Better default schedule/storage type inference

[tbennun]

(Fixes #1185)

Rewrite the default schedule/storage type inference based on the discussion in #1185. The schedule type of a scope (e.g., a Map) is now also determined by the surrounding storage. If the surrounding storage is ambiguous, dace will fail with a nice exception. This means that:

• Codes such as the one below:

@dace.program
def add(a: dace.float32[10, 10] @ dace.StorageType.GPU_Global, 
        b: dace.float32[10, 10] @ dace.StorageType.GPU_Global):
    return a + b @ b

will now automatically run the + and @ operators on the GPU.

• Users can now write codes such as

@dace.program(device=dace.DeviceType.GPU)
def myprogram(a: dace.float64[20], b: dace.float64[20]):
    a += b
    for i in dace.map[0:20]:
        out >> b[i]
        out = 5

without GPUTransform. The arrays a and b will automatically be set to reside on the GPU, and the map will be scheduled as a GPU_Device map.

The often-confusing GPU_Default schedule has now been removed as well.