remove databatch dependency to tbb

WeBankBlockchain · Sep 14, 2024 · 8d8ffc8 · 8d8ffc8
1 parent d1d7637
commit 8d8ffc8
Show file tree

Hide file tree

Showing 5 changed files with 18 additions and 16 deletions.
diff --git a/cpp/ppc-framework/io/DataBatch.h b/cpp/ppc-framework/io/DataBatch.h
@@ -20,7 +20,6 @@
 #pragma once
 #include "../Common.h"
 #include <bcos-utilities/Common.h>
-#include <tbb/tbb.h>
 #include <boost/lexical_cast.hpp>
 #include <algorithm>
 #include <memory>
@@ -199,18 +198,8 @@ class DataBatch
 
     DataSchema getDataSchema() const { return m_dataSchema; }
 
-    uint32_t dedup()
-    {
-        if (!m_data || m_data->empty())
-        {
-            return 0;
-        }
 
-        tbb::parallel_sort(m_data->begin(), m_data->end());
-        auto unique_end = std::unique(m_data->begin(), m_data->end());
-        m_data->erase(unique_end, m_data->end());
-        return m_data->size();
-    }
+    std::shared_ptr<std::vector<DataType>>& mutableData() const { return m_data; }
 
 private:
     std::shared_ptr<std::vector<DataType>> m_data;

diff --git a/cpp/wedpr-computing/ppc-psi/src/cm2020-psi/Common.h b/cpp/wedpr-computing/ppc-psi/src/cm2020-psi/Common.h
@@ -21,9 +21,9 @@
 
 #include "ppc-framework/Common.h"
 #include "ppc-framework/protocol/PPCMessageFace.h"
-
 #include <bcos-utilities/Common.h>
 #include <bcos-utilities/Log.h>
+#include <tbb/tbb.h>
 
 
 namespace ppc::psi
@@ -65,4 +65,17 @@ enum class CM2020PSIRetCode : int
     INVALID_TASK_PARAM = -3002
 };
 
+inline uint32_t dedupDataBatch(ppc::io::DataBatch::Ptr dataBatch)
+{
+    if (!dataBatch || dataBatch->mutableData() == nullptr || dataBatch->mutableData()->empty())
+    {
+        return 0;
+    }
+    auto& data = dataBatch->mutableData();
+    tbb::parallel_sort(data->begin(), data->end());
+    auto unique_end = std::unique(data->begin(), data->end());
+    data->erase(unique_end, m_data->end());
+    return data->size();
+}
+
 }  // namespace ppc::psi
diff --git a/cpp/wedpr-computing/ppc-psi/src/cm2020-psi/core/CM2020PSIReceiver.cpp b/cpp/wedpr-computing/ppc-psi/src/cm2020-psi/core/CM2020PSIReceiver.cpp
@@ -908,7 +908,7 @@ void CM2020PSIReceiver::saveResults()
             }
             CM2020_PSI_LOG(INFO) << LOG_BADGE("before dedup") << LOG_KV("taskID", m_taskID)
                                  << LOG_KV("originCount", finalResults->size());
-            m_resultCount = finalResults->dedup();
+            m_resultCount = dedupDataBatch(finalResults);
             CM2020_PSI_LOG(INFO) << LOG_BADGE("after dedup") << LOG_KV("taskID", m_taskID)
                                  << LOG_KV("resultCount", m_resultCount);
             m_taskState->writeLines(finalResults, DataSchema::Bytes);

diff --git a/cpp/wedpr-computing/ppc-psi/src/cm2020-psi/core/CM2020PSISender.cpp b/cpp/wedpr-computing/ppc-psi/src/cm2020-psi/core/CM2020PSISender.cpp
@@ -808,7 +808,7 @@ void CM2020PSISender::saveResults()
             }
             CM2020_PSI_LOG(INFO) << LOG_BADGE("before dedup") << LOG_KV("taskID", m_taskID)
                                  << LOG_KV("originCount", finalResults->size());
-            m_resultCount = finalResults->dedup();
+            m_resultCount = dedupDataBatch(finalResults);
             CM2020_PSI_LOG(INFO) << LOG_BADGE("after dedup") << LOG_KV("taskID", m_taskID)
                                  << LOG_KV("resultCount", m_resultCount);
             m_taskState->writeLines(finalResults, DataSchema::Bytes);

diff --git a/cpp/wedpr-storage/ppc-io/src/CMakeLists.txt b/cpp/wedpr-storage/ppc-io/src/CMakeLists.txt
@@ -1,4 +1,4 @@
 file(GLOB_RECURSE SRCS *.cpp)
 add_library(${IO_TARGET} ${SRCS})
 # Note: the DataBatch depends on tbb
-target_link_libraries(${IO_TARGET} PUBLIC ${BCOS_UTILITIES_TARGET} ${BCOS_BOOSTSSL_TARGET} ${HDFS_LIB} ${CPU_FEATURES_LIB} TBB::tbb)
+target_link_libraries(${IO_TARGET} PUBLIC ${BCOS_UTILITIES_TARGET} ${BCOS_BOOSTSSL_TARGET} ${HDFS_LIB} ${CPU_FEATURES_LIB})