Timing of Dw() and Dwhat()
--------------------------

64x32x32x32 lattice, 4x2x2x2 process grid, 16x16x16x16 local lattice

There are 32 MPI processes
The local size of the gauge field is 18 MB
The local size of a quark field is 6 MB
Using AVX instructions
Assuming SSE prefetch instructions fetch 64 bytes

Lattice parameters:
beta = 5.5
c0 = 1.0, c1 = 0.0
csw = 1.978

Open boundary conditions
cG = 0.55
cF = 0.9012

Time per lattice point for Dw():
0.362 micro sec (5305 Mflops)

Time per lattice point for Dwhat():
0.329 micro sec (5794 Mflops)