From c609c031f19260ce53f3808c1df014d0965ecb57 Mon Sep 17 00:00:00 2001
From: Megvii Engine Team <megengine@megvii.com>
Date: Fri, 14 Jan 2022 13:23:14 +0800
Subject: [PATCH] refactor(dispatch): implement symbol

GitOrigin-RevId: c7bd86f5c1a412205c801226fdd8ba01ce02d4c1
---
 .../imperative/transformations/symbol.h       | 131 ++++++++++++++++++
 1 file changed, 131 insertions(+)
 create mode 100644 imperative/src/include/megbrain/imperative/transformations/symbol.h

diff --git a/imperative/src/include/megbrain/imperative/transformations/symbol.h b/imperative/src/include/megbrain/imperative/transformations/symbol.h
new file mode 100644
index 000000000..2032ef42a
--- /dev/null
+++ b/imperative/src/include/megbrain/imperative/transformations/symbol.h
@@ -0,0 +1,131 @@
+/**
+ * \file imperative/src/include/megbrain/imperative/symbol.h
+ * MegEngine is Licensed under the Apache License, Version 2.0 (the "License")
+ *
+ * Copyright (c) 2014-2021 Megvii Inc. All rights reserved.
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT ARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ */
+
+#pragma once
+
+#include <future>
+#include <variant>
+
+#include "megbrain/imperative/dispatch.h"
+#include "megbrain/imperative/interpreter.h"
+#include "megbrain/imperative/opr_utility.h"
+#include "megbrain/imperative/utils/helper.h"
+#include "megbrain/opr/io.h"
+
+namespace mgb::imperative {
+
+class SymbolValue final : public ValueImpl<SymbolValue> {
+private:
+    VarNode* m_node = nullptr;
+
+public:
+    SymbolValue(VarNode* node) : m_node(node) {}
+
+    VarNode* node() const { return m_node; }
+
+    std::string to_string() const override { return ssprintf("VarNode{%p}", m_node); }
+
+    void clear() override { m_node = nullptr; }
+};
+
+/**
+ * \brief this transformation is used to handle VarNode.
+ *
+ * Unlike other transformations, this transformation is not used in Tensor evaluation.
+ * when user calls py_apply(SymbolVar), we'll switch current transformation context to a
+ * special symbol context. The advantage is that we can handle scalar by
+ * ScalarTransformation.
+ */
+class SymbolTransformation final : public Transformation {
+private:
+    ComputingGraph* m_graph = nullptr;
+
+public:
+    SymbolTransformation(ComputingGraph* graph) : m_graph(graph) {}
+    std::vector<ValueRef> apply_transformation(
+            const Operator& op, Span<ValueRef> inputs) override {
+        if (auto* apply_op = op.as<ApplyOp>()) {
+            SmallVector<VarNode*> input_nodes;
+            for (auto&& input : inputs) {
+                input_nodes.push_back(input.cast<SymbolValue>().node());
+            }
+            auto output_nodes = OpDef::apply_on_var_node(apply_op->op(), input_nodes);
+            std::vector<ValueRef> outputs;
+            for (auto&& output_node : output_nodes) {
+                outputs.push_back(SymbolValue::make(output_node));
+            }
+            return outputs;
+        } else if (auto* create_tensor = op.as<CreateTensor>()) {
+            auto&& args = create_tensor->parse(inputs);
+            mgb_assert(
+                    args.kind == CreateTensor::Const,
+                    "only const value is allowed here");
+            auto* node = opr::ImmutableTensor::make(*m_graph, *args.host, {}).node();
+            return {SymbolValue::make(node)};
+        } else if (auto* get_attr = op.as<GetAttr>()) {
+            auto* node = inputs.as_array<1>()[0].cast<SymbolValue>().node();
+            switch (get_attr->attr()) {
+                case GetAttr::DType:
+                    return {DTypeValue::make(node->dtype())};
+                case GetAttr::Device:
+                    return {CompNodeValue::make(node->comp_node())};
+                case GetAttr::Shape: {
+                    if (!cg::is_static_var_shape(node)) {
+                        mgb_log_debug(
+                                "shape inference invalid for %s", node->name().c_str());
+                        return {ValueRef()};
+                    }
+                    auto shape = m_graph->static_infer_manager().infer_shape(node);
+                    return {ShapeValue::make(ValueShape::from(shape))};
+                }
+                case GetAttr::Value: {
+                    if (!cg::is_static_var_value(node)) {
+                        mgb_log_debug(
+                                "value inference invalid for %s", node->name().c_str());
+                        return {ValueRef()};
+                    }
+                    auto inferred_value =
+                            m_graph->static_infer_manager().infer_value(node);
+                    HostTensorND host_value(node->comp_node(), node->dtype());
+                    host_value.copy_from(inferred_value);
+                    return {HostValue::make(host_value)};
+                }
+                case GetAttr::Data: {
+                    if (!cg::is_static_var_value(node)) {
+                        mgb_log_debug(
+                                "value inference invalid for %s", node->name().c_str());
+                        return {ValueRef()};
+                    }
+                    auto inferred_value =
+                            m_graph->static_infer_manager().infer_value(node);
+                    DeviceTensorND dev_value(node->comp_node(), node->dtype());
+                    dev_value.copy_from(inferred_value);
+                    return {DeviceValue::make(dev_value)};
+                }
+                default:
+                    mgb_throw(
+                            MegBrainError, "Symbol: malformed GetAttr: %s",
+                            op.to_string().c_str());
+            }
+        } else {
+            return op.fallback(inputs);
+        }
+    }
+
+    ValueRef unwrap(ValueRef value) override {
+        mgb_assert(!value.is<SymbolValue>(), "SymbolValue doesn't support unwrap");
+        return value;
+    }
+
+    std::string name() const override { return "SymbolTransformation"; }
+};
+
+}  // namespace mgb::imperative
-- 
GitLab