Optimize reshape idx lmbda lowering

[kaushikcfd]

On main, reshape lowering generates sub-optimal indexing:

(py314_env) kaushikggg@line:~/temp$ cat hello.py 
import pytato as pt


x = pt.make_placeholder("x", (10, 4))
print(pt.generate_loopy(x.reshape(-1)).program)
(py314_env) kaushikggg@line:~/temp$ python hello.py 
---------------------------------------------------------------------------
KERNEL: _pt_kernel
---------------------------------------------------------------------------
ARGUMENTS:
x: type: np:dtype('float64'), shape: (10, 4), dim_tags: (N1:stride:4, N0:stride:1), offset: <class 'loopy.typing.auto'> in aspace: global
_pt_out: type: np:dtype('float64'), shape: (40), dim_tags: (N0:stride:1) out aspace: global
---------------------------------------------------------------------------
DOMAINS:
{  :  }
{ [_pt_out_dim0] : 0 <= _pt_out_dim0 <= 39 }
---------------------------------------------------------------------------
INAME TAGS:
_pt_out_dim0: None
---------------------------------------------------------------------------
INSTRUCTIONS:
for _pt_out_dim0
  _pt_out[_pt_out_dim0] = x[((0 + _pt_out_dim0*1) % 40) // 4, ((0 + _pt_out_dim0*1) % 4) // 1]  {id=_pt_out_store}
end _pt_out_dim0
---------------------------------------------------------------------------

After this patch, the assignment becomes:

for _pt_out_dim0
  _pt_out[_pt_out_dim0] = x[_pt_out_dim0 // 4, _pt_out_dim0 % 4]  {id=_pt_out_store}
end _pt_out_dim0

Notice the reduction in the redundant product with one and modulo with the size operations.