Hi, I'm use GetDP (on Ubuntu 13.04) for solving large problems. GetDP was configured with PETSc and MUMPS. Usually, I use LU-decomposition as preconditioner for GMRES. For this, as default, using BLAS. In this time I try to use OpenBLAS (GotoBLAS) instead BLAS, it so much faster, but ... Can I use GPU (HD7770) with APPML (clBLAS) instead BLAS(OpenBLAS/ATLAS) and CPU? If yes, how can I replace BLAS with APPML?
The Accelerated Parallel Processing Math Library. APPML contains an OpenCL implementation of BLAS and FFT routines. The library enables developers to accelerate common scientific and engineering computations on APUs and discrete graphics accelerators.
Please check the below link for the API. which may help you in replacing BLAS APIs to APPML APIs