1]), Ww[i][2]), -4.0), recip576); + + // V[i][2] = -4*Ww[i][0] + 4*Ww[i][1] + -4*Ww[i][2]; + V[i][2] = vmul_n_f32(vmul_n_f32(vsub_f32(vsub_f32(Ww[i][1], Ww[i][0]), Ww[i][2]), 4.0), recip576); + + // V[i][3] = 1*Ww[i][0] + 2*Ww[i][1] + 4*Ww[i][2]; + V[i][3] = vmul_n_f32(vmla_n_f32(vmla_n_f32(Ww[i][0], Ww[i][1], 2.0f), Ww[i][2], 4.0f), recip576); + + // V[i][4] = 1*Ww[i][0] + -2*Ww[i][1] + 4*Ww[i][2]; + V[i][4] = vmul_n_f32(vmla_n_f32(vmls_n_f32(Ww[i][0], Ww[i][1], 2.0f), Ww[i][2], 4.0f), recip576); + + // V[i][5] = 24*Ww[i][2]; + V[i][5] = vmul_n_f32(vmul_n_f32(Ww[i][2], 24.0f), recip576); + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + vst1_f32(outptr + m*matrix_stride, V[i][j]); + } + } + + inptr += 2; + outptr += 2; + } + for (; n_channels; n_channels--) + { + // Matrices used and computed in this kernel + float w[3][3], Ww[6][3], V[6][6]; + + // Read weights + for (int i = 0; i < 3; i++) + { + for (int j = 0; j < 3; j++) + { + w[i][j] = *(inptr + i*ld_weight_row + j*ld_weight_col); + } + } + + // Compute the matrix W w + for (int j = 0; j < 3; j++) + { + Ww[0][j] = 6*w[0][j]; + Ww[1][j] = -4*w[0][j] + -4*w[1][j] + -4*w[2][j]; + Ww[2][j] = -4*w[0][j] + 4*w[1][j] + -4*w[2][j]; + Ww[3][j] = 1*w[0][j] + 2*w[1][j] + 4*w[2][j]; + Ww[4][j] = 1*w[0][j] + -2*w[1][j] + 4*w[2][j]; + Ww[5][j] = 24*w[2][j]; + } + + // Compute V = W w WT + for (int i = 0; i < 6; i++) + { + V[i][0] = ( 6*Ww[i][0]) / 576.0; + V[i][1] = (-4*Ww[i][0] + -4*Ww[i][1] + -4*Ww[i][2]) / 576.0; + V[i][2] = (-4*Ww[i][0] + 4*Ww[i][1] + -4*Ww[i][2]) / 576.0; + V[i][3] = ( 1*Ww[i][0] + 2*Ww[i][1] + 4*Ww[i][2]) / 576.0; + V[i][4] = ( 1*Ww[i][0] + -2*Ww[i][1] + 4*Ww[i][2]) / 576.0; + V[i][5] = (24*Ww[i][2]) / 576.0; + } + + // Store the transformed weights + for (int i = 0, m = 0; i < 6; i++) + { + for (int j = 0; j < 6; j++, m++) + { + *(outptr + m*matrix_stride) = V[i][j]; + } + } + + inptr++; + outptr++; + } +} + +} // namespace weight_transform +} // namespace winograd +} // namespace arm_conv diff --git a/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x2_1x7.cpp b/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x2_1x7.cpp new file mode 100644 index 0000000000..834f982f37 --- /dev/null +++ b/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x2_1x7.cpp @@ -0,0 +1,71 @@ +/* + * Copyright (c) 2022 ARM Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include + +namespace arm_conv { +namespace winograd { +namespace weight_transform { + +void cpp_fp32_1x2_1x7( + unsigned int n_channels, + const float* inptr, size_t, size_t ld_weight_col, + float *outptr, size_t matrix_stride +) +{ + for (; n_channels; n_channels--) + { + // Matrices used and computed in this kernel + float w[7], V[8]; + + // Read weights + for (int j = 0; j < 7; j++) + { + w[j] = *(inptr + j*ld_weight_col); + } + + // Compute V = w WT + V[0] = (w[0]*-1) / 36.0f; + V[1] = (w[1]*-1 + w[3]*-1 + w[5]*-1 + w[0]*1 + w[2]*1 + w[4]*1 + w[6]*1) / 48.0f; + V[2] = (w[0]*1 + w[1]*1 + w[2]*1 + w[3]*1 + w[4]*1 + w[5]*1 + w[6]*1) / 48.0f; + V[3] = (w[0]*-1 + w[6]*-64 + w[4]*-16 + w[2]*-4 + w[1]*2 + w[3]*8 + w[5]*32) / 120.0f; + V[4] = (w[0]*-1 + w[6]*-64 + w[5]*-32 + w[4]*-16 + w[3]*-8 + w[2]*-4 + w[1]*-2) / 120.0f; + V[5] = (w[5]*-243 + w[3]*-27 + w[1]*-3 + w[2]*9 + w[4]*81 + w[6]*729 + w[0]*1) / 720.0f; + V[6] = (w[1]*3 + w[2]*9 + w[3]*27 + w[4]*81 + w[5]*243 + w[6]*729 + w[0]*1) / 720.0f; + V[7] = (w[6]*1) / 1.0f; + + // Store the transformed weights + for (int j = 0; j < 8; j++) + { + *(outptr + j*matrix_stride) = V[j]; + } + + inptr++; + outptr++; + } +} + +} // namespace output_transform +} // namespace winograd +} // namespace arm_conv diff --git a/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x4_1x5.cpp b/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x4_1x5.cpp new file mode 100644 index 0000000000..585fb2516b --- /dev/null +++ b/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x4_1x5.cpp @@ -0,0 +1,77 @@ +/* + * Copyright (c) 2022 ARM Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include + +namespace arm_conv { +namespace winograd { +namespace weight_transform { + +void cpp_fp32_1x4_1x5( + unsigned int n_channels, + const float *inptr, + size_t, // ld_weight_row + size_t ld_weight_col, + float *outptr, + size_t matrix_stride +) +{ + constexpr auto kernel_cols = 5u, inner_tile_cols = 8u; + + // For each output channel + for (; n_channels; n_channels--) + { + // Matrices used and computed in this kernel + float w[kernel_cols], V[inner_tile_cols]; + + // Read weights + for (auto j = 0u; j < kernel_cols; j++) + { + w[j] = *(inptr + j * ld_weight_col); + } + + // Compute V = w WT + V[0] = (w[0]*-1) / 36; + V[1] = (w[1]*-1 + w[3]*-1 + w[0]*1 + w[2]*1 + w[4]*1) / 48; + V[2] = (w[0]*1 + w[1]*1 + w[2]*1 + w[3]*1 + w[4]*1) / 48; + V[3] = (w[0]*-1 + w[4]*-16 + w[2]*-4 + w[1]*2 + w[3]*8) / 120; + V[4] = (w[0]*-1 + w[4]*-16 + w[3]*-8 + w[2]*-4 + w[1]*-2) / 120; + V[5] = (w[3]*-27 + w[1]*-3 + w[2]*9 + w[4]*81 + w[0]*1) / 720; + V[6] = (w[1]*3 + w[2]*9 + w[3]*27 + w[4]*81 + w[0]*1) / 720; + V[7] = (w[4]*1) / 1; + + // Store the transformed weights + for (auto j = 0u; j < inner_tile_cols; j++) + { + *(outptr + j*matrix_stride) = V[j]; + } + + inptr++; + outptr++; + } +} + +} // namespace weight_transform +} // namespace winograd +} // namespace arm_conv diff --git a/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x6_1x3.cpp b/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x6_1x3.cpp new file mode 100644 index 0000000000..63754e529c --- /dev/null +++ b/src/core/NEON/kernels/convolution/winograd/weight_transforms/cpp_fp32_1x6_1x3.cpp @@ -0,0 +1,71 @@ +/* + * Copyright (c) 2022 ARM Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include + +namespace arm_conv { +namespace winograd { +namespace weight_transform { + +void cpp_fp32_1x6_1x3( + unsigned int n_channels, + const float *inptr, size_t, size_t ld_weight_col, + float *outptr, size_t matrix_stride +) +{ + for (; n_channels; n_channels--) + { + // Matrices used and computed in this kernel + float w[3], V[8]; + + // Read weights + for (int j = 0; j < 3; j++) + { + w[j] = *(inptr + j * ld_weight_col); + } + + // Compute V = w WT + V[0] = (w[0]*-1) / 36.0f; + V[1] = (w[1]*-1 + w[0]*1 + w[2]*1) / 48.0f; + V[2] = (w[0]*1 + w[1]*1 + w[2]*1) / 48.0f; + V[3] = (w[0]*-1 + w[2]*-4 + w[1]*2) / 120.0f; + V[4] = (w[0]*-1 + w[2]*-4 + w[1]*-2) / 120.0f; + V[5] = (w[1]*-3 + w[2]*9 + w[0]*1) / 720.0f; + V[6] = (w[1]*3 + w[2]*9 + w[0]*1) / 720.0f; + V[7] = (w[2]*1) / 1; + + // Store the transformed weights + for (int j = 0; j < 8; j++) + { + *(outptr + j*matrix_stride) = V[j]; + } + + inptr++; + outptr++; + } +} + +} // namespace weight_transform +} // namespace winograd +} // namespace arm_conv -- cgit v1.2.1