Stencil benchmarking & Test refactoring

[samkellerhals]

Enables stencil by stencil benchmarking for different gt4py backends and meshes in unit tests. Benchmarks can be disabled by passing --benchmark-disable to pytest.

Edit (@tehrengruber):
As part of this PR many tests have been refactored to use the new StencilTest base class for writing test cases.

Example (pseudo-code):

class TestMultiplyByTwo(StencilTest): # doctest: +SKIP
    PROGRAM = multiply_by_two  # noqa: F821
    OUTPUTS = ("some_output",)

    @pytest.fixture
    def input_data(self):
        return {"some_input": ..., "some_output": ...}

    @staticmethod
    def reference(some_input, **kwargs):
        return dict(some_output=np.asarray(some_input)*2)

Test cases still to be refactored:

Sparse Field Tests (need more work):

• test_calculate_nabla2_of_theta
• test_calculate_nabla4
• test_mo_advection_traj_btraj_compute_o1_dsl.py
• mo_velocity_advection_stencil_19.py

Custom test patterns:

• test_mo_nh_diffusion_stencil_15.py
• test_mo_solve_nonhydro_stencil_20.py
• test_mo_solve_nonhydro_stencil_21.py
• mo_velocity_advection_stencil_02.py
• mo_velocity_advection_stencil_16.py

np.roll stencils:

• mo_velocity_advection_stencil_03.py
• mo_velocity_advection_stencil_06.py
• mo_velocity_advection_stencil_10.py
• mo_velocity_advection_stencil_20.py
• test_mo_solve_nonhydro_stencil_05.py
• test_mo_solve_nonhydro_stencil_08.py
• test_mo_solve_nonhydro_stencil_09.py
• test_mo_solve_nonhydro_stencil_10.py
• test_mo_solve_nonhydro_stencil_11_upper.py
• test_mo_solve_nonhydro_stencil_36.py
• test_mo_solve_nonhydro_stencil_38.py
• test_mo_solve_nonhydro_stencil_39.py
• test_mo_solve_nonhydro_stencil_40.py
• test_mo_solve_nonhydro_stencil_51.py

more complex cases:

• test_mo_solve_nonhydro_stencil_16_fused_btraj_traj_o1.py
• test_mo_solve_nonhydro_stencil_52.py
• test_truly_horizontal_diffusion_nabla_of_theta_over_steep_points.py

[tehrengruber]

Just small things.