# oMLX OMLX_DISABLE_PRESSURE_RECLAIM and request_pressure_reclaim scheduler path > Whether 5 grace polls cover a slow prefill chunk, since the chunk-boundary drain waits for the end of the current chunk. Parent: [Mac local LLMs: oMLX, Rapid-MLX and related internals](https://llms-explorer.com/tree/mac-local-llms-omlx-and-rapid-mlx-internals/) · 1 facets · 12 facts · page: https://llms-explorer.com/tree/omlx-omlx-disable-pressure-reclaim-and-request-p/ ## Facts - Whether 5 grace polls cover a slow prefill chunk, since the chunk-boundary drain waits for the end of the current chunk. — source: `asserted` - `Scheduler` keeps `_pending_reclaim_request` for idle reclaim and `_pending_pressure_clear` for hard-pressure reclaim; the second "drains at the next step boundary even under load". — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/scheduler.py) - `request_pressure_reclaim` only sets `_pending_pressure_clear = True`, and its docstring says the actual `_sync_and_clear_cache` runs on the inference thread at the next `step()` boundary. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/scheduler.py) - `_consume_pressure_clear` returns False when the flag is unset, otherwise resets it and returns True, and it is deliberately not idle-gated. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/scheduler.py) - `has_requests()` includes `_pending_pressure_clear`, so an idle engine loop keeps calling `step()` until the clear fires. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/scheduler.py) - The external prefill loop calls `_consume_pressure_clear()` after each chunk and then `Scheduler._clear_cache(self)`, because a single `step()` can last minutes. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/scheduler.py) - A comment at that site says the abort escalation was measured firing 0.1 GB over the watermark with 3.7 GB pooled. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/scheduler.py) - `Scheduler._clear_cache` calls `_sync_and_clear_cache(self._stream)` and records the pooled bytes for the prefill transient tracker when flat-overhead tracking is on. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/scheduler.py) - `_process_pending_reclaim` returns without consuming the idle request while any request is running, prefilling or waiting, and cites issue 2179. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/scheduler.py) - The enforcer's grace gate requires level not ok, not emergency, `OMLX_DISABLE_PRESSURE_RECLAIM != "1"` and fewer than `_PRESSURE_RECLAIM_GRACE_POLLS_MAX` (5) polls, and returns early after a successful request. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/process_memory_enforcer.py) - At hard level the enforcer calls `_request_scheduler_cache_reclaim(freed_hot)` after the hot-cache shrink unless the env var is "1", and then recomputes the pressure level from current usage. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/process_memory_enforcer.py) - `_request_scheduler_cache_reclaim` calls `request_pressure_reclaim` on each pooled engine's scheduler, tolerates exceptions per scheduler, and returns the count of accepted requests. — [source](https://raw.githubusercontent.com/jundot/omlx/main/omlx/process_memory_enforcer.py)