Cooperative High-Performance Computing With Fpgas - Matrix Multiply Case-Study