fix(mgb/tensor): do tensor overlap check only when d2d and h2h

GitOrigin-RevId: 9125936a35

src/core/impl/comp_node/atlas/comp_node.h

@@ -18,7 +18,9 @@
 namespace mgb {
 class AtlasCompNode final : public CompNodeImplHelper {
 public:
-    static constexpr Flag sm_flag = Flag::QUEUE_LIMITED | Flag::HAS_COPY_STREAM;
+    static constexpr Flag sm_flag = Flag::QUEUE_LIMITED |
+                                    Flag::HAS_COPY_STREAM |
+                                    Flag::SUPPORT_UNIFIED_ADDRESS;
 
     class CompNodeImpl;
     class EventImpl;

src/core/impl/comp_node/cambricon/comp_node.h

@@ -16,8 +16,9 @@
 namespace mgb {
     class CambriconCompNode final: public CompNodeImplHelper {
         public:
-            static constexpr Flag sm_flag =
-                    Flag::QUEUE_LIMITED | Flag::HAS_COPY_STREAM;
+            static constexpr Flag sm_flag = Flag::QUEUE_LIMITED |
+                                            Flag::HAS_COPY_STREAM |
+                                            Flag::SUPPORT_UNIFIED_ADDRESS;
 
             class CompNodeImpl;
             class EventImpl;

src/core/impl/comp_node/cpu/comp_node.h

@@ -38,7 +38,8 @@ namespace mgb {
             static constexpr Flag sm_flag =
                     Flag::SUPPORT_RECORDER |
                     Flag::RECORDER_SUPPORT_DYNAMIC_ALLOC |
-                    Flag::EVENT_DTOR_UNSAFE;
+                    Flag::EVENT_DTOR_UNSAFE |
+                    Flag::SUPPORT_UNIFIED_ADDRESS;
 
             //! base class for comp nodes that can be dispatched on CPU.
             //! This is currently used by CPU, FPGA and CADENCE

src/core/impl/comp_node/cuda/comp_node.h

@@ -16,8 +16,9 @@
 namespace mgb {
     class CudaCompNode final: public CompNodeImplHelper {
         public:
-            static constexpr Flag sm_flag =
-                    Flag::QUEUE_LIMITED | Flag::HAS_COPY_STREAM;
+            static constexpr Flag sm_flag = Flag::QUEUE_LIMITED |
+                                            Flag::HAS_COPY_STREAM |
+                                            Flag::SUPPORT_UNIFIED_ADDRESS;
 
             class CompNodeImpl;
             class EventImpl;

src/core/impl/comp_node/rocm/comp_node.h

@@ -16,7 +16,9 @@
 namespace mgb {
 class ROCmCompNode final : public CompNodeImplHelper {
 public:
-    static constexpr Flag sm_flag = Flag::QUEUE_LIMITED | Flag::HAS_COPY_STREAM;
+    static constexpr Flag sm_flag = Flag::QUEUE_LIMITED |
+                                    Flag::HAS_COPY_STREAM |
+                                    Flag::SUPPORT_UNIFIED_ADDRESS;
 
     class CompNodeImpl;
     class EventImpl;

src/core/impl/tensor.cpp

@@ -518,6 +518,60 @@ DEF(sub, )(const SubTensorSpec &spec) const {
 // def }
 
 /* ===================== TensorND::copy_from ===================== */
+namespace {
+/**
+ * \brief determine whether to check overlap of two tensors.
+ * \return true : when HostStorage || (DeviceStorage && SUPPORT_UNIFIED_ADDRESS)
+ * \note when both support unified address, we can treat them both on CPU. So,
+ * overlap check should be done
+ */
+template <typename TensorStorage, typename RStorage>
+inline bool should_check_overlap(const TensorND<TensorStorage>& dst,
+                                 const TensorND<RStorage>& src) {
+    return true;
+}
+
+template <>
+inline bool should_check_overlap<HostTensorStorage, DeviceTensorStorage>(
+        const HostTensorND& dst, const DeviceTensorND& src) {
+    return src.comp_node().contain_flag(
+            CompNode::Flag::SUPPORT_UNIFIED_ADDRESS);
+}
+
+template <>
+inline bool should_check_overlap<DeviceTensorStorage, HostTensorStorage>(
+        const DeviceTensorND& dst, const HostTensorND& src) {
+    return dst.comp_node().contain_flag(
+            CompNode::Flag::SUPPORT_UNIFIED_ADDRESS);
+}
+
+/**
+ * \brief D2D tensor copy should check overlap when
+ * 1. They are on the same mem node. But note that the address must be logical
+ * comparable. i.e. the original address alloc on enflame is uncomparable.
+ * 2. They both support unified address, so can be treated as CPU address.
+ */
+template <>
+inline bool should_check_overlap<DeviceTensorStorage, DeviceTensorStorage>(
+        const DeviceTensorND& dst, const DeviceTensorND& src) {
+    bool is_same_memnode =
+            dst.comp_node().mem_node() == src.comp_node().mem_node();
+    bool unified_address = src.comp_node().contain_flag(
+                                   CompNode::Flag::SUPPORT_UNIFIED_ADDRESS) &&
+                           dst.comp_node().contain_flag(
+                                   CompNode::Flag::SUPPORT_UNIFIED_ADDRESS);
+    return is_same_memnode || unified_address;
+}
+
+/**
+ * \brief check overlap of two tensors. throw exception when overlapped
+ */
+inline void check_overlapped(const dt_byte* dst_min, const dt_byte* dst_max,
+                             const dt_byte* src_min, const dt_byte* src_max) {
+    mgb_throw_if(src_min < dst_max && dst_min < src_max, TensorCopyOverlapError,
+                 "cound not perform copy between overlapped tensors");
+}
+}  // namespace
 
 template<class TensorStorage>
 template<class RStorage>
@@ -539,12 +593,12 @@ TensorND<TensorStorage>::copy_from(const TensorND<RStorage> &src) {
         return static_cast<ChainReturnType&>(*this);
     }
     if (src.layout().is_physical_contiguous()) {
-        const dt_byte
-            *dst_min = m_storage.ptr(), *dst_max = dst_min + size_bytes,
-            *src_min = src.storage().ptr(), *src_max = src_min + size_bytes;
-        mgb_throw_if(src_max > dst_min && dst_max > src_min,
-                TensorCopyOverlapError,
-                "cound not perform copy between overlapped tensors");
+        if (should_check_overlap(*this, src)) {
+            check_overlapped(m_storage.ptr(),
+                             m_storage.ptr() + size_bytes,
+                             src.storage().ptr(),
+                             src.storage().ptr() + size_bytes);
+        }
         m_storage.copy_from(src.storage(), size_bytes);
         return static_cast<ChainReturnType&>(*this);
     }
@@ -574,15 +628,12 @@ TensorND<TensorStorage>::copy_from_fixlayout(
         src_span = src.layout().span(),
         dst_span = layout().span();
 
-    const dt_byte
-        *src_ptr_min = src.raw_ptr() + src_span.low_byte,
-        *src_ptr_max = src.raw_ptr() + src_span.high_byte,
-        *dst_ptr_min = this->raw_ptr() + dst_span.low_byte,
-        *dst_ptr_max = this->raw_ptr() + dst_span.high_byte;
-
-    mgb_throw_if(src_ptr_max > dst_ptr_min && dst_ptr_max > src_ptr_min,
-            TensorCopyOverlapError,
-            "cound not perform copy between overlapped tensors");
+    if (should_check_overlap(*this, src)) {
+        check_overlapped(this->raw_ptr() + dst_span.low_byte,
+                         this->raw_ptr() + dst_span.high_byte,
+                         src.raw_ptr() + src_span.low_byte,
+                         src.raw_ptr() + src_span.high_byte);
+    }
 
     bool self_contig = m_layout.is_physical_contiguous(),
          src_contig = src.layout().is_physical_contiguous();

src/core/include/megbrain/comp_node.h

@@ -436,6 +436,10 @@ class CompNode {
             //! MGB_HAVE_THREAD=0. Usually this means that execution on the
             //! CompNode is synchronous, i.e. behaves like cpu:default
             SUPPORT_NO_THREAD = 1 << 5,
+
+            //! Whether this comp node supports unified address. i.e. CPU and
+            //! CUDA supports unified address.
+            SUPPORT_UNIFIED_ADDRESS = 1 << 6,
         };
 
         bool contain_flag(Flag flag) {