Merge branch 'itikhono/bug_fix/moc_in_prepostprocessing' of https://g…

…ithub.com/itikhono/openvino into itikhono/bug_fix/moc_in_prepostprocessing

src/common/snippets/CMakeLists.txt

@@ -16,10 +16,6 @@ source_group("src" FILES ${LIBRARY_SRC})
 source_group("include" FILES ${PUBLIC_HEADERS})
 
 # Create static library
-if((CMAKE_COMPILER_IS_GNUCXX OR OV_COMPILER_IS_CLANG) AND CMAKE_CXX_STANDARD GREATER_EQUAL 20)
-    set(CMAKE_CXX_FLAGS "-Wno-error=deprecated ${CMAKE_CXX_FLAGS}")
-endif()
-
 add_library(${TARGET_NAME} STATIC
             ${LIBRARY_SRC}
             ${PUBLIC_HEADERS})

src/common/snippets/include/snippets/kernel_executor_table.hpp

@@ -4,6 +4,7 @@
 
 #pragma once
 
+#include "openvino/util/pp.hpp"
 #include "snippets/lowered/linear_ir.hpp"
 #include <typeinfo>
 #if defined(SNIPPETS_DEBUG_CAPS) && !defined(_WIN32)
@@ -147,7 +148,7 @@ class KernelExecutorTable {
     /*** Returns lambda function that contains current state of the table, and restores this state when called  */
     std::function<void()> get_state_reset() {
         auto current_state = get_state();
-        return [=]() { reset_state(current_state); };
+        return [OV_CAPTURE_CPY_AND_THIS]() { reset_state(current_state); };
     }
 
     virtual ~KernelExecutorTable() = default;

src/common/transformations/src/transformations/convert_precision.cpp

@@ -1417,13 +1417,7 @@ bool fuse_type_to_constant(const std::shared_ptr<ov::Node>& node,
         new_const->validate_and_infer_types();
         new_const->set_friendly_name(constant->get_friendly_name());
         ov::copy_runtime_info(constant, new_const);
-
-        const auto& rt_info = node->get_rt_info();
-        auto weightless_caching_attr = rt_info.find(ov::WeightlessCacheAttribute::get_type_info_static());
-        if (weightless_caching_attr != rt_info.end()) {
-            new_const->get_rt_info()[ov::WeightlessCacheAttribute::get_type_info_static()] =
-                weightless_caching_attr->second;
-        }
+        ov::copy_weightless_cache_attr(constant, new_const);
         return true;
     }
     return false;

src/core/dev_api/openvino/core/rt_info/weightless_caching_attributes.hpp

@@ -10,6 +10,8 @@
 
 namespace ov {
 
+OPENVINO_API void copy_weightless_cache_attr(const std::shared_ptr<Node>& from, const std::shared_ptr<Node>& to);
+
 /**
  * @brief Holds weightless caching attributes of a single constant.
  *

src/core/src/op/util/weightless_caching_attributes.cpp

@@ -7,3 +7,13 @@
 bool ov::WeightlessCacheAttribute::is_copyable() const {
     return false;
 }
+
+OPENVINO_API void ov::copy_weightless_cache_attr(const std::shared_ptr<ov::Node>& from,
+                                                 const std::shared_ptr<ov::Node>& to) {
+    const auto& rt_info = from->get_rt_info();
+    auto weightless_caching_attr = rt_info.find(ov::WeightlessCacheAttribute::get_type_info_static());
+
+    if (weightless_caching_attr != rt_info.end()) {
+        to->get_rt_info()[ov::WeightlessCacheAttribute::get_type_info_static()] = weightless_caching_attr->second;
+    }
+}

src/core/src/pass/constant_folding.cpp

@@ -7,6 +7,7 @@
 #include "openvino/cc/pass/itt.hpp"
 #include "openvino/core/constant_fold_utils.hpp"
 #include "openvino/core/rt_info.hpp"
+#include "openvino/core/rt_info/weightless_caching_attributes.hpp"
 #include "openvino/op/constant.hpp"
 #include "openvino/op/convert.hpp"
 #include "openvino/op/util/op_types.hpp"
@@ -153,6 +154,7 @@ bool ov::pass::ConstantFolding::run_on_model(const std::shared_ptr<ov::Model>& m
                     copy_runtime_info_from_input_values(original_node);
                     // Propagate runtime info attributes to replacement
                     copy_runtime_info(original_node, replacement_ptr);
+                    ov::copy_weightless_cache_attr(original_node, replacement_ptr);
 
                     rewritten = true;
                 }

src/core/src/preprocess/pre_post_process.cpp

@@ -245,7 +245,7 @@ std::shared_ptr<Model> PrePostProcessor::build() {
     // MO path : [fw model conversion -> PrePostProcessing -> MOC] -> nncf
     // OVC path: [fw model conversion -> MOC] -> PrePostProcessing -> nncf
     //
-    // Since nncf is applied with a not fully optimized model, extra FQ ops might appear,
+    // Since nncf is applied to a not fully optimized model, extra FQ ops might appear,
     // which can affect both accuracy and performance.
     // PrePostProcessing is not part of OVC, so we have to insert an additional
     // Transformation calls inside PrePostProcessing.

src/inference/dev_api/openvino/runtime/system_conf.hpp

@@ -295,14 +295,6 @@ OPENVINO_RUNTIME_API void reserve_available_cpus(const std::vector<std::vector<i
  */
 OPENVINO_RUNTIME_API void set_cpu_used(const std::vector<int>& cpu_ids, const int used);
 
-/**
- * @brief      Get socket id by current numa node id
- * @ingroup    ov_dev_api_system_conf
- * @param[in]  numa_node_id numa node id
- * @return     socket id
- */
-OPENVINO_RUNTIME_API int get_socket_by_numa_node(int numa_node_id);
-
 /**
  * @brief      Get original socket id by current socket id, the input socket id is recalculated after filtering (like
  * numactl), while the original socket id is the original id before filtering

src/inference/dev_api/openvino/runtime/threading/cpu_streams_executor_internal.hpp

@@ -42,6 +42,7 @@ void get_cur_stream_info(const int stream_id,
                          int& concurrency,
                          int& core_type,
                          int& numa_node_id,
+                         int& socket_id,
                          int& max_threads_per_core);
 
 /**

src/inference/src/dev/threading/cpu_streams_executor.cpp

@@ -104,10 +104,11 @@ struct CPUStreamsExecutor::Impl {
                                    const int concurrency,
                                    const int core_type,
                                    const int numa_node_id,
+                                   const int socket_id,
                                    const int max_threads_per_core) {
             auto stream_processors = _impl->_config.get_stream_processor_ids();
-            _numaNodeId = std::max(0, numa_node_id);
-            _socketId = get_socket_by_numa_node(_numaNodeId);
+            _numaNodeId = numa_node_id;
+            _socketId = socket_id;
             if (stream_type == STREAM_WITHOUT_PARAM) {
                 _taskArena.reset(new custom::task_arena{custom::task_arena::constraints{}
                                                             .set_max_concurrency(concurrency)
@@ -159,6 +160,7 @@ struct CPUStreamsExecutor::Impl {
             int concurrency;
             int cpu_core_type;
             int numa_node_id;
+            int socket_id;
             int max_threads_per_core;
             StreamCreateType stream_type;
             const auto org_proc_type_table = get_org_proc_type_table();
@@ -173,6 +175,7 @@ struct CPUStreamsExecutor::Impl {
                                 concurrency,
                                 cpu_core_type,
                                 numa_node_id,
+                                socket_id,
                                 max_threads_per_core);
             if (concurrency <= 0) {
                 return;
@@ -182,6 +185,7 @@ struct CPUStreamsExecutor::Impl {
                                   concurrency,
                                   cpu_core_type,
                                   numa_node_id,
+                                  socket_id,
                                   max_threads_per_core);
         }
 #endif

src/inference/src/dev/threading/cpu_streams_executor_internal.cpp

@@ -21,6 +21,7 @@ void get_cur_stream_info(const int stream_id,
                          int& concurrency,
                          int& core_type,
                          int& numa_node_id,
+                         int& socket_id,
                          int& max_threads_per_core) {
     int stream_total = 0;
     size_t stream_info_id = 0;
@@ -36,6 +37,7 @@ void get_cur_stream_info(const int stream_id,
     concurrency = streams_info_table[stream_info_id][THREADS_PER_STREAM];
     core_type = streams_info_table[stream_info_id][PROC_TYPE];
     numa_node_id = streams_info_table[stream_info_id][STREAM_NUMA_NODE_ID];
+    socket_id = streams_info_table[stream_info_id][STREAM_SOCKET_ID];
     max_threads_per_core = 1;
     if (core_type == ALL_PROC) {
         for (size_t i = stream_info_id + 1; i < streams_info_table.size(); i++) {

src/inference/src/system_conf.cpp

@@ -285,10 +285,6 @@ void reserve_available_cpus(const std::vector<std::vector<int>> streams_info_tab
                             const int cpu_status) {}
 void set_cpu_used(const std::vector<int>& cpu_ids, const int used) {}
 
-int get_socket_by_numa_node(int numa_node_id) {
-    return -1;
-};
-
 int get_org_socket_id(int socket_id) {
     return -1;
 }
@@ -352,16 +348,6 @@ void reserve_available_cpus(const std::vector<std::vector<int>> streams_info_tab
                             const int cpu_status) {}
 void set_cpu_used(const std::vector<int>& cpu_ids, const int used) {}
 
-int get_socket_by_numa_node(int numa_node_id) {
-    CPU& cpu = cpu_info();
-    for (size_t i = 0; i < cpu._proc_type_table.size(); i++) {
-        if (cpu._proc_type_table[i][PROC_NUMA_NODE_ID] == numa_node_id) {
-            return cpu._proc_type_table[i][PROC_SOCKET_ID];
-        }
-    }
-    return -1;
-};
-
 int get_org_socket_id(int socket_id) {
     CPU& cpu = cpu_info();
     auto iter = cpu._socketid_mapping_table.find(socket_id);
@@ -522,16 +508,6 @@ void set_cpu_used(const std::vector<int>& cpu_ids, const int used) {
     }
 }
 
-int get_socket_by_numa_node(int numa_node_id) {
-    CPU& cpu = cpu_info();
-    for (int i = 0; i < cpu._processors; i++) {
-        if (cpu._cpu_mapping_table[i][CPU_MAP_NUMA_NODE_ID] == numa_node_id) {
-            return cpu._cpu_mapping_table[i][CPU_MAP_SOCKET_ID];
-        }
-    }
-    return -1;
-}
-
 int get_number_of_logical_cpu_cores(bool bigCoresOnly) {
     int logical_cores = parallel_get_max_threads();
 #    if (OV_THREAD == OV_THREAD_TBB || OV_THREAD == OV_THREAD_TBB_AUTO)

src/inference/tests/unit/cpu_stream_info_test.cpp

@@ -63,6 +63,7 @@ class LinuxCpuStreamTypeTests : public ov::test::TestsCommon,
             int test_concurrency;
             int test_core_type;
             int test_numa_node_id;
+            int test_socket_id = 0;
             int test_max_threads_per_core;
             get_cur_stream_info(i,
                                 test_data._cpu_pinning,
@@ -72,6 +73,7 @@ class LinuxCpuStreamTypeTests : public ov::test::TestsCommon,
                                 test_concurrency,
                                 test_core_type,
                                 test_numa_node_id,
+                                test_socket_id,
                                 test_max_threads_per_core);
             test_stream_types.push_back(test_stream_type);
             test_concurrencys.push_back(test_concurrency);

src/plugins/intel_cpu/CMakeLists.txt

@@ -8,10 +8,6 @@ endif()
 
 set(TARGET_NAME "openvino_intel_cpu_plugin")
 
-if((CMAKE_COMPILER_IS_GNUCXX OR OV_COMPILER_IS_CLANG) AND CMAKE_CXX_STANDARD GREATER_EQUAL 20)
-    set(CMAKE_CXX_FLAGS "-Wno-error=deprecated ${CMAKE_CXX_FLAGS}")
-endif()
-
 if(CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
     # C4267, 4244 issues from oneDNN headers conversion from 'XXX' to 'YYY', possible loss of data
     ov_add_compiler_flags(/wd4018)

src/plugins/intel_cpu/src/.clang-tidy

@@ -31,7 +31,6 @@
 # -readability-implicit-bool-conversion,
 # -readability-magic-numbers, cppcoreguidelines-avoid-magic-numbers
 # -readability-function-cognitive-complexity. Reasonable way to enforce splitting complex code into simple functions
-# -modernize-concat-nested-namespaces. More compact way when C++17 is available
 
 Checks: >
   -*,
@@ -58,7 +57,6 @@ Checks: >
   -cppcoreguidelines-avoid-magic-numbers,
   -readability-uppercase-literal-suffix,
   -readability-function-cognitive-complexity,
-  -modernize-concat-nested-namespaces,
 # Treat warnings as errors
 WarningsAsErrors: '*'
 # Use clang-format for applied fixes

src/plugins/intel_cpu/src/allocation_context.hpp

@@ -8,8 +8,7 @@
 #include <unordered_map>
 #include <vector>
 
-namespace ov {
-namespace intel_cpu {
+namespace ov::intel_cpu {
 
 class Node;
 class Edge;
@@ -22,5 +21,4 @@ struct AllocationContext {
     std::vector<size_t> syncPoints;
 };
 
-}  // namespace intel_cpu
-}  // namespace ov
+}  // namespace ov::intel_cpu

src/plugins/intel_cpu/src/cache/multi_cache.cpp

@@ -4,10 +4,8 @@
 
 #include "multi_cache.h"
 
-namespace ov {
-namespace intel_cpu {
+namespace ov::intel_cpu {
 
 std::atomic_size_t MultiCache::_typeIdCounter{0};
 
-}  // namespace intel_cpu
-}  // namespace ov
+}  // namespace ov::intel_cpu

src/plugins/intel_cpu/src/compiled_model.cpp

@@ -34,8 +34,7 @@
 
 using namespace ov::threading;
 
-namespace ov {
-namespace intel_cpu {
+namespace ov::intel_cpu {
 
 struct ImmediateSerialExecutor : public ov::threading::ITaskExecutor {
     void run(ov::threading::Task task) override {
@@ -153,7 +152,7 @@ CompiledModel::GraphGuard::Lock CompiledModel::get_graph() const {
     auto streamsExecutor = std::dynamic_pointer_cast<IStreamsExecutor>(m_task_executor);
     if (nullptr != streamsExecutor) {
         streamId = streamsExecutor->get_stream_id();
-        socketId = streamsExecutor->get_socket_id();
+        socketId = std::max(0, streamsExecutor->get_socket_id());
     }
     auto graphLock = GraphGuard::Lock(m_graphs[streamId % m_graphs.size()]);
     if (!graphLock._graph.IsReady()) {
@@ -362,5 +361,4 @@ void CompiledModel::release_memory() {
     }
 }
 
-}  // namespace intel_cpu
-}  // namespace ov
+}  // namespace ov::intel_cpu

src/plugins/intel_cpu/src/config.cpp

@@ -18,8 +18,7 @@
 #include "utils/debug_capabilities.h"
 #include "utils/precision_support.h"
 
-namespace ov {
-namespace intel_cpu {
+namespace ov::intel_cpu {
 
 using namespace ov::threading;
 using namespace dnnl::impl::cpu::x64;
@@ -509,5 +508,4 @@ void Config::applyRtInfo(const std::shared_ptr<const ov::Model>& model) {
     }
 }
 
-}  // namespace intel_cpu
-}  // namespace ov
+}  // namespace ov::intel_cpu

src/plugins/intel_cpu/src/cpu_map_scheduling.cpp

@@ -9,8 +9,7 @@
 #include "openvino/runtime/system_conf.hpp"
 #include "openvino/runtime/threading/cpu_streams_info.hpp"
 
-namespace ov {
-namespace intel_cpu {
+namespace ov::intel_cpu {
 
 std::vector<std::vector<int>> apply_scheduling_core_type(ov::hint::SchedulingCoreType& input_type,
                                                          const std::vector<std::vector<int>>& proc_type_table) {
@@ -106,5 +105,4 @@ bool get_cpu_pinning(bool& input_value,
     return result_value;
 }
 
-}  // namespace intel_cpu
-}  // namespace ov
+}  // namespace ov::intel_cpu

src/plugins/intel_cpu/src/cpu_map_scheduling.hpp

@@ -15,8 +15,7 @@
 #include "openvino/runtime/properties.hpp"
 #include "openvino/runtime/threading/istreams_executor.hpp"
 
-namespace ov {
-namespace intel_cpu {
+namespace ov::intel_cpu {
 
 /**
  * @brief      Limit available CPU resource in processors type table according to scheduling core type property
@@ -55,5 +54,4 @@ bool get_cpu_pinning(bool& input_value,
                      const std::vector<std::vector<int>>& proc_type_table,
                      const std::vector<std::vector<int>>& streams_info_table);
 
-}  // namespace intel_cpu
-}  // namespace ov
+}  // namespace ov::intel_cpu

src/plugins/intel_cpu/src/cpu_memory.cpp

@@ -18,8 +18,7 @@
 #    include <utility>
 #endif
 
-namespace ov {
-namespace intel_cpu {
+namespace ov::intel_cpu {
 template <>
 DnnlMemoryDescPtr IMemory::getDescWithType<DnnlMemoryDesc, 0, 0>() const {
     return MemoryDescUtils::convertToDnnlMemoryDesc(getDescPtr());
@@ -729,5 +728,4 @@ MemoryPtr split_vertical(const dnnl::engine& eng,
     return ptr;
 }
 
-}  // namespace intel_cpu
-}  // namespace ov
+}  // namespace ov::intel_cpu

src/plugins/intel_cpu/src/cpu_shape.cpp

@@ -6,8 +6,7 @@
 
 #include "utils/general_utils.h"
 
-namespace ov {
-namespace intel_cpu {
+namespace ov::intel_cpu {
 
 bool Shape::isCompatible(const VectorDims& vecDims) const {
     if (getRank() != vecDims.size()) {
@@ -77,5 +76,4 @@ Shape mergeShapes(const Shape& lhs, const Shape& rhs) {
     return Shape{resultMinDims, resultMaxDims};
 }
 
-}  // namespace intel_cpu
-}  // namespace ov
+}  // namespace ov::intel_cpu