diff --git a/loopy/schedule/tools.py b/loopy/schedule/tools.py
index a7860e343..3858462b1 100644
--- a/loopy/schedule/tools.py
+++ b/loopy/schedule/tools.py
@@ -886,7 +886,7 @@ def _raise_loopy_err(x):
     _update_nesting_constraints(relaxed_priorities, warn)
 
     # ordered_loop_nests: A mapping from the unordered loop nests to their
-    # ordered couterparts. For example. If we had only one loop nest
+    # ordered counterparts. For example. If we had only one loop nest
     # `frozenset({"i", "j", "k"})`, and the prioirities said added the
     # constraint that "i" must be nested within "k", then `ordered_loop_nests`
     # would be: `{frozenset({"i", "j", "k"}): ["j", "k", "i"]}` i.e. the loop
diff --git a/loopy/statistics.py b/loopy/statistics.py
index 2d0537fdb..99b163f80 100755
--- a/loopy/statistics.py
+++ b/loopy/statistics.py
@@ -422,16 +422,16 @@ def to_bytes(self):
             bytes_map = get_mem_access_map(knl).to_bytes()
             params = {"n": 512, "m": 256, "l": 128}
 
-            s1_g_ld_byt = bytes_map.filter_by(
+            s1_g_ld_bytes = bytes_map.filter_by(
                                 mtype=["global"], lid_strides={0: 1},
                                 direction=["load"]).eval_and_sum(params)
-            s2_g_ld_byt = bytes_map.filter_by(
+            s2_g_ld_bytes = bytes_map.filter_by(
                                 mtype=["global"], lid_strides={0: 2},
                                 direction=["load"]).eval_and_sum(params)
-            s1_g_st_byt = bytes_map.filter_by(
+            s1_g_st_bytes = bytes_map.filter_by(
                                 mtype=["global"], lid_strides={0: 1},
                                 direction=["store"]).eval_and_sum(params)
-            s2_g_st_byt = bytes_map.filter_by(
+            s2_g_st_bytes = bytes_map.filter_by(
                                 mtype=["global"], lid_strides={0: 2},
                                 direction=["store"]).eval_and_sum(params)
 
diff --git a/loopy/transform/precompute.py b/loopy/transform/precompute.py
index d04fa5b2d..2c91643ac 100644
--- a/loopy/transform/precompute.py
+++ b/loopy/transform/precompute.py
@@ -47,6 +47,7 @@
     RuleAwareIdentityMapper,
     RuleAwareSubstitutionMapper,
     SubstitutionRuleMappingContext,
+    flatten,
     get_dependencies,
 )
 from loopy.transform.array_buffer_map import (
@@ -296,7 +297,7 @@ def map_substitution(self, name, tag, arguments, expn_state):
 
         new_outer_expr = var(self.temporary_name)
         if stor_subscript:
-            new_outer_expr = new_outer_expr.index(tuple(stor_subscript))
+            new_outer_expr = new_outer_expr[tuple(stor_subscript)]
 
         # Can't possibly be nested, and no need to traverse
         # further as compute expression has already been seen
@@ -928,7 +929,7 @@ def add_assumptions(d):
 
         storage_axis_subst_dict[
                 prior_storage_axis_name_dict.get(arg_name, arg_name)] = \
-                        arg+base_index
+                        flatten(arg+base_index)
 
     rule_mapping_context = SubstitutionRuleMappingContext(
             kernel.substitutions, kernel.get_var_name_generator())
diff --git a/loopy/transform/privatize.py b/loopy/transform/privatize.py
index ef878c90f..ca31368d2 100644
--- a/loopy/transform/privatize.py
+++ b/loopy/transform/privatize.py
@@ -43,7 +43,7 @@
 
 # {{{ privatize temporaries with iname
 
-from loopy.symbolic import IdentityMapper
+from loopy.symbolic import IdentityMapper, flatten
 
 
 class ExtraInameIndexInserter(IdentityMapper):
@@ -66,7 +66,7 @@ def map_subscript(self, expr):
 
             self.seen_priv_axis_inames.update(v.name for v in extra_idx)
 
-            new_idx = index + tuple(v - self.iname_to_lbound[v.name]
+            new_idx = index + tuple(flatten(v - self.iname_to_lbound[v.name])
                             for v in extra_idx)
 
             if len(new_idx) == 1:
@@ -81,7 +81,7 @@ def map_variable(self, expr):
         else:
             self.seen_priv_axis_inames.update(v.name for v in new_idx)
 
-            new_idx = tuple(v - self.iname_to_lbound[v.name]
+            new_idx = tuple(flatten(v - self.iname_to_lbound[v.name])
                             for v in new_idx)
 
             if len(new_idx) == 1:
diff --git a/loopy/transform/realize_reduction.py b/loopy/transform/realize_reduction.py
index 8aea6541a..7d1f3c870 100644
--- a/loopy/transform/realize_reduction.py
+++ b/loopy/transform/realize_reduction.py
@@ -711,7 +711,7 @@ def _add_to_depends_on(insn_id, new_depends_on_params):
 
             needs_replacement = True
 
-            # {{{ generate a new assignent instruction
+            # {{{ generate a new assignment instruction
 
             new_assignee_name = var_name_gen(
                     "{insn_id}_retval_{assignee_nr}"