Parallel Computing question

Hi! Can anyone help me with this.
I am trying to find out arithmetic intensity of following code, whether the vector is amenable to vector or SIMD execution and if the kernel is executed on a processor with 30GB/sec memory bandwidth will it be memory bound or compute bound.

The code given is:
for (int x=0 x<NX-1 x++)
{ for (int y=0, y<NY-1 ,y++)
{ for (int Z=0, z<NZ-1, z++)
int index xNYNZ + y*NZ +z ;
if ( y>0 && x>0)
{
material=IDx [index];
dH1=(Hz[index]-Hz[index-incrementY]/dy[y];
dh2=(Hy[index]-Hy[index-incrementz]/dz[z];
Ex[index]= ca[material]Ex[index]+cb[material](dh2-dh1);
}}}
Assume that dH1,dH2,Hy,Hz,dy,dz,Ca,Cb, and Ex are all single-precision floating-point arrays.Assume IDx is an array of unsigned int

This is more lookup bound it seems, thus memory latency/hiding will play a greater role.