dugagjin / parallel-benchmark

different approaches to perform the same thing.

Synopsis

This project try to parallelize the reduction of an array with different approaches. The project is made for Windows.

What is already implemented

• Sequential
• Auto-parallelize
• No vector
• OpenMP
• Threads (with futures)

What has to be implemented

• OpenCL
• CUDA
• SIMD (written but has a bug)

Remarks

• Auto-parallelize from Microsoft does not work - error 1007: reduction of array to scalar - so it's normal.

To run

In developer command line:

cl /EHsc /O2 /Qpar main.cpp && main

Author

Dugagjin Lashi

License

MIT