SharedKernelContext Class Reference

#include <ExecutionKernel.h>

Collaboration diagram for SharedKernelContext:

Public Member Functions
	SharedKernelContext (const std::vector< InputTableInfo > &query_infos)
const std::vector< uint64_t > &	getFragOffsets ()
void	addDeviceResults (ResultSetPtr &&device_results, std::vector< size_t > outer_table_fragment_ids)
std::vector< std::pair < ResultSetPtr, std::vector < size_t > > > &	getFragmentResults ()
const std::vector < InputTableInfo > &	getQueryInfos () const
void	setNumAllocatedThreads (size_t num_threads)
size_t	getNumAllocatedThreads ()

Public Attributes
std::atomic_flag	dynamic_watchdog_set = ATOMIC_FLAG_INIT

Private Attributes
std::mutex	reduce_mutex_
std::vector< std::pair < ResultSetPtr, std::vector < size_t > > >	all_fragment_results_
std::vector< uint64_t >	all_frag_row_offsets_
std::mutex	all_frag_row_offsets_mutex_
const std::vector < InputTableInfo > &	query_infos_
const RegisteredQueryHint	query_hint_
size_t	num_allocated_threads_ {1}

Detailed Description

Definition at line 29 of file ExecutionKernel.h.

Constructor & Destructor Documentation

SharedKernelContext::SharedKernelContext ( const std::vector< InputTableInfo > &  query_infos )

inline

Definition at line 31 of file ExecutionKernel.h.

      : query_infos_(query_infos)
#ifdef HAVE_TBB
      , task_group_(nullptr)
#endif
  {
  }

Member Function Documentation

void SharedKernelContext::addDeviceResults	(	ResultSetPtr &&	device_results,
		std::vector< size_t >	outer_table_fragment_ids
	)

Definition at line 115 of file ExecutionKernel.cpp.

References all_fragment_results_, anonymous_namespace{ExecutionKernel.cpp}::needs_skip_result(), and reduce_mutex_.

Referenced by Executor::launchKernelsImpl(), and ExecutionKernel::runImpl().

                                                                                       {
  std::lock_guard<std::mutex> lock(reduce_mutex_);
  if (!needs_skip_result(device_results)) {
    all_fragment_results_.emplace_back(std::move(device_results),
                                       outer_table_fragment_ids);
  }
}

Here is the call graph for this function:

Here is the caller graph for this function:

std::vector< std::pair< ResultSetPtr, std::vector< size_t > > > & SharedKernelContext::getFragmentResults

(

)

Definition at line 125 of file ExecutionKernel.cpp.

References all_fragment_results_.

Referenced by Executor::collectAllDeviceResults(), Executor::collectAllDeviceShardedTopResults(), Executor::executeUpdate(), Executor::executeWorkUnitImpl(), Executor::executeWorkUnitPerFragment(), and Executor::resultsUnion().

                                        {
  return all_fragment_results_;
}

Here is the caller graph for this function:

const std::vector< uint64_t > & SharedKernelContext::getFragOffsets

(

)

Definition at line 102 of file ExecutionKernel.cpp.

References all_frag_row_offsets_, all_frag_row_offsets_mutex_, and query_infos_.

Referenced by Executor::createKernels(), Executor::executeUpdate(), and ExecutionKernel::runImpl().

                                                               {
  std::lock_guard<std::mutex> lock(all_frag_row_offsets_mutex_);
  if (all_frag_row_offsets_.empty()) {
    all_frag_row_offsets_.resize(query_infos_.front().info.fragments.size() + 1);
    for (size_t i = 1; i <= query_infos_.front().info.fragments.size(); ++i) {
      all_frag_row_offsets_[i] =
          all_frag_row_offsets_[i - 1] +
          query_infos_.front().info.fragments[i - 1].getNumTuples();
    }
  }
  return all_frag_row_offsets_;
}

Here is the caller graph for this function:

size_t SharedKernelContext::getNumAllocatedThreads ( )

inline

Definition at line 54 of file ExecutionKernel.h.

References num_allocated_threads_.

Referenced by anonymous_namespace{ExecutionKernel.cpp}::get_available_cpu_threads_per_task().

                                  {
    return num_allocated_threads_;
  }

Here is the caller graph for this function:

const std::vector<InputTableInfo>& SharedKernelContext::getQueryInfos ( ) const

inline

Definition at line 46 of file ExecutionKernel.h.

References query_infos_.

Referenced by Executor::launchKernelsViaResourceMgr(), and ExecutionKernel::runImpl().

                                                         {
    return query_infos_;
  }

Here is the caller graph for this function:

void SharedKernelContext::setNumAllocatedThreads ( size_t  num_threads )

inline

Definition at line 50 of file ExecutionKernel.h.

References num_allocated_threads_.

Referenced by Executor::launchKernelsImpl().

                                                  {
    num_allocated_threads_ = num_threads;
  }

Here is the caller graph for this function:

Member Data Documentation

std::vector<uint64_t> SharedKernelContext::all_frag_row_offsets_

private

Definition at line 76 of file ExecutionKernel.h.

Referenced by getFragOffsets().

std::mutex SharedKernelContext::all_frag_row_offsets_mutex_

private

Definition at line 77 of file ExecutionKernel.h.

Referenced by getFragOffsets().

std::vector<std::pair<ResultSetPtr, std::vector<size_t> > > SharedKernelContext::all_fragment_results_

private

Definition at line 74 of file ExecutionKernel.h.

Referenced by addDeviceResults(), and getFragmentResults().

std::atomic_flag SharedKernelContext::dynamic_watchdog_set = ATOMIC_FLAG_INIT

Definition at line 58 of file ExecutionKernel.h.

Referenced by ExecutionKernel::runImpl().

size_t SharedKernelContext::num_allocated_threads_ {1}

private

Definition at line 83 of file ExecutionKernel.h.

Referenced by getNumAllocatedThreads(), and setNumAllocatedThreads().

const RegisteredQueryHint SharedKernelContext::query_hint_

private

Definition at line 79 of file ExecutionKernel.h.

const std::vector<InputTableInfo>& SharedKernelContext::query_infos_

private

Definition at line 78 of file ExecutionKernel.h.

Referenced by getFragOffsets(), and getQueryInfos().

std::mutex SharedKernelContext::reduce_mutex_

private

Definition at line 73 of file ExecutionKernel.h.

Referenced by addDeviceResults().

---

The documentation for this class was generated from the following files:

• /home/jenkins-slave/workspace/core-os-doxygen/QueryEngine/ExecutionKernel.h
• /home/jenkins-slave/workspace/core-os-doxygen/QueryEngine/ExecutionKernel.cpp