In addition to the HPX-based shared memory parallelization, starting with Blaze 2.1, Blaze also provides a shared memory parallelization based on C++11 threads.

C++11 Thread Setup

In order to enable the C++11 thread-based parallelization, first the according C++11-specific compiler flags have to be used and second the BLAZE_USE_CPP_THREADS command line argument has to be explicitly specified. For instance, in case of the GNU C++ and Clang compilers the compiler flags have to be extended by

... -std=c++11 -DBLAZE_USE_CPP_THREADS ...

This simple action will cause the Blaze library to automatically try to run all operations in parallel with the specified number of C++11 threads. Note that in case both HPX and C++11 threads are enabled on the command line, the HPX-based parallelization has priority and is preferred.

The number of threads can be either specified via the environment variable BLAZE_NUM_THREADS

export BLAZE_NUM_THREADS=4  // Unix systems
set BLAZE_NUM_THREADS=4     // Windows systems

or alternatively via the setNumThreads() function provided by the Blaze library:

blaze::setNumThreads( 4 );

Please note that the Blaze library does not limit the available number of threads. Therefore it is in YOUR responsibility to choose an appropriate number of threads. The best performance, though, can be expected if the specified number of threads matches the available number of cores.

In order to query the number of threads used for the parallelization of operations, the getNumThreads() function can be used:

const size_t threads = blaze::getNumThreads();

In the context of C++11 threads, the function will return the previously specified number of threads.

C++11 Thread Configuration

As in case of the OpenMP-based parallelization Blaze is not unconditionally running an operation in parallel. In case Blaze deems the parallel execution as counterproductive for the overall performance, the operation is executed serially. One of the main reasons for not executing an operation in parallel is the size of the operands. For instance, a vector addition is only executed in parallel if the size of both vector operands exceeds a certain threshold. Otherwise, the performance could seriously decrease due to the overhead caused by the thread setup. However, in order to be able to adjust the Blaze library to a specific system, it is possible to configure these thresholds manually. All thresholds are contained within the configuration file <blaze/config/Thresholds.h>.

Please note that these thresholds are highly sensitiv to the used system architecture and the shared memory parallelization technique. Therefore the default values cannot guarantee maximum performance for all possible situations and configurations. They merely provide a reasonable standard for the current CPU generation. Also note that the provided defaults have been determined using the OpenMP parallelization and require individual adaption for the C++11 thread parallelization.

Known Issues

There is a known issue in Visual Studio 2012 and 2013 that may cause C++11 threads to hang if their destructor is executed after the main() function. Unfortunately, the C++11 parallelization of the Blaze library is affected from this bug. In order to circumvent this problem, Blaze provides the shutDownThreads() function, which can be used to manually destroy all threads at the end of the main() function:

int main()
{
   // ... Using the C++11 thread parallelization of Blaze

   shutDownThreads();
}

Please note that this function may only be used at the end of the main() function. After this function no further computation may be executed! Also note that this function has an effect for Visual Studio compilers only and doesn't need to be used with any other compiler.

Previous: HPX Parallelization ---- Next: Boost Thread Parallelization