diff --git a/examples/llm.c/export.py b/examples/llm.c/export.py
index ee1e9650e1..eafe4c3998 100755
--- a/examples/llm.c/export.py
+++ b/examples/llm.c/export.py
@@ -5,9 +5,8 @@ from tinygrad import Device, nn, Tensor, dtypes, Variable
 Device.DEFAULT = "CLANG"
 from train_gpt2 import GPT, GPTConfig
 from tinygrad.helpers import dedup, to_function_name, flatten, getenv, GRAPH, GlobalCounters, ansilen, to_function_name
-from tinygrad.engine.schedule import create_schedule
+from tinygrad.engine.schedule import create_schedule, memory_planner
 from tinygrad.engine.realize import get_linearizer, run_schedule
-from tinygrad.engine.memory import memory_planner
 from tinygrad.ops import BufferOps, LoadOps
 
 TIMING = getenv("TIMING")
diff --git a/examples/openpilot/compile2.py b/examples/openpilot/compile2.py
index 678e657764..bfa9bac0a7 100644
--- a/examples/openpilot/compile2.py
+++ b/examples/openpilot/compile2.py
@@ -18,8 +18,7 @@ from tinygrad.dtype import ImageDType
 from tinygrad.device import Buffer
 from tinygrad.helpers import partition, Context, fetch, getenv, DEBUG
 from tinygrad.engine.realize import run_schedule, lower_schedule, ExecItem, CompiledRunner
-from tinygrad.engine.memory import memory_planner
-from tinygrad.engine.schedule import ScheduleItem, create_schedule
+from tinygrad.engine.schedule import ScheduleItem, create_schedule, memory_planner
 from tinygrad.ops import LoadOps
 Device.DEFAULT = "GPU"
 
diff --git a/tinygrad/engine/jit.py b/tinygrad/engine/jit.py
index 3b6d08eafb..02908e8743 100644
--- a/tinygrad/engine/jit.py
+++ b/tinygrad/engine/jit.py
@@ -9,7 +9,7 @@ from tinygrad.dtype import DType
 from tinygrad.shape.shapetracker import ShapeTracker
 from tinygrad.shape.symbolic import Variable, sint
 from tinygrad.engine.realize import ExecItem, capturing, EmptyOp, ViewOp, BufferXfer, CompiledRunner, Runner
-from tinygrad.engine.memory import _internal_memory_planner
+from tinygrad.engine.schedule import _internal_memory_planner
 from tinygrad.nn.state import get_parameters
 from weakref import WeakKeyDictionary
 
diff --git a/tinygrad/engine/memory.py b/tinygrad/engine/memory.py
deleted file mode 100644
index 6a33476d80..0000000000
--- a/tinygrad/engine/memory.py
+++ /dev/null
@@ -1,43 +0,0 @@
-from typing import List, Dict, DefaultDict, Tuple, Union
-from collections import defaultdict
-from tinygrad.dtype import DType
-from tinygrad.device import Buffer
-from tinygrad.helpers import getenv, DEBUG, dedup
-from tinygrad.engine.schedule import ScheduleItem
-
-def _internal_memory_planner(buffers:List[Union[List[Buffer], Tuple[Buffer, ...]]], debug_prefix="") -> Dict[Buffer, Buffer]:
-  if getenv("NO_MEMORY_PLANNER"): return {}
-  last_appearance = {}
-  for i,u in enumerate(buffers):
-    for buf in u: last_appearance[buf] = i
-
-  # LRU algorithm
-  assigned: Dict[Buffer, Buffer] = {}
-  local_cache: DefaultDict[Tuple[str, int, DType], List[Buffer]] = defaultdict(list)
-
-  def handle_buffer(buf):
-    key = (buf.device, buf.size, buf.dtype)
-    if buf not in assigned:
-      if len(ll:=local_cache[key]): assigned[buf] = ll.pop()
-      else: assigned[buf] = Buffer(*key)
-    if i == last_appearance[buf]:
-      if assigned[buf] not in local_cache[key]: local_cache[key].append(assigned[buf])
-
-  for i,u in enumerate(buffers):
-    for buf in u:
-      # all unallocated unparented buffers are fair game to replace
-      if buf.is_allocated() or buf.lb_refcount > 0: continue
-      # handle view buffers
-      if buf._base is not None:
-        assigned[buf] = Buffer(buf.device, buf.size, buf.dtype, base=assigned.get(buf._base, buf._base), offset=buf.offset)
-      else:
-        handle_buffer(buf)
-
-  if DEBUG >= 1 and len(ak:=dedup(assigned.keys())) != len(av:=dedup(assigned.values())):
-    print(debug_prefix+f"memory reduced from {sum([x.nbytes for x in ak])/1e6:.2f} MB -> {sum([x.nbytes for x in av])/1e6:.2f} MB,",
-          f"{len(ak)} -> {len(av)} bufs")
-  return assigned
-
-def memory_planner(schedule:List[ScheduleItem]) -> List[ScheduleItem]:
-  assigned = _internal_memory_planner([si.bufs for si in schedule])
-  return [ScheduleItem(si.ast, tuple(assigned.get(x, x) for x in si.bufs)) for si in schedule]
diff --git a/tinygrad/engine/schedule.py b/tinygrad/engine/schedule.py
index 8a0e00a0f9..d8719b5fb6 100644
--- a/tinygrad/engine/schedule.py
+++ b/tinygrad/engine/schedule.py
@@ -1,12 +1,12 @@
 import sys, pickle, atexit
 from collections import defaultdict, deque
 from dataclasses import dataclass
-from typing import Tuple, List, Dict, Optional, Set, DefaultDict
+from typing import Tuple, List, Dict, Optional, Set, DefaultDict, Union
 from tinygrad.ops import LoadOps, BufferOps, LazyOp, ReduceOps, ConstBuffer, MemBuffer, UNSAFE_PAD_OPS, UnaryOps
 from tinygrad.engine.graph import log_lazybuffer, realized_lazybuffer
 from tinygrad.helpers import GRAPH, DEBUG, MULTIOUTPUT, SAVE_SCHEDULE, GlobalCounters, prod, dedup, all_int, merge_dicts, getenv
 from tinygrad.shape.symbolic import Variable
-from tinygrad.dtype import ImageDType, dtypes
+from tinygrad.dtype import ImageDType, dtypes, DType
 from tinygrad.lazy import LazyBuffer
 from tinygrad.shape.shapetracker import ShapeTracker
 from tinygrad.device import Buffer
@@ -318,3 +318,42 @@ def create_schedule(outs:List[LazyBuffer], seen:Optional[Set[LazyBuffer]]=None)
   schedule, var_vals = create_schedule_with_vars(outs, seen)
   assert len(var_vals) == 0
   return schedule
+
+# *** memory planning ***
+
+def _internal_memory_planner(buffers:List[Union[List[Buffer], Tuple[Buffer, ...]]], debug_prefix="") -> Dict[Buffer, Buffer]:
+  if getenv("NO_MEMORY_PLANNER"): return {}
+  last_appearance = {}
+  for i,u in enumerate(buffers):
+    for buf in u: last_appearance[buf] = i
+
+  # LRU algorithm
+  assigned: Dict[Buffer, Buffer] = {}
+  local_cache: DefaultDict[Tuple[str, int, DType], List[Buffer]] = defaultdict(list)
+
+  def handle_buffer(buf):
+    key = (buf.device, buf.size, buf.dtype)
+    if buf not in assigned:
+      if len(ll:=local_cache[key]): assigned[buf] = ll.pop()
+      else: assigned[buf] = Buffer(*key)
+    if i == last_appearance[buf]:
+      if assigned[buf] not in local_cache[key]: local_cache[key].append(assigned[buf])
+
+  for i,u in enumerate(buffers):
+    for buf in u:
+      # all unallocated unparented buffers are fair game to replace
+      if buf.is_allocated() or buf.lb_refcount > 0: continue
+      # handle view buffers
+      if buf._base is not None:
+        assigned[buf] = Buffer(buf.device, buf.size, buf.dtype, base=assigned.get(buf._base, buf._base), offset=buf.offset)
+      else:
+        handle_buffer(buf)
+
+  if DEBUG >= 1 and len(ak:=dedup(assigned.keys())) != len(av:=dedup(assigned.values())):
+    print(debug_prefix+f"memory reduced from {sum([x.nbytes for x in ak])/1e6:.2f} MB -> {sum([x.nbytes for x in av])/1e6:.2f} MB,",
+          f"{len(ak)} -> {len(av)} bufs")
+  return assigned
+
+def memory_planner(schedule:List[ScheduleItem]) -> List[ScheduleItem]:
+  assigned = _internal_memory_planner([si.bufs for si in schedule])
+  return [ScheduleItem(si.ast, tuple(assigned.get(x, x) for x in si.bufs)) for si in schedule]
diff --git a/tinygrad/tensor.py b/tinygrad/tensor.py
index 34109ea040..fcc665df07 100644
--- a/tinygrad/tensor.py
+++ b/tinygrad/tensor.py
@@ -16,8 +16,7 @@ from tinygrad.device import Buffer, BufferOptions
 from tinygrad.device import Device
 from tinygrad.shape.symbolic import sint, Variable, MulNode, Node
 from tinygrad.engine.realize import run_schedule
-from tinygrad.engine.memory import memory_planner
-from tinygrad.engine.schedule import ScheduleItem, create_schedule_with_vars
+from tinygrad.engine.schedule import ScheduleItem, create_schedule_with_vars, memory_planner
 
 # **** start with two base classes, Tensor and Function ****