derekeverett/cpu-vh

A viscous hydrodynamic code for the description of ultrarelativistic heavy ion collisions. Accelerated via OpenMP.

cpu-vh's Issues

NS initial conditions

check default formula for navier stokes initial conditions

Classes and Wrappers

Begun writing a simple c++ wrapper class for cpu-vh. We need additional class methods for the setting of parameters, passing in initial conditions, reading out the freeze out surface and spacetime history, etc...
Can we use the same wrappers for gpu-vh and cpu-vah also?

libconfig

Can we find a simple way to parse the parameters files without requiring libconfig? If so, try removing libconfig dependence... If not, so be it.

vorticity tensor

check terms in vorticity tensor in SourceTerms.cpp

CMake

Replace all Makefiles for different systems (Linux, Mac,...) and different installs of libconfig with one CMake file.

approximate derivative

add option for use of approximate derivative in calculation of source terms (SourceTerms.cpp)

Libconfig?

necessary?

Cache Tiling

Investigate the work required to employ tiling/cache blocking to reduce cache misses

https://software.intel.com/en-us/articles/cache-blocking-techniques

https://software.intel.com/en-us/articles/efficient-use-of-tiling

Vectorization

Check if we can good get good acceleration via vector instructions by either explicit (omp directives) and/or compiler flags (-march=...)
See functions in FullyDiscretexxx.cpp etc...