Merge pull request !27297 from hwjiaorui/mindrt-bug-fix

Merge pull request !27272 from liangzelang/sugbraph

Merge pull request !27067 from liangzelang/dev_mindrt

Merge pull request !27299 from gaoyong10/runtime_second12

Merge pull request !27264 from hwjiaorui/mindrt-bug-fix

Merge pull request !27240 from liangzelang/fix_weight_input_in_subgraph_bug

Merge pull request !26595 from caifubi/master-pynative-mindrt-lazy-build-with-ascend

Merge pull request !27239 from gaoyong10/runtime_second12

Merge pull request !27209 from parastooashtari/offline_dbg_bug

Merge pull request !27092 from gaoyong10/runtime_second12

Merge pull request !27084 from parastooashtari/gpu_dump_sink_issue

Merge pull request !26742 from chengbin/master

Merge pull request !27030 from gaoyong10/runtime_second12

Merge pull request !27012 from gaoyong10/runtime_second14

Merge pull request !27034 from limingqi107/bug_fix4

Merge pull request !22320 from chujinjin/support_heterogeneous_for_pynative

'LoadInputData', which is replaced by DataPrepareActor::PrepareData on
GPU(MindRT). This PR do the update of dynamic input in PrepareData and will
work for all platforms when they switch to MindRT.

Merge pull request !26981 from gaoyong10/runtime_second12

fix debugger, mindarmour, audio, gnn, grpc, ps/fl.

Merge pull request !26831 from caifubi/master-pynative-lazy-build-bug

Merge pull request !26752 from gaoyong10/runtime_second12

Merge pull request !26734 from limingqi107/bug_fix

Merge pull request !26712 from limingqi107/new_actor_runtime

Robin-hood-hashing (https://github.com/martinus/robin-hood-hashing)
is considered faster then std::unordered_map/set,
so we use it to improve mindspore performance.

1. robin_hood head file in `third_party/robin_hood/include`;
2. In `utils/hash_map.h` and `utils/hash_set.h`, we define:
 - mindspore::HashMap as an alias of robin_hood::unordered_map;
 - mindspore::HashSet as an alias of robin_hood::unordered_set;
3. Replace:
 - `#include <unordered_map>` --> `#include "utils/hash_map.h"`;
 - `#include <unordered_set>` --> `#include "utils/hash_set.h"`;
 - `std::unordered_map` --> `mindspore::HashMap`;
 - `std::unordered_set` --> `mindspore::HashSet`;
 - `map.insert(std::pair(key, value))` --> `map.emplace(key, value)`;
 - `[] (const std::pair<K, V> &p) {..} ` --> `[] (const auto &p) {..} `;
4. Fix issues found by switch to robin_hood:
 - AnfNodeConfig hash and equal;
 - Fix a bug in `Slice::operator==()`;
 - Fix a bug in `CNode::HasPrimalAttr()`;
 - Fix map.erase() usage bugs: `map.erase(iter++)` --> `iter = map.erase(iter)`;
 - Fix some iterator invalidated problem;
5. Some std::unordered_map/set can not replace by robin_hood:
 - As parameter of functions that exposed to python by pybind11;
 - Use bad hash that cause robin_hood::map over_flow, such as AbstractBasePtrListHasher;
6. Update cpp unit tests;
7. Add build option '-F' to enable robin_hood, default on.

Merge pull request !25132 from caifubi/master-pynative-mindrt-gpu-async-build