fix(dnn/gopt): add convolution nchw44-dot format gopt

GitOrigin-RevId: e8e1e96379

dnn/src/naive/convolution/helper.h

@@ -287,8 +287,7 @@ void compute2d(_megdnn_tensor_in src, ftype* __restrict fptr,
             }
         } else if (filter_meta.format == Format::NCHW44 ||
                    filter_meta.format == Format::NCHW44_DOT) {
-            if (filter_meta.format == Format::NCHW44 && !is_output &&
-                src.layout.ndim == 4) {
+            if (!is_output && src.layout.ndim == 4) {
                 return n * layout.stride[0] + c * layout.stride[1] +
                        h * layout.stride[2] + w * layout.stride[3];
             } else {

python_module/megengine/_internal/__init__.py

@@ -554,6 +554,7 @@ def optimize_for_inference(
     use_nchw4=False,
     use_nchw88=False,
     use_nchw44=False,
+    use_nchw44_dot=False,
     use_chwn4=False
 ):
     """optimize computing graph for inference
@@ -577,6 +578,8 @@ def optimize_for_inference(
         times.
     :param use_nchw44: whether to use NCHW44 tensor format. This maybe faster some
         times.
+    :param use_nchw44_dot: whether to use NCHW44_DOT tensor format. This format is
+        optimized for inference in armv8.2
     :param use_nchw32: whether to use NCHW32 tensor format. Mainly used for
         nvidia tensorcore.
     :param use_chwn4: whether to use CHWN4 tensor format. Mainly used for
@@ -605,6 +608,7 @@ def optimize_for_inference(
         "use_nchw32": "nchw32",
         "use_nchw88": "nchw88",
         "use_nchw44": "nchw44",
+        "use_nchw44_dot": "nchw44_dot",
         "use_chwn4": "chwn4",
     }.items():
         if settings[k]:

python_module/src/swig/misc.i

@@ -84,6 +84,7 @@ struct _OptimizeForInferenceOptions {
         SET(nhwcd4, NHWCD4);
         SET(nchw88, NCHW88);
         SET(nchw44, NCHW44);
+        SET(nchw44_dot, NCHW44_DOT);
         SET(nchw32, NCHW32);
         SET(chwn4, CHWN4);
 #undef SET

sdk/load-and-run/dump_with_testcase_mge.py

@@ -255,6 +255,7 @@ def optimize_for_inference(args, outputs):
         'enable_nchw4': 'use_nchw4',
         'enable_nchw88': 'use_nchw88',
         'enable_nchw44': 'use_nchw44',
+        'enable_nchw44_dot': 'use_nchw44_dot',
         'enable_nchw32': 'use_nchw32',
         'enable_chwn4': 'use_chwn4',
         'enable_fuse_conv_bias_nonlinearity': 'fuse_conv_bias_nonlinearity',
@@ -401,6 +402,12 @@ def main():
         'for inference'
     )
     parser.add_argument(
+        '--enable-nchw44-dot',
+        action='store_true',
+        help='transform the model format from NCHW to NCHW44_DOT '
+        'for optimizing armv8.2 dot in inference'
+    )
+    parser.add_argument(
         '--enable-nchw32',
         action='store_true',
         help='transform the model format from NCHW4 to NCHW32 '

src/core/include/megbrain/graph/cg.h

@@ -97,14 +97,15 @@ struct GraphCommonOptimizeOptions {
     bool fuse_conv_bias_with_z = false;
     enum LayoutTransform : uint32_t {
         DEFAULT,
-        NCHW4,   ///< compute using NCHW4 tensor format
-        NHWCD4,  ///< compute using NHWCD4 tensor format
-        NCHW88,  ///< compute using NCHW88 tensor format
-        NCHW44,  ///< compute using NCHW44 tensor format
-        NCHW32,  ///< compute using NCHW32 tensor format, used for
-                 ///< tensorcore
-        CHWN4,   ///< compute using CHWN4 tensor format, transformed mainly
-                 ///< used for cuda
+        NCHW4,       ///< compute using NCHW4 tensor format
+        NHWCD4,      ///< compute using NHWCD4 tensor format
+        NCHW88,      ///< compute using NCHW88 tensor format
+        NCHW44,      ///< compute using NCHW44 tensor format
+        NCHW44_DOT,  ///< compute using NCHW44_DOT tensor format
+        NCHW32,      ///< compute using NCHW32 tensor format, used for
+                     ///< tensorcore
+        CHWN4,       ///< compute using CHWN4 tensor format, transformed mainly
+                     ///< used for cuda
     };
     LayoutTransform layout_transform = LayoutTransform::DEFAULT;
 
@@ -142,6 +143,7 @@ struct GraphCommonOptimizeOptions {
     SET(nhwcd4, NHWCD4);
     SET(nchw88, NCHW88);
     SET(nchw44, NCHW44);
+    SET(nchw44_dot, NCHW44_DOT);
     SET(nchw32, NCHW32);
     SET(chwn4, CHWN4);
 #undef SET

src/gopt/impl/framework.cpp

@@ -738,6 +738,8 @@ const GraphOptimizer& GraphOptimizer::add_passes_for_optimize_options(
     });
     cb(nchw88, { add_pass(EnableNchwxxPass::make_nchwxx_converter(8)); });
     cb(nchw44, { add_pass(EnableNchwxxPass::make_nchwxx_converter(4)); });
+    cb(nchw44_dot,
+       { add_pass(EnableNchw44DotPass::make_nchw44_dot_converter()); });
     cb(nchw32, {
         add_pass<FuseConvBiasNonlinPass>();
         add_pass<FuseConvBiasZPass>();

src/gopt/impl/tensor_reformat.cpp

@@ -28,6 +28,7 @@
 #include "megbrain/opr/imgproc.h"
 #include "megbrain/opr/nn_int.h"
 
+#include "megdnn/opr_param_defs.h"
 #include "megdnn/tensor_format.h"
 
 #if MGB_ENABLE_TENSOR_RT
@@ -59,19 +60,19 @@ MGB_DEFINE_OPR_CLASS(TensorReformatPass::RelayoutPlaceholder,
 public:
     //! relayout type of this opr
     enum class LayoutType {
-        NCHW4_TO_NCHW32,              //!< from nchw4 layout to nchw32 layout
-        NCHW32_TO_NCHW4,              //!< from nchw32 layout to nchw4 layout
-        NCHW4_TO_CHWN4,               //!< from nchw4 layout to chwn4 layout
-        CHWN4_TO_NCHW4,               //!< from chwn4 layout to nchw4 layout
-        NCHW_TO_NCHW4,                //!< from nchw layout to nchw4 layout
-        NCHW4_TO_NCHW,                //!< from nchw4 layout to nchw layout
-        NCHW_TO_NCHW88,               //!< from nchw layout to nchw88 layout
-        NCHW88_TO_NCHW,               //!< from nchw88 layout to nchw layout
-
-        WEIGHT_NCHW_TO_NCHW4_DENSE,   //!< weight from nchw layout to nchw4
-                                      //!< layout
-        WEIGHT_NCHW_TO_NCHW4_GROUP,   //!< group weight from nchw layout to
-                                      //!< nchw4 layout
+        NCHW4_TO_NCHW32,  //!< from nchw4 layout to nchw32 layout
+        NCHW32_TO_NCHW4,  //!< from nchw32 layout to nchw4 layout
+        NCHW4_TO_CHWN4,   //!< from nchw4 layout to chwn4 layout
+        CHWN4_TO_NCHW4,   //!< from chwn4 layout to nchw4 layout
+        NCHW_TO_NCHW4,    //!< from nchw layout to nchw4 layout
+        NCHW4_TO_NCHW,    //!< from nchw4 layout to nchw layout
+        NCHW_TO_NCHW88,   //!< from nchw layout to nchw88 layout
+        NCHW88_TO_NCHW,   //!< from nchw88 layout to nchw layout
+
+        WEIGHT_NCHW_TO_NCHW4_DENSE,  //!< weight from nchw layout to nchw4
+                                     //!< layout
+        WEIGHT_NCHW_TO_NCHW4_GROUP,  //!< group weight from nchw layout to
+                                     //!< nchw4 layout
 
         WEIGHT_NCHW_TO_NCHW88_DENSE,  //!< weight from nchw layout to nchw88
                                       //!< layout
@@ -92,6 +93,10 @@ public:
         //!< the weight layout of input is nchw output is nchw44, special for
         //!< shape weight in nchw like {64, 2, 3, 3} to {16, 3, 3, 2, 4}
         WEIGHT_HYBIRD_NCHW_NCHW44,
+        WEIGHT_NCHW_TO_NCHW44_DOT_DENSE,  //!< weight from NCHW44 layout to
+                                          //!< NCHW44_DOT layout dense
+        WEIGHT_NCHW_TO_NCHW44_DOT_GROUP,  //!< weight from NCHW44 layout to
+                                          //!< NCHW44_DOT layout group
     };
 
     RelayoutPlaceholder(VarNode* src_var, LayoutType layout_type);
@@ -268,7 +273,9 @@ void TensorReformatPass::RelayoutPlaceholder::init_output_static_infer_desc() {
             dst[3] = inp_shape[1];
             dst[4] = 8;
         } else if (layout_type() == RelayoutPlaceholder::LayoutType::
-                                            WEIGHT_NCHW_TO_NCHW44_DENSE) {
+                                            WEIGHT_NCHW_TO_NCHW44_DENSE ||
+                   layout_type() == RelayoutPlaceholder::LayoutType::
+                                            WEIGHT_NCHW_TO_NCHW44_DOT_DENSE) {
             mgb_assert(inp_shape.ndim == 4 && inp_shape[0] % 4 == 0 &&
                        inp_shape[1] % 4 == 0);
             dst.ndim = 6;
@@ -279,7 +286,9 @@ void TensorReformatPass::RelayoutPlaceholder::init_output_static_infer_desc() {
             dst[4] = 4;
             dst[5] = 4;
         } else if (layout_type() == RelayoutPlaceholder::LayoutType::
-                                            WEIGHT_NCHW_TO_NCHW44_GROUP) {
+                                            WEIGHT_NCHW_TO_NCHW44_GROUP ||
+                   layout_type() == RelayoutPlaceholder::LayoutType::
+                                            WEIGHT_NCHW_TO_NCHW44_DOT_GROUP) {
             mgb_assert(inp_shape.ndim == 5 && inp_shape[1] % 4 == 0 &&
                        inp_shape[2] % 4 == 0);
             dst.ndim = 7;
@@ -646,6 +655,42 @@ void TensorReformatPass::translate_pass(OptState& opt) const {
         auto y2 = opr::Reshape::make(y1, tshp1);
         return y2.node();
     };
+    reformat[LayoutType::WEIGHT_NCHW_TO_NCHW44_DOT_DENSE] =
+            [](VarNode* inp) -> VarNode* {
+        auto x = SymbolVar(inp);
+        auto xshp = opr::GetVarShape::make(x);
+        auto cv = [&x](int v) { return x.make_scalar(v); };
+        auto sub = [&xshp, &cv](int idx) {
+            return opr::IndexAt::make(xshp, {{0, cv(idx)}});
+        };
+        auto tshp0 = opr::Concat::make(
+                     {sub(0) / 4, cv(4), sub(1) / 4, cv(4), sub(2), sub(3)}, 0),
+             tshp1 = opr::Concat::make(
+                     {sub(0) / 4, sub(1) / 4, sub(2), sub(3), cv(4), cv(4)}, 0);
+        auto y0 = opr::Reshape::make(x, tshp0);
+        auto y1 = opr::Dimshuffle::make(y0, {0, 2, 4, 5, 1, 3});
+        auto y2 = opr::Reshape::make(y1, tshp1);
+        return y2.node();
+    };
+    reformat[LayoutType::WEIGHT_NCHW_TO_NCHW44_DOT_GROUP] =
+            [](VarNode* inp) -> VarNode* {
+        auto x = SymbolVar(inp);
+        auto xshp = opr::GetVarShape::make(x);
+        auto cv = [&x](int v) { return x.make_scalar(v); };
+        auto sub = [&xshp, &cv](int idx) {
+            return opr::IndexAt::make(xshp, {{0, cv(idx)}});
+        };
+        auto tshp0 = opr::Concat::make({sub(0), sub(1) / 4, cv(4), sub(2) / 4,
+                                        cv(4), sub(3), sub(4)},
+                                       0),
+             tshp1 = opr::Concat::make({sub(0), sub(1) / 4, sub(2) / 4, sub(3),
+                                        sub(4), cv(4), cv(4)},
+                                       0);
+        auto y0 = opr::Reshape::make(x, tshp0);
+        auto y1 = opr::Dimshuffle::make(y0, {0, 1, 3, 5, 6, 2, 4});
+        auto y2 = opr::Reshape::make(y1, tshp1);
+        return y2.node();
+    };
 
     auto rewriter = opt.graph().make_rewriter();
     auto on_opr = [&reformat, &rewriter](OperatorNodeBase* opr) {
@@ -1601,12 +1646,7 @@ VarNode* EnableNchwxxPass::on_graph_endpoint_var(VarNode* new_var,
     return new_var;
 }
 
-std::unique_ptr<EnableNchwxxPass> EnableNchwxxPass::make_nchwxx_converter(
-        size_t pack_c_size) {
-    auto ret = std::make_unique<EnableNchwxxPass>(pack_c_size);
-    ret->set_var_replace_check_flag(VarReplaceCheckFlag::NOCHECK);
-    //! First is whether the conv can trans to nchwxx, second is the filter
-    //! trans mode
+void EnableNchwxxPass::fill_opr_convert_fun(size_t pack_c_size){
     using RelayoutMode = RelayoutPlaceholder::LayoutType;
     using TestFilterResult = std::pair<TransType, RelayoutMode>;
     RelayoutMode weight_to_nchwxx_mode_dense =
@@ -1954,8 +1994,7 @@ std::unique_ptr<EnableNchwxxPass> EnableNchwxxPass::make_nchwxx_converter(
         return serialization::copy_opr_shallow(*opr, temp_inp, opr->config());
     };
 
-    ret->set_name(convter_pass_name);
-    auto&& replace_func = ret->m_opr_replace_func;
+    auto&& replace_func = m_opr_replace_func;
     //! supportted nchwxx
     replace_func[opr::Convolution::typeinfo()] = replace_conv_opr;
     replace_func[opr::ConvBias::typeinfo()] = replace_conv_bias_opr;
@@ -1978,6 +2017,246 @@ std::unique_ptr<EnableNchwxxPass> EnableNchwxxPass::make_nchwxx_converter(
     replace_func[opr::WarpPerspectiveForward::typeinfo()] =
             relayout_inp_to_nchw;
     replace_func[opr::WarpAffineForward::typeinfo()] = relayout_inp_to_nchw;
+}
+
+std::unique_ptr<EnableNchwxxPass> EnableNchwxxPass::make_nchwxx_converter(
+        size_t pack_c_size) {
+    auto ret = std::make_unique<EnableNchwxxPass>(pack_c_size);
+    ret->set_var_replace_check_flag(VarReplaceCheckFlag::NOCHECK);
+    std::string convter_pass_name = "conv_format_nchw88";
+    if (pack_c_size == 4) {
+        convter_pass_name = "conv_format_nchw44";
+    }
+    ret->fill_opr_convert_fun(pack_c_size);
+    ret->set_name(convter_pass_name);
+    return ret;
+}
+
+/* ================ EnableNchw44DotPass =============== */
+VarNode* EnableNchw44DotPass::on_graph_endpoint_var(VarNode* new_var,
+                                                    VarNode* orig_var) const {
+    if (!orig_var->shape().eq_shape(new_var->shape())) {
+        return RelayoutPlaceholder::make(
+                       new_var, RelayoutPlaceholder::LayoutType::NCHW4_TO_NCHW)
+                .node();
+    }
+    return new_var;
+}
+
+std::unique_ptr<EnableNchw44DotPass>
+EnableNchw44DotPass::make_nchw44_dot_converter() {
+    auto ret = std::make_unique<EnableNchw44DotPass>();
+    ret->set_var_replace_check_flag(VarReplaceCheckFlag::NOCHECK);
+    //! First is whether the conv can trans to nchwxx, second is the filter
+    //! trans mode
+    using RelayoutMode = RelayoutPlaceholder::LayoutType;
+    using TestTransResult = std::pair<TransType, RelayoutMode>;
+    megdnn::param::ConvolutionV0::Format conv_dot_format =
+            megdnn::param::ConvBias::Format::NCHW44_DOT;
+    constexpr size_t pack_c_size = 4_z;
+    auto test_trans_nchw44_dot =
+            [](const megdnn::param::Convolution::Sparse conv_mode,
+               const VarNode* filter) -> TestTransResult {
+        TestTransResult ret{TransType::TRANS_NONE, {}};
+        if (conv_mode == megdnn::param::Convolution::Sparse::DENSE) {
+            size_t IC = filter->shape()[1];
+            size_t OC = filter->shape()[0];
+            if ((IC % pack_c_size == 0) && (OC % pack_c_size == 0)) {
+                ret.first = TransType::TRANS_PURE_NCHWXX;
+                ret.second = RelayoutMode::WEIGHT_NCHW_TO_NCHW44_DOT_DENSE;
+            } else if (IC < pack_c_size && OC % pack_c_size == 0) {
+                ret.first = TransType::TRANS_HYBIRD_NCHWXX;
+                ret.second = RelayoutMode::WEIGHT_HYBIRD_NCHW_NCHW44;
+            }
+        } else {
+            mgb_assert(conv_mode == megdnn::param::Convolution::Sparse::GROUP);
+            size_t group = filter->shape()[0];
+            size_t ocpg = filter->shape()[1];
+            size_t icpg = filter->shape()[2];
+            if (icpg == 1 && ocpg == 1 && (group % pack_c_size == 0)) {
+                ret.first = TransType::TRANS_NONE;
+            } else if ((icpg % pack_c_size == 0) && (ocpg % pack_c_size == 0)) {
+                ret.first = TransType::TRANS_PURE_NCHWXX;
+                ret.second = RelayoutMode::WEIGHT_NCHW_TO_NCHW44_DOT_GROUP;
+            }
+        }
+        return ret;
+    };
+    auto replace_conv_opr = [test_trans_nchw44_dot, conv_dot_format](
+                                    OperatorNodeBase* opr,
+                                    const VarNodeArray& new_inp) {
+        mgb_assert(opr->input().size() == new_inp.size());
+        auto& conv_opr = opr->cast_final_safe<opr::ConvolutionForward>();
+        mgb_assert(conv_opr.param().format ==
+                           megdnn::param::Convolution::Format::NCHW,
+                   "ConvertFormat Pass only support converting NCHW to "
+                   "NCHW44_DOT");
+        auto is_trans =
+                test_trans_nchw44_dot(conv_opr.param().sparse, new_inp[1]);
+        //! can not trans to nchwxx
+        if (is_trans.first == TransType::TRANS_NONE) {
+            mgb_assert(new_inp[1]->shape().ndim == 4 ||
+                               new_inp[1]->shape().ndim == 5,
+                       "The origin filter is not NCHW mode");
+            VarNodeArray temp_inp = new_inp;
+            //! if src is nchwxx, should RelayoutPlaceholder to nchw
+            if (temp_inp[0]->shape().ndim == 5) {
+                auto new_src = RelayoutPlaceholder::make(
+                        new_inp[0], RelayoutMode::NCHW4_TO_NCHW);
+                temp_inp[0] = new_src.node();
+            }
+            auto new_opr = serialization::copy_opr_shallow(*opr, temp_inp,
+                                                           opr->config());
+            return new_opr;
+        } else if (is_trans.first == TransType::TRANS_PURE_NCHWXX) {
+            //! filter trans to nchwxx mode
+            mgb_assert(new_inp[1]->shape().ndim == 4 ||
+                               new_inp[1]->shape().ndim == 5,
+                       "The origin filter is not NCHW mode");
+            VarNode *conv_src = new_inp[0], *conv_filter = new_inp[1];
+            auto new_filter =
+                    RelayoutPlaceholder::make(new_inp[1], is_trans.second);
+            conv_filter = new_filter.node();
+            //! src trans to nchwxx mode
+            if (new_inp[0]->shape().ndim != 5) {
+                mgb_assert(new_inp[0]->shape().ndim == 4);
+                auto new_src = RelayoutPlaceholder::make(
+                        new_inp[0], RelayoutMode::NCHW_TO_NCHW4);
+                conv_src = new_src.node();
+            }
+            auto new_param = conv_opr.param();
+            new_param.format = conv_dot_format;
+            mgb_assert(conv_src->shape().ndim == 5 &&
+                               conv_filter->shape().ndim >= 6,
+                       "The conv src dim is not trans to nchwxx");
+            auto new_conv_opr = opr::Convolution::make(
+                    conv_src, conv_filter, new_param,
+                    conv_opr.execution_policy(), conv_opr.config());
+            OperatorNodeBase* new_opr = new_conv_opr.node()->owner_opr();
+            mgb_assert(new_conv_opr.shape().ndim == 5,
+                       "The conv dst dim is not trans to nchwxx");
+            return new_opr;
+        } else {
+            mgb_assert(is_trans.first == TransType::TRANS_HYBIRD_NCHWXX);
+            VarNode *conv_src = new_inp[0], *conv_filter = new_inp[1];
+            auto new_filter =
+                    RelayoutPlaceholder::make(new_inp[1], is_trans.second);
+            conv_filter = new_filter.node();
+            mgb_assert(conv_src->shape().ndim == 4 &&
+                               conv_filter->shape().ndim == 5,
+                       "The src and filter is OK");
+            auto new_param = conv_opr.param();
+            new_param.format = conv_dot_format;
+            auto new_conv_opr = opr::Convolution::make(
+                    conv_src, conv_filter, new_param,
+                    conv_opr.execution_policy(), conv_opr.config());
+            OperatorNodeBase* new_opr = new_conv_opr.node()->owner_opr();
+            mgb_assert(new_conv_opr.shape().ndim == 5,
+                       "The conv dst dim is not trans to nchwxx");
+            return new_opr;
+        }
+    };
+
+    auto replace_conv_bias_opr = [test_trans_nchw44_dot, conv_dot_format](
+                                         OperatorNodeBase* opr,
+                                         const VarNodeArray& new_inp) {
+        mgb_assert(opr->input().size() == new_inp.size());
+        auto& conv_bias_opr = opr->cast_final_safe<opr::ConvBiasForward>();
+        mgb_assert(conv_bias_opr.param().format ==
+                           megdnn::param::ConvBias::Format::NCHW,
+                   "ConvertFormat Pass only support converting NCHW to NCHWXX");
+        auto is_trans =
+                test_trans_nchw44_dot(conv_bias_opr.param().sparse, new_inp[1]);
+        //! can not trans to nchwxx
+        if (is_trans.first == TransType::TRANS_NONE) {
+            mgb_assert(new_inp[1]->shape().ndim == 4 ||
+                               new_inp[1]->shape().ndim == 5,
+                       "The origin filter is not NCHW mode");
+            VarNodeArray temp_inp = new_inp;
+            //! if src is nchwxx, should RelayoutPlaceholder to nchw
+            if (temp_inp[0]->shape().ndim == 5) {
+                auto new_src = RelayoutPlaceholder::make(
+                        new_inp[0], RelayoutMode::NCHW4_TO_NCHW);
+                temp_inp[0] = new_src.node();
+            }
+            //! the bias is nchwxx
+            if (temp_inp[2]->shape().ndim == 5) {
+                auto new_bias = RelayoutPlaceholder::make(
+                        new_inp[2], RelayoutMode::NCHW4_TO_NCHW);
+                temp_inp[2] = new_bias.node();
+            }
+            auto new_opr = serialization::copy_opr_shallow(*opr, temp_inp,
+                                                           opr->config());
+            return new_opr;
+        } else if (is_trans.first == TransType::TRANS_PURE_NCHWXX) {
+            VarNode *conv_bias_src = new_inp[0], *conv_bias_filter = new_inp[1],
+                    *conv_bias_bias = new_inp[2];
+            //! filter trans to nchwxx mode
+            mgb_assert(new_inp[1]->shape().ndim == 4 ||
+                               new_inp[1]->shape().ndim == 5,
+                       "The origin filter is not NCHW mode");
+            auto new_filter =
+                    RelayoutPlaceholder::make(new_inp[1], is_trans.second);
+            conv_bias_filter = new_filter.node();
+            //! src trans to nchwxx mode
+            if (new_inp[0]->shape().ndim != 5) {
+                mgb_assert(new_inp[0]->shape().ndim == 4);
+                auto new_src = RelayoutPlaceholder::make(
+                        new_inp[0], RelayoutMode::NCHW_TO_NCHW4);
+                conv_bias_src = new_src.node();
+            }
+            //! bias trans to nchwxx mode, bias may be scale
+            if (new_inp[2]->shape().ndim == 4) {
+                auto new_bias = RelayoutPlaceholder::make(
+                        new_inp[2], RelayoutMode::NCHW_TO_NCHW4);
+                conv_bias_bias = new_bias.node();
+            }
+
+            auto new_param = conv_bias_opr.param();
+            new_param.format = conv_dot_format;
+            mgb_assert(conv_bias_src->shape().ndim == 5 &&
+                               conv_bias_filter->shape().ndim >= 6,
+                       "The conv_bias src dim is not trans to nchwxx");
+            auto new_conv_bias_opr = opr::ConvBias::make(
+                    conv_bias_src, conv_bias_filter, conv_bias_bias, new_param,
+                    conv_bias_opr.execution_policy(), conv_bias_opr.config());
+            OperatorNodeBase* new_opr = new_conv_bias_opr.node()->owner_opr();
+            mgb_assert(new_conv_bias_opr.shape().ndim == 5,
+                       "The conv_bias dst dim is not trans to nchwxx");
+            return new_opr;
+        } else {
+            mgb_assert(is_trans.first == TransType::TRANS_HYBIRD_NCHWXX);
+            VarNode *conv_bias_src = new_inp[0], *conv_bias_filter = new_inp[1],
+                    *conv_bias_bias = new_inp[2];
+            auto new_filter =
+                    RelayoutPlaceholder::make(new_inp[1], is_trans.second);
+            conv_bias_filter = new_filter.node();
+            //! bias trans to nchwxx mode, bias may be scale
+            if (new_inp[2]->shape().ndim == 4) {
+                auto new_bias = RelayoutPlaceholder::make(
+                        new_inp[2], RelayoutMode::NCHW_TO_NCHW4);
+                conv_bias_bias = new_bias.node();
+            }
+            mgb_assert(conv_bias_src->shape().ndim == 4 &&
+                       conv_bias_filter->shape().ndim == 5);
+            mgb_assert((conv_bias_bias->shape().ndim == 5) ||
+                       conv_bias_bias->shape().is_scalar());
+            auto new_param = conv_bias_opr.param();
+            new_param.format = conv_dot_format;
+            auto new_conv_bias_opr = opr::ConvBias::make(
+                    conv_bias_src, conv_bias_filter, conv_bias_bias, new_param,
+                    conv_bias_opr.execution_policy(), conv_bias_opr.config());
+            OperatorNodeBase* new_opr = new_conv_bias_opr.node()->owner_opr();
+            mgb_assert(new_conv_bias_opr.shape().ndim == 5,
+                       "The conv dst dim is not trans to nchwxx");
+            return new_opr;
+        }
+    };
+    ret->fill_opr_convert_fun(4);
+    auto&& replace_func = ret->m_opr_replace_func;
+    //! supportted nchwxx
+    replace_func[opr::Convolution::typeinfo()] = replace_conv_opr;
+    replace_func[opr::ConvBias::typeinfo()] = replace_conv_bias_opr;
     return ret;
 }
 

src/gopt/include/megbrain/gopt/inference.h

@@ -236,8 +236,10 @@ namespace gopt {
         VarNode* on_graph_endpoint_var(VarNode* new_var,
                                        VarNode* orig_var) const override;
     public:
-        const char* name() const override { return mgb_cstr_log("tensor_format_nchw4"); }
-        
+        const char* name() const override {
+            return mgb_cstr_log("tensor_format_nchw4");
+        }
+
         //! make nchw -> nchw4 converter opt pass
         static std::unique_ptr<EnableNCHW4Pass> make_nchw4_converter();
     };
@@ -246,30 +248,48 @@ namespace gopt {
      * \brief convert tensor format to nchwxx to speed up inference on certain
      * devices
      */
-    class EnableNchwxxPass final : public TensorReformatPass {
+    class EnableNchwxxPass : public TensorReformatPass {
         std::string m_name = "tensor_format_nchwxx";
         size_t m_pack_c_size;
         VarNode* on_graph_endpoint_var(VarNode* new_var,
                                        VarNode* orig_var) const override;
+    public:
+        EnableNchwxxPass(size_t pack_c_size) : m_pack_c_size(pack_c_size) {}
+
         //! the flag for conv to transform to nchwxx
         enum class TransType {
             TRANS_PURE_NCHWXX,    //!< weight and src all trans to nchwxx
             TRANS_HYBIRD_NCHWXX,  //!< input is nchw, output is nchwxx
             TRANS_NONE,           //!< no need trans
         };
-
-    public:
-        EnableNchwxxPass(size_t pack_c_size) : m_pack_c_size(pack_c_size) {}
         const char* name() const override {
             return mgb_cstr_log(m_name.c_str());
         }
         void set_name(std::string in_name) { m_name = in_name; }
+
+        void fill_opr_convert_fun(size_t pack_c_size);
+
         //! make nchw -> nchwxx converter opt pass, pack_c_size is the x, like
         //! 4,8,16
         static std::unique_ptr<EnableNchwxxPass> make_nchwxx_converter(
                 size_t pack_c_size);
     };
 
+    /*!
+     * \brief convert tensor format from nchw44 to nchw44_dot to speed up
+     * inference on armv8.2
+     */
+    class EnableNchw44DotPass final : public EnableNchwxxPass {
+        std::string m_name = "tensor_format_nchw44_dot";
+        VarNode* on_graph_endpoint_var(VarNode* new_var,
+                                       VarNode* orig_var) const override;
+
+    public:
+        EnableNchw44DotPass() : EnableNchwxxPass(4) {}
+        //! make nchw44 -> nchw44_dot converter opt pass
+        static std::unique_ptr<EnableNchw44DotPass> make_nchw44_dot_converter();
+    };
+
     struct OptimizeForInferenceOptions : cg::GraphCommonOptimizeOptions {};
 
     /*!

src/gopt/test/inference.cpp

@@ -2356,7 +2356,7 @@ TEST(TestGoptInference, ConvertFormatNCHW4GPU) {
                         .rename(name),
                 dtype);
     };
-    
+
     auto x = mkvar("x", {2, 4, 16, 16}, dtype::QuantizedS8(2.5f));
     opr::ConvBias::Param param_conv_bias;
     param_conv_bias.format = opr::ConvBias::Param::Format::NCHW;
@@ -2376,7 +2376,7 @@ TEST(TestGoptInference, ConvertFormatNCHW4GPU) {
          b2 = mkcvar("b2", {1, 8, 1, 1}, dtype::QuantizedS32(6.25f));
     auto conv2 = opr::ConvBiasForward::make(conv1, w2, b2,
             param_conv_bias, {}, OperatorNodeConfig{dtype::QuantizedS8{2.5f}});
-    
+
     auto y = opr::TypeCvt::make(conv2, dtype::Float32());
 
     SymbolVar y_opt;
@@ -2617,4 +2617,86 @@ TEST(TestGoptInference, ConvertFormatNCHW44) {
     MGB_ASSERT_TENSOR_NEAR(host_y, host_y_opt, 1e-1);
 }
 
+TEST(TestGoptInference, ConvertFormatNCHW44_DOT) {
+    HostTensorGenerator<> gen;
+    auto cn = CompNode::load("cpu0");
+    auto graph = ComputingGraph::make();
+    graph->options().graph_opt_level = 0;
+    auto mkvar = [&](const char* name, const TensorShape& shp) {
+        return opr::Host2DeviceCopy::make(*graph, gen(shp, cn)).rename(name);
+    };
+    auto mkcvar = [&](const char* name, const TensorShape& shp) {
+        return opr::SharedDeviceTensor::make(*graph, *gen(shp, cn))
+                .rename(name);
+    };
+
+    auto host_x = gen({2, 3, 16, 16}, cn);
+    auto x = opr::Host2DeviceCopy::make(*graph, host_x);
+    //!Hybrid nchw88 mode
+    opr::Convolution::Param param_conv;
+    param_conv.pad_h = param_conv.pad_w = 1;
+    auto w1 = mkcvar("w1", {8, 3, 3, 3}),
+         conv1 = opr::Convolution::make(x, w1, param_conv);
+    //!channel wise
+    opr::ConvBias::Param param_conv_bias;
+    param_conv_bias.pad_h = param_conv_bias.pad_w = 1;
+    param_conv_bias.sparse = opr::ConvBias::Param::Sparse::GROUP;
+    auto w2 = mkcvar("w2", {8, 1, 1, 3, 3}), b2 = mkcvar("b2", {1, 8, 1, 1}),
+         conv2 = opr::ConvBias::make(conv1, w2, b2, param_conv_bias);
+    //! group
+    auto w3 = mkcvar("w3", {2, 4, 4, 3, 3}), b3 = mkcvar("b3", {1, 8, 1, 1}),
+         conv3 = opr::ConvBias::make(conv2, w3, b3, param_conv_bias);
+
+    auto shape_of = opr::GetVarShape::make(conv3);
+    auto subtensor = opr::Subtensor::make(
+            shape_of, {opr::Subtensor::AxisIndexer::make_interval(
+                              0, x.make_scalar(2), None, x.make_scalar(1))});
+    opr::Resize::Param param_resize;
+    param_resize.format = opr::Resize::Param::Format::NCHW;
+    auto resize = opr::ResizeForward::make(conv3, subtensor * 2, param_resize);
+    auto mat = mkcvar("mat", {2, 3, 3}),
+         warp = opr::WarpPerspectiveForward::make(
+                 resize, mat, nullptr, cg::var_from_tensor_shape(x, {4, 4}));
+
+    auto b = mkvar("b", {1, 8, 1, 1}),
+         elem = opr::Elemwise::make({warp + b},
+                                    opr::Elemwise::Param::Mode::RELU);
+    //! Dense
+    param_conv_bias.sparse = opr::ConvBias::Param::Sparse::DENSE;
+    param_conv_bias.pad_h = param_conv_bias.pad_w = 1;
+    auto w4 = mkcvar("w4", {4, 8, 3, 3}), b4 = mkcvar("b4", {1, 4, 1, 1}),
+         conv4 = opr::ConvBias::make(elem, w4, b4, param_conv_bias);
+    auto w5 = mkcvar("w5", {6, 4, 3, 3}), b5 = mkcvar("b5", {1, 6, 1, 1}),
+         conv5 = opr::ConvBias::make(conv4, w5, b5, param_conv_bias);
+    auto w6 = mkcvar("w6", {4, 6, 3, 3}), b6 = mkcvar("b6", {1, 4, 1, 1}),
+         y = opr::ConvBias::make(conv5, w6, b6, param_conv_bias);
+
+    SymbolVar y_opt;
+    auto options = gopt::OptimizeForInferenceOptions{};
+    options.enable_nchw44_dot();
+    unpack_vector(gopt::optimize_for_inference({y}, options), y_opt);
+
+    ASSERT_EQ(opr::ConvBias::Param::Format::NCHW44_DOT,
+              find_opr<opr::Convolution>(y_opt).param().format);
+    ASSERT_EQ(opr::Convolution::Param::Format::NCHW,
+              find_opr<opr::ConvBias>(y_opt).param().format);
+
+    graph->compile({{y_opt, {}}})
+            ->to_json()
+            ->writeto_fpath(
+                    output_file("TestGoptInference.ConvertFormatNCHW44.json"));
+
+    HostTensorND host_y_opt, host_y;
+    auto func = graph->compile({make_callback_copy(y, host_y),
+                                make_callback_copy(y_opt, host_y_opt)});
+    func->execute();
+    //! meybe go to winograd in x86-32, so set error 1e-1
+    MGB_ASSERT_TENSOR_NEAR(host_y, host_y_opt, 1e-1);
+
+    *host_x = *gen({2, 3, 32, 32}, cn);
+    func->execute();
+    //! meybe go to winograd in x86-32, so set error 1e-1
+    MGB_ASSERT_TENSOR_NEAR(host_y, host_y_opt, 1e-1);
+}
+
 // vim: syntax=cpp.doxygen foldmethod=marker foldmarker=f{{{,f}}}