59a0a3844c
PR Test AMD / cancel-on-close (push) Has been skipped
PR Test NVIDIA ARM / scan (push) Has been skipped
PR Test NVIDIA / cancel-on-close (push) Has been skipped
PR Test AMD / scan (push) Has been skipped
PR Test NVIDIA ARM / cancel-on-close (push) Has been skipped
PR Test NVIDIA / scan (push) Has been skipped
Release Docker Images / build (cu129-torch-2.11.0) (push) Has been skipped
Release Docker Images / build (cu130-torch-2.11.0) (push) Has been skipped
Release PyPI / publish (push) Has been skipped
Scheduler Python Test / test (push) Successful in 27m19s
Docs / build (push) Successful in 28m8s
Scheduler C++ Test / test (push) Successful in 28m19s
Scheduler C++ Test / test-flat (push) Successful in 28m18s
Docs / deploy (push) Has been cancelled
PR Test AMD / finish (push) Has been cancelled
PR Test NVIDIA / finish (push) Has been cancelled
PR Test NVIDIA ARM / finish (push) Has been cancelled
PR Test NVIDIA ARM / ${{ matrix.name }} (${{ matrix.runner }}) (push) Has been cancelled
PR Test AMD / ${{ matrix.name }} (${{ matrix.runner }}) (push) Has been cancelled
PR Test NVIDIA / ${{ matrix.name }} (${{ matrix.runner }}) (push) Has been cancelled
188 lines
6.4 KiB
Python
188 lines
6.4 KiB
Python
# Copyright (c) 2026 LightSeek Foundation
|
|
#
|
|
# Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
# of this software and associated documentation files (the "Software"), to deal
|
|
# in the Software without restriction, including without limitation the rights
|
|
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
# copies of the Software, and to permit persons to whom the Software is
|
|
# furnished to do so, subject to the following conditions:
|
|
#
|
|
# The above copyright notice and this permission notice shall be included in
|
|
# all copies or substantial portions of the Software.
|
|
#
|
|
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
# SOFTWARE.
|
|
|
|
"""Unit tests for EventLoop._handle_flat_oom_terminals.
|
|
|
|
The C++ flat scheduler terminalizes a never-fits request internally
|
|
(AbortEvent; the reaper reclaims its resources) and reports the id on
|
|
``ExecutionPlan.flat_oom_request_ids``. The Python event loop must stream a
|
|
terminating abort finish to the client — otherwise the client hangs forever
|
|
— and must NOT send a ForwardEvent.Abort back (the request is already gone
|
|
scheduler-side).
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
from tokenspeed.runtime.engine.event_loop import EventLoop
|
|
from tokenspeed.runtime.engine.generation_output_processor import (
|
|
OutputProcesser,
|
|
RequestState,
|
|
)
|
|
from tokenspeed.runtime.engine.request_types import FINISH_ABORT
|
|
from tokenspeed.runtime.sampling.sampling_params import SamplingParams
|
|
|
|
|
|
class _Sender:
|
|
def __init__(self):
|
|
self.items = []
|
|
|
|
def send_pyobj(self, obj):
|
|
self.items.append(obj)
|
|
|
|
|
|
class _Tokenizer:
|
|
eos_token_id = None
|
|
additional_stop_token_ids = None
|
|
|
|
def decode(self, ids):
|
|
return "".join(str(i) for i in ids)
|
|
|
|
|
|
class _Metrics:
|
|
enabled = False
|
|
|
|
def record_nan_abort(self):
|
|
pass
|
|
|
|
|
|
class _Plan:
|
|
"""ExecutionPlan stub carrying only the flat OOM binding field."""
|
|
|
|
def __init__(self, flat_oom_request_ids):
|
|
self.flat_oom_request_ids = flat_oom_request_ids
|
|
|
|
|
|
class _RadixPlan:
|
|
"""ExecutionPlan stub WITHOUT the field (older/radix binding shape)."""
|
|
|
|
|
|
def _state(input_ids: list[int]) -> RequestState:
|
|
return RequestState(
|
|
prompt_input_ids=input_ids,
|
|
sampling_params=SamplingParams(max_new_tokens=8, stop=[], ignore_eos=True),
|
|
stream=False,
|
|
tokenizer=_Tokenizer(),
|
|
)
|
|
|
|
|
|
def _make_loop(sender: _Sender) -> EventLoop:
|
|
# The handler only touches self.output_processor; skip full engine init.
|
|
loop = EventLoop.__new__(EventLoop)
|
|
loop.output_processor = OutputProcesser(sender, attn_tp_rank=0, metrics=_Metrics())
|
|
return loop
|
|
|
|
|
|
def test_flat_oom_terminal_streams_abort_finish_to_client():
|
|
sender = _Sender()
|
|
loop = _make_loop(sender)
|
|
state = _state([1, 2, 3, 4])
|
|
loop.output_processor.rid_to_state["oom"] = state
|
|
|
|
loop._handle_flat_oom_terminals(_Plan(["oom"]))
|
|
|
|
# Request finished with an error abort carrying the OOM message.
|
|
assert isinstance(state.finished_reason, FINISH_ABORT)
|
|
assert state.finished_reason.is_error
|
|
assert "cannot fit" in state.finished_reason.message
|
|
|
|
# The finish was streamed to the client and the state dropped from
|
|
# tracking (publish_finished_at_admission pops it).
|
|
assert "oom" not in loop.output_processor.rid_to_state
|
|
assert len(sender.items) == 1
|
|
out = sender.items[0]
|
|
idx = out.rids.index("oom")
|
|
reason = out.finished_reasons[idx]
|
|
assert reason["type"] == "abort"
|
|
assert "cannot fit" in reason["message"]
|
|
|
|
|
|
def test_flat_oom_terminal_leaves_other_requests_running():
|
|
sender = _Sender()
|
|
loop = _make_loop(sender)
|
|
oom_state = _state([1, 2, 3, 4])
|
|
live_state = _state([5, 6, 7])
|
|
loop.output_processor.rid_to_state["oom"] = oom_state
|
|
loop.output_processor.rid_to_state["live"] = live_state
|
|
|
|
loop._handle_flat_oom_terminals(_Plan(["oom"]))
|
|
|
|
assert not live_state.finished
|
|
assert "live" in loop.output_processor.rid_to_state
|
|
|
|
|
|
def test_flat_oom_terminal_missing_rid_is_skipped():
|
|
sender = _Sender()
|
|
loop = _make_loop(sender)
|
|
|
|
# rid already gone (e.g. client abort raced ahead): no crash, no stream.
|
|
loop._handle_flat_oom_terminals(_Plan(["ghost"]))
|
|
|
|
assert sender.items == []
|
|
|
|
|
|
def test_flat_oom_terminal_already_finished_state_is_reaped():
|
|
sender = _Sender()
|
|
loop = _make_loop(sender)
|
|
state = _state([1, 2, 3])
|
|
state.set_finish_with_abort("earlier abort")
|
|
loop.output_processor.rid_to_state["done"] = state
|
|
|
|
loop._handle_flat_oom_terminals(_Plan(["done"]))
|
|
|
|
# Not re-finished, not streamed (client-initiated aborts tore down their
|
|
# own stream), message unchanged — but the state must be reaped: C++
|
|
# reports the rid exactly once, no future op will pop it.
|
|
assert state.finished_reason.message == "earlier abort"
|
|
assert sender.items == []
|
|
assert "done" not in loop.output_processor.rid_to_state
|
|
|
|
|
|
def test_flat_oom_terminal_already_finished_notify_client_is_published():
|
|
sender = _Sender()
|
|
loop = _make_loop(sender)
|
|
state = _state([1, 2, 3])
|
|
# Pause-initiated abort: the passive client still waits on the stream.
|
|
state.set_finish_with_abort("pause abort", notify_client=True)
|
|
loop.output_processor.rid_to_state["paused"] = state
|
|
|
|
loop._handle_flat_oom_terminals(_Plan(["paused"]))
|
|
|
|
# The existing finish is streamed once and the state reaped; the OOM
|
|
# terminal must not overwrite the earlier finish reason.
|
|
assert state.finished_reason.message == "pause abort"
|
|
assert "paused" not in loop.output_processor.rid_to_state
|
|
assert len(sender.items) == 1
|
|
out = sender.items[0]
|
|
idx = out.rids.index("paused")
|
|
assert out.finished_reasons[idx]["message"] == "pause abort"
|
|
|
|
|
|
def test_plan_without_flat_oom_field_is_noop():
|
|
sender = _Sender()
|
|
loop = _make_loop(sender)
|
|
state = _state([1, 2, 3])
|
|
loop.output_processor.rid_to_state["r"] = state
|
|
|
|
loop._handle_flat_oom_terminals(_RadixPlan())
|
|
loop._handle_flat_oom_terminals(_Plan([]))
|
|
|
|
assert not state.finished
|
|
assert sender.items == []
|