• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // Copyright 2015 The Gemmlowp Authors. All Rights Reserved.
2 //
3 // Licensed under the Apache License, Version 2.0 (the "License");
4 // you may not use this file except in compliance with the License.
5 // You may obtain a copy of the License at
6 //
7 //     http://www.apache.org/licenses/LICENSE-2.0
8 //
9 // Unless required by applicable law or agreed to in writing, software
10 // distributed under the License is distributed on an "AS IS" BASIS,
11 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12 // See the License for the specific language governing permissions and
13 // limitations under the License.
14 
15 // pack_SSE.h: optimized SSE specializations of the templates in pack.h.
16 
17 #ifndef GEMMLOWP_INTERNAL_PACK_SSE_H_
18 #define GEMMLOWP_INTERNAL_PACK_SSE_H_
19 
20 #include <smmintrin.h>
21 #include "pack.h"
22 
23 namespace gemmlowp {
24 
25 // TODO: Add DepthMajorUint8SideMap
26 
27 typedef SideMap<const std::uint8_t, SideMapOrder::WidthMajor>
28     WidthMajorUint8SideMap;
29 
30 template <int Cells>
31 using WidthMajorSideFormatNCells4x2 =
32     KernelSideFormat<CellFormat<4, 2, CellOrder::WidthMajor>, Cells>;
33 
34 template <int Cells>
35 class PackingRegisterBlock<
36     WidthMajorUint8SideMap,
37     PackedSideBlock<WidthMajorSideFormatNCells4x2<Cells> > >
38     : public PackingRegisterBlockBase<
39           WidthMajorUint8SideMap,
40           PackedSideBlock<WidthMajorSideFormatNCells4x2<Cells> > > {
41  public:
42   typedef WidthMajorSideFormatNCells4x2<Cells> KernelSideFormat;
43   typedef typename KernelSideFormat::Cell CellFormat;
44   static constexpr int kCells = KernelSideFormat::kCells;
45   static constexpr int kCellWidth = CellFormat::kWidth;
46   static constexpr int kKernelWidth = CellFormat::kWidth * kCells;
47   static constexpr int kCellDepth = CellFormat::kDepth;
48   static constexpr int kCellSize = CellFormat::kSize;
49 
Pack(PackedSideBlock<KernelSideFormat> * dst,int start_width)50   void Pack(PackedSideBlock<KernelSideFormat>* dst, int start_width) {
51     std::uint8_t* dst_ptr = dst->current_data();
52     const int width_stride = this->complete_src_.width_stride();
53     int depth_step = 8;
54 
55     __m128i one = _mm_set1_epi16(1);
56     for (int cell_start_depth = 0; cell_start_depth < kRegisterSize;
57          cell_start_depth += depth_step) {
58       for (int cell_start_width = 0; cell_start_width < kKernelWidth;
59            cell_start_width += kCellWidth) {
60         std::int32_t* cell_sums_of_each_slice_ptr =
61             dst->sums_of_each_slice() + start_width + cell_start_width;
62         const std::uint8_t* src_data =
63             this->complete_src_.data(cell_start_width, cell_start_depth);
64 
65         __m128i xmm1 =
66             _mm_loadl_epi64(reinterpret_cast<const __m128i*>(&src_data[0]));
67         __m128i xmm2 = _mm_loadl_epi64(
68             reinterpret_cast<const __m128i*>(&src_data[1 * width_stride]));
69         __m128i xmm3 = _mm_loadl_epi64(
70             reinterpret_cast<const __m128i*>(&src_data[2 * width_stride]));
71         __m128i xmm4 = _mm_loadl_epi64(
72             reinterpret_cast<const __m128i*>(&src_data[3 * width_stride]));
73 
74         __m128i xmm5 = _mm_unpacklo_epi16(xmm1, xmm2);
75         __m128i xmm8 = _mm_shuffle_epi32(xmm5, 0x31);
76 
77         __m128i xmm6 = _mm_unpacklo_epi16(xmm3, xmm4);
78         __m128i xmm7 = _mm_shuffle_epi32(xmm6, 0x80);
79 
80         __m128i xmm9 = _mm_blend_epi16(xmm5, xmm7, 0xcc);
81         __m128i xmm10 = _mm_blend_epi16(xmm8, xmm6, 0xcc);
82 
83         _mm_storel_epi64(reinterpret_cast<__m128i*>(&dst_ptr[0]), xmm9);
84         _mm_storel_epi64(
85             reinterpret_cast<__m128i*>(&dst_ptr[kCellSize * kCells]), xmm10);
86 
87         __m128i xmm11 = _mm_shuffle_epi32(xmm9, 0xee);
88         __m128i xmm12 = _mm_shuffle_epi32(xmm10, 0xee);
89 
90         _mm_storel_epi64(
91             reinterpret_cast<__m128i*>(&dst_ptr[2 * kCellSize * kCells]),
92             xmm11);
93         _mm_storel_epi64(
94             reinterpret_cast<__m128i*>(&dst_ptr[3 * kCellSize * kCells]),
95             xmm12);
96 
97         xmm1 = _mm_cvtepu8_epi16(xmm9);
98         xmm2 = _mm_madd_epi16(xmm1, one);
99         __m128i sums_of_each_slice_xmm = _mm_loadu_si128(
100             reinterpret_cast<const __m128i*>(&cell_sums_of_each_slice_ptr[0]));
101         sums_of_each_slice_xmm = _mm_add_epi32(sums_of_each_slice_xmm, xmm2);
102 
103         xmm1 = _mm_cvtepu8_epi16(xmm10);
104         xmm2 = _mm_madd_epi16(xmm1, one);
105         sums_of_each_slice_xmm = _mm_add_epi32(sums_of_each_slice_xmm, xmm2);
106 
107         xmm1 = _mm_cvtepu8_epi16(xmm11);
108         xmm2 = _mm_madd_epi16(xmm1, one);
109         sums_of_each_slice_xmm = _mm_add_epi32(sums_of_each_slice_xmm, xmm2);
110 
111         xmm1 = _mm_cvtepu8_epi16(xmm12);
112         xmm2 = _mm_madd_epi16(xmm1, one);
113         sums_of_each_slice_xmm = _mm_add_epi32(sums_of_each_slice_xmm, xmm2);
114 
115         _mm_storeu_si128(
116             reinterpret_cast<__m128i*>(&cell_sums_of_each_slice_ptr[0]),
117             sums_of_each_slice_xmm);
118         dst_ptr += kCellSize;
119       }
120       dst_ptr += 3 * kCellSize * kCells;
121     }
122     dst->seek_forward_n_cells(kCells * kRegisterSize / kCellDepth);
123   }
124 };
125 
126 }  // namespace gemmlowp
127 
128 #endif  // GEMMLOWP_INTERNAL_PACK_SSE_H_
129