Executor 2.0: Stream assignment

dali/pipeline/executor/executor2/stream_assignment.h

@@ -0,0 +1,170 @@
+// Copyright (c) 2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef DALI_PIPELINE_EXECUTOR_EXECUTOR2_STREAM_ASSIGNMENT_H_
+#define DALI_PIPELINE_EXECUTOR_EXECUTOR2_STREAM_ASSIGNMENT_H_
+
+#include <algorithm>
+#include <cassert>
+#include <functional>
+#include <optional>
+#include <queue>
+#include <unordered_map>
+#include <set>
+#include <utility>
+#include <vector>
+#include "dali/pipeline/graph/graph_util.h"
+#include "dali/pipeline/executor/executor2/exec_graph.h"
+// TODO(michalz): This is here for review process only. Remove when exec2.h is available
+// #include "dali/pipeline/executor/executor2/exec2.h"
+#include "dali/pipeline/graph/op_graph2.h"
+
+namespace dali {
+namespace exec2 {
+
+// TODO(michalz): This is here for review process only. Remove when exec2.h is available
+enum class StreamPolicy : int {
+  Single,       //< There's just one stream that's used by all operators
+  PerBackend,   //< Operators are scheduled on a stream specific to their backend (mixed or GPU)
+  PerOperator   //< Independent operators are executed on separate streams.
+
+  // TODO(michalz): implement minimal assignment for PerOperator policy
+};
+
+
+template <StreamPolicy policy>
+class StreamAssignment;
+
+inline bool NeedsStream(const ExecNode *node) {
+  if (node->is_pipeline_output) {
+    for (auto &pipe_out : node->inputs) {
+      if (pipe_out->device == StorageDevice::GPU)
+        return true;
+    }
+  } else {
+    return node->backend != OpType::CPU;
+  }
+  return false;
+}
+
+inline OpType NodeType(const ExecNode *node) {
+  if (node->is_pipeline_output) {
+    OpType type = OpType::CPU;
+    for (auto &pipe_out : node->inputs) {
+      if (pipe_out->device == StorageDevice::GPU) {
+        auto producer_type = pipe_out->producer->backend;
+        if (producer_type == OpType::GPU) {
+          return OpType::GPU;
+        } else if (producer_type == OpType::MIXED) {
+          type = OpType::MIXED;
+        }
+      }
+    }
+    return type;
+  } else {
+    return node->backend;
+  }
+}
+
+/** A trivial stream policy, with just one stream shared by all non-CPU operaotrs. */
+template <>
+class StreamAssignment<StreamPolicy::Single> {
+ public:
+  explicit StreamAssignment(ExecGraph &graph) {
+    for (auto &node : graph.Nodes()) {
+      if (NeedsStream(&node)) {
+        needs_stream_ = true;
+        break;
+      }
+    }
+  }
+
+  std::optional<int> operator[](const ExecNode *node) const {
+    if (NeedsStream(node))
+      return 0;
+    else
+      return std::nullopt;
+  }
+
+  int NumStreams() const {
+    return needs_stream_ ? 1 : 0;
+  }
+
+ private:
+  bool needs_stream_ = false;
+};
+
+
+/** A simple stream policy where all mixed and GPU operators share their respective streams.
+ *
+ * In this policy there are 0..2 streams, depending on the number of mixed and GPU nodes:
+ * 0 - only CPU nodes
+ * 1 - there are some mixed or some GPU nodes, but not both
+ * 2 - there are both mixed and CPU nodes present.
+ */
+template <>
+class StreamAssignment<StreamPolicy::PerBackend> {
+ public:
+  explicit StreamAssignment(ExecGraph &graph) {
+    for (auto &node : graph.Nodes()) {
+      switch (NodeType(&node)) {
+      case OpType::GPU:
+        has_gpu_ = true;
+        if (has_mixed_)
+          return;  // we already have both, nothing more can happen
+        break;
+      case OpType::MIXED:
+        has_mixed_ = true;
+        if (has_gpu_)
+          return;  // we already have both, nothing more can happen
+        break;
+      default:
+        break;
+      }
+    }
+  }
+
+  /** Returns a stream index for a non-CPU operator.
+   *
+   * If the node is a Mixed node, it gets stream index 0.
+   * If the node is a GPU node it gets stream index 1 if there are any mixed nodes, otherwise
+   * the only stream is the GPU stream and the returned index is 0.
+   */
+  std::optional<int> operator[](const ExecNode *node) const {
+    switch (NodeType(node)) {
+    case OpType::CPU:
+      return std::nullopt;
+    case OpType::GPU:
+      return has_mixed_ ? 1 : 0;
+    case OpType::MIXED:
+      return 0;
+    default:
+      assert(false && "Unreachable");
+      return std::nullopt;
+    }
+  }
+
+  int NumStreams() const {
+    return has_gpu_ + has_mixed_;
+  }
+
+ private:
+  bool has_gpu_ = false;
+  bool has_mixed_ = false;
+};
+
+}  // namespace exec2
+}  // namespace dali
+
+#endif  // DALI_PIPELINE_EXECUTOR_EXECUTOR2_STREAM_ASSIGNMENT_H_

dali/pipeline/executor/executor2/stream_assignment_test.cc

@@ -0,0 +1,196 @@
+// Copyright (c) 2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include <gtest/gtest.h>
+#include <functional>
+#include <map>
+#include <optional>
+#include <string>
+#include <vector>
+#include "dali/pipeline/executor/executor2/stream_assignment.h"
+#include "dali/pipeline/operator/operator.h"
+#include "dali/pipeline/operator/operator_factory.h"
+
+namespace dali {
+
+template <typename Backend>
+class StreamAssignmentDummyOp : public Operator<Backend> {
+ public:
+  using Operator<Backend>::Operator;
+  USE_OPERATOR_MEMBERS();
+
+  void RunImpl(Workspace &ws) override {}
+  bool SetupImpl(std::vector<OutputDesc> &output_desc, const Workspace &ws) override {
+    return false;
+  }
+};
+
+DALI_SCHEMA(StreamAssignmentDummyOp)
+  .NumInput(0, 999)
+  .NumOutput(0)
+  .AdditionalOutputsFn([](const OpSpec &spec) {
+    return spec.NumOutput();
+  });
+
+DALI_REGISTER_OPERATOR(StreamAssignmentDummyOp, StreamAssignmentDummyOp<CPUBackend>, CPU);
+DALI_REGISTER_OPERATOR(StreamAssignmentDummyOp, StreamAssignmentDummyOp<MixedBackend>, Mixed);
+DALI_REGISTER_OPERATOR(StreamAssignmentDummyOp, StreamAssignmentDummyOp<GPUBackend>, GPU);
+
+namespace exec2 {
+
+namespace {
+
+OpSpec SpecDev(const std::string &device) {
+  return OpSpec("StreamAssignmentDummyOp")
+    .AddArg("device", device)
+    .AddArg("num_threads", 1)
+    .AddArg("max_batch_size", 1);
+}
+
+OpSpec SpecGPU() {
+  return SpecDev("gpu");
+}
+
+OpSpec SpecCPU() {
+  return SpecDev("cpu");
+}
+
+OpSpec SpecMixed() {
+  return SpecDev("mixed");
+}
+
+auto MakeNodeMap(const ExecGraph &graph) {
+  std::map<std::string_view, const ExecNode *, std::less<>> map;
+  for (auto &n : graph.Nodes())
+    if (!n.instance_name.empty()) {
+      map[n.instance_name] = &n;
+    }
+  return map;
+}
+
+}  // namespace
+
+TEST(Exec2Test, StreamAssignment_Single_OnlyCPU) {
+  graph::OpGraph::Builder b;
+  b.Add("a",
+        SpecCPU()
+        .AddOutput("a->out", "cpu"));
+  b.AddOutput("a->out_cpu");
+  auto g = std::move(b).GetGraph(true);
+  ExecGraph eg;
+  eg.Lower(g);
+
+  StreamAssignment<StreamPolicy::Single> assignment(eg);
+  auto map = MakeNodeMap(eg);
+  EXPECT_EQ(assignment[map["a"]], std::nullopt);
+}
+
+TEST(Exec2Test, StreamAssignment_Single_CPUMixedGPU) {
+  graph::OpGraph::Builder b;
+  b.Add("a",
+        SpecCPU()
+        .AddOutput("a->b", "cpu"));
+  b.Add("b",
+        SpecMixed()
+        .AddInput("a->b", "cpu")
+        .AddOutput("b->c", "gpu"));
+  b.Add("c",
+        SpecGPU()
+        .AddInput("b->c", "gpu")
+        .AddOutput("c->out", "gpu"));
+  b.AddOutput("c->out_gpu");
+  auto g = std::move(b).GetGraph(true);
+  ExecGraph eg;
+  eg.Lower(g);
+
+  StreamAssignment<StreamPolicy::Single> assignment(eg);
+  auto map = MakeNodeMap(eg);
+  EXPECT_EQ(assignment[map["a"]], std::nullopt);
+  EXPECT_EQ(assignment[map["b"]], 0);
+  EXPECT_EQ(assignment[map["c"]], 0);
+}
+
+
+TEST(Exec2Test, StreamAssignment_PerBackend_OnlyCPU) {
+  graph::OpGraph::Builder b;
+  b.Add("a",
+        SpecCPU()
+        .AddOutput("a->out", "cpu"));
+  b.AddOutput("a->out_cpu");
+  auto g = std::move(b).GetGraph(true);
+  ExecGraph eg;
+  eg.Lower(g);
+
+  StreamAssignment<StreamPolicy::Single> assignment(eg);
+  auto map = MakeNodeMap(eg);
+  EXPECT_EQ(assignment[map["a"]], std::nullopt);
+}
+
+
+TEST(Exec2Test, StreamAssignment_PerBackend_CPUMixed) {
+  graph::OpGraph::Builder b;
+  b.Add("a",
+        SpecCPU()
+        .AddOutput("a->b", "cpu")
+        .AddOutput("a->c", "cpu"));
+  b.Add("b",
+        SpecMixed()
+        .AddInput("a->b", "cpu")
+        .AddOutput("b->out", "gpu"));
+  b.Add("c",
+        SpecMixed()
+        .AddInput("a->c", "cpu")
+        .AddOutput("c->out", "gpu"));
+  b.AddOutput("b->out_gpu");
+  b.AddOutput("c->out_gpu");
+  auto g = std::move(b).GetGraph(true);
+  ExecGraph eg;
+  eg.Lower(g);
+
+  StreamAssignment<StreamPolicy::PerBackend> assignment(eg);
+  auto map = MakeNodeMap(eg);
+  EXPECT_EQ(assignment[map["a"]], std::nullopt);
+  EXPECT_EQ(assignment[map["b"]], 0);
+  EXPECT_EQ(assignment[map["c"]], 0);
+}
+
+TEST(Exec2Test, StreamAssignment_PerBackend_CPUMixedGPU) {
+  graph::OpGraph::Builder b;
+  b.Add("a",
+        SpecCPU()
+        .AddOutput("a->b", "cpu")
+        .AddOutput("a->c", "cpu"));
+  b.Add("b",
+        SpecGPU()
+        .AddInput("a->b", "cpu")
+        .AddOutput("b->out", "gpu"));
+  b.Add("c",
+        SpecMixed()
+        .AddInput("a->c", "cpu")
+        .AddOutput("c->out", "gpu"));
+  b.AddOutput("b->out_gpu");
+  b.AddOutput("c->out_gpu");
+  auto g = std::move(b).GetGraph(true);
+  ExecGraph eg;
+  eg.Lower(g);
+
+  StreamAssignment<StreamPolicy::PerBackend> assignment(eg);
+  auto map = MakeNodeMap(eg);
+  EXPECT_EQ(assignment[map["a"]], std::nullopt);
+  EXPECT_EQ(assignment[map["b"]], 1);
+  EXPECT_EQ(assignment[map["c"]], 0);
+}
+
+}  // namespace exec2
+}  // namespace dali