Implementing the gridding algorithm from the SKA Science Data Processor in Accelerate, a domain specific language for high performance computing embedded in Haskell.
- Make sure you have a version 1.10.0 of hdf5 installed. (The development library). (Other versions might, work but I didn't test with them)
- Also alter the "extra-lib-dirs:" and "include-dirs" in package.yaml to point to the right directories for hdf5. For my ubuntu this is: "extra-lib-dirs: /usr/lib/x86_64-linux-gnu/hdf5/serial" and "include-dirs: /usr/include/hdf5/serial"
- When building and executing on ubuntu, I ran into the following linking bug for libffi. To fix it, I created the symbolic link libffi.so.7 pointing to libffi.so.6 in the systems lib directory
- The general accelerate instalation is used, with the CPU backend. The backend is build with LLVM 7.0, so install that one aswell. See Accelerate for more details.