InverseMatrix

Inverting Matrices in Parallel with Eigen and C++

Matrix Definition Methodology

• Given 7 numbers choose 4
• From chosen 4 numbers choose 3
• Combine both selections to form 7 number combination
• Each row in the matrix is a permutation of the original combination
• Create a Matrix<float, 7, 7> matrix from the combinations

BDD/TDD using Catch

• Single file inclusion that allows a very descriptive syntax to write tests
• Supported out of the box with CLion 2017

Eigen Explicit Vectorization / Matrix Decomposition (Leveraging SIMD - SSE, AVX512 etc)

• Within-Thread Parallelization due to memory-aligned data structures and ability to utilize SIMD instructions
• Smart matrix decomposition used for inversion calculation i.e. LU and QR

STL Async for Parallelism

• Running the inversions in parallel without having to explicitly create STL::Thread
• Exception handled concurrency constructs