Poor scaling with many calls to add_prefetch

Even with #755, attempting to prefetch many arrays scales poorly. By the 19th add_prefetch operation it takes around 5 seconds for add_prefetch to complete on one fused Mirgecom kernels with 100+ einsums. Profiling shows a lot of time is spent in `get_grid_sizes_for_insn_ids_as_dicts`.

```
add_prefetch time
0.10026049613952637
0.11436939239501953
0.13449859619140625
0.15656232833862305
0.18714141845703125
0.22336697578430176
0.2757580280303955
0.34452342987060547
0.4413919448852539
0.5716948509216309
0.7412521839141846
0.9670014381408691
1.2657301425933838
1.6620988845825195
2.180263042449951
2.8522701263427734
3.696044683456421
4.755332946777344
6.091721773147583
```

```
452893 function calls (368789 primitive calls) in 8.962 seconds

Ordered by: cumulative time
List reduced from 557 to 30 due to restriction <30>

ncalls tottime percall cumtime percall filename:lineno(function)
1 0.000 0.000 8.962 8.962 __init__.py:779(prefetch_and_project)
1 0.000 0.000 8.837 8.837 data.py:302(add_prefetch)
1 0.000 0.000 8.837 8.837 data.py:153(add_prefetch_for_single_kernel)
1 0.001 0.001 8.784 8.784 precompute.py:353(precompute_for_single_kernel)
2879/1595 0.003 0.000 8.018 0.005 __init__.py:752(wrapper)
5/1 0.000 0.000 7.958 7.958 tools.py:800(assign_automatic_axes)
44 0.002 0.000 7.890 0.179 __init__.py:802(get_iname_bounds)
1 0.000 0.000 7.818 7.818 __init__.py:1031(get_grid_size_upper_bounds_as_exprs)
1 0.000 0.000 7.818 7.818 __init__.py:990(get_grid_sizes_for_insn_ids_as_exprs)
1 0.000 0.000 7.817 7.817 __init__.py:939(get_grid_sizes_for_insn_ids)
1 0.002 0.002 7.817 7.817 __init__.py:845(get_grid_sizes_for_insn_ids_as_dicts)
6948 7.439 0.001 7.482 0.001 __init__.py:925(wrapper)
92 0.001 0.000 7.469 0.081 tools.py:352(op)
46 0.000 0.000 3.740 0.081 tools.py:364(dim_min)
46 0.038 0.001 3.729 0.081 tools.py:343(_get_dim_min)
46 0.000 0.000 3.729 0.081 tools.py:370(dim_max)
46 0.043 0.001 3.719 0.081 tools.py:339(_get_dim_max)
68309/37887 0.635 0.000 0.656 0.000 __init__.py:936(wrapper)
1 0.001 0.001 0.512 0.512 array_buffer_map.py:196(__init__)
154 0.002 0.000 0.479 0.003 __init__.py:1263(align_spaces)
54 0.000 0.000 0.467 0.009 __init__.py:1312(align_two)
1 0.000 0.000 0.467 0.467 array_buffer_map.py:173(compute_bounds)
1 0.000 0.000 0.456 0.456 array_buffer_map.py:162(find_var_base_indices_and_shape_from_inames)
1 0.000 0.000 0.456 0.456 array_buffer_map.py:165()
2 0.000 0.000 0.456 0.228 tools.py:379(base_index_and_length)
462 0.069 0.000 0.434 0.001 __init__.py:1182(_align_dim_type)
81 0.001 0.000 0.321 0.004 __init__.py:801(expr_like_add)
84 0.313 0.004 0.320 0.004 __init__.py:769(_number_to_expr_like)
6773 0.031 0.000 0.220 0.000 __init__.py:1178(_set_dim_id)
49 0.000 0.000 0.215 0.004 __init__.py:1061(obj_project_out_except)
```