Skip to content

Commit c93b48b

Browse files
DwyaneShiHaiyang Shi
DwyaneShi
and
Haiyang Shi
authored
[Feature] AIBrix KVCache common (vllm-project#1057)
- project init - common module - memory module - spec/status/cache handle --------- Signed-off-by: Haiyang Shi <[email protected]> Co-authored-by: Haiyang Shi <[email protected]>
1 parent b835bab commit c93b48b

24 files changed

+3728
-3
lines changed

‎.github/workflows/python-aibrix-kvcache-tests.yml‎

Lines changed: 45 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,45 @@
1+
name: Python AIBrix KVCache Tests
2+
3+
on:
4+
push:
5+
branches: [ "main", "release-*" ]
6+
paths:
7+
- 'python/aibrix_kvcache/**'
8+
pull_request:
9+
branches: [ "main" ]
10+
paths:
11+
- 'python/aibrix_kvcache/**'
12+
jobs:
13+
lint:
14+
runs-on: ubuntu-latest
15+
strategy:
16+
matrix:
17+
python-version: ["3.10", "3.11", "3.12"]
18+
name: Lint
19+
steps:
20+
- name: Check out source repository
21+
uses: actions/checkout@v4
22+
- name: Set up Python environment ${{ matrix.python-version }}
23+
uses: actions/setup-python@v5
24+
with:
25+
python-version: ${{ matrix.python-version }}
26+
- name: Install dependencies
27+
run: |
28+
cd python/aibrix_kvcache
29+
python -m pip install --upgrade pip
30+
pip install -U pip poetry
31+
poetry config virtualenvs.create false
32+
poetry install --no-root --with dev
33+
- name: Run Ruff
34+
run: |
35+
cd python/aibrix_kvcache
36+
python -m ruff check .
37+
python -m ruff format --check .
38+
- name: Run mypy
39+
run: |
40+
cd python/aibrix_kvcache
41+
python -m mypy .
42+
- name: Run Test
43+
run: |
44+
cd python/aibrix_kvcache/tests
45+
python -m pytest .

.github/workflows/python-tests.yml renamed to .github/workflows/python-aibrix-tests.yml

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,14 +1,14 @@
1-
name: Python Tests
1+
name: Python AIBrix Tests
22

33
on:
44
push:
55
branches: [ "main", "release-*" ]
66
paths:
7-
- 'python/**'
7+
- 'python/aibrix/**'
88
pull_request:
99
branches: [ "main" ]
1010
paths:
11-
- 'python/**'
11+
- 'python/aibrix/**'
1212
jobs:
1313
lint:
1414
runs-on: ubuntu-latest

‎python/aibrix_kvcache/.gitignore‎

Lines changed: 12 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,12 @@
1+
# python
2+
__pycache__
3+
4+
# pytest
5+
.benchmarks
6+
.pytest_cache
7+
8+
# ruff
9+
.ruff_cache
10+
11+
# mypy
12+
.mypy_cache

‎python/aibrix_kvcache/LICENSE‎

Lines changed: 201 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,201 @@
1+
Apache License
2+
Version 2.0, January 2004
3+
http://www.apache.org/licenses/
4+
5+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
6+
7+
1. Definitions.
8+
9+
"License" shall mean the terms and conditions for use, reproduction,
10+
and distribution as defined by Sections 1 through 9 of this document.
11+
12+
"Licensor" shall mean the copyright owner or entity authorized by
13+
the copyright owner that is granting the License.
14+
15+
"Legal Entity" shall mean the union of the acting entity and all
16+
other entities that control, are controlled by, or are under common
17+
control with that entity. For the purposes of this definition,
18+
"control" means (i) the power, direct or indirect, to cause the
19+
direction or management of such entity, whether by contract or
20+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
21+
outstanding shares, or (iii) beneficial ownership of such entity.
22+
23+
"You" (or "Your") shall mean an individual or Legal Entity
24+
exercising permissions granted by this License.
25+
26+
"Source" form shall mean the preferred form for making modifications,
27+
including but not limited to software source code, documentation
28+
source, and configuration files.
29+
30+
"Object" form shall mean any form resulting from mechanical
31+
transformation or translation of a Source form, including but
32+
not limited to compiled object code, generated documentation,
33+
and conversions to other media types.
34+
35+
"Work" shall mean the work of authorship, whether in Source or
36+
Object form, made available under the License, as indicated by a
37+
copyright notice that is included in or attached to the work
38+
(an example is provided in the Appendix below).
39+
40+
"Derivative Works" shall mean any work, whether in Source or Object
41+
form, that is based on (or derived from) the Work and for which the
42+
editorial revisions, annotations, elaborations, or other modifications
43+
represent, as a whole, an original work of authorship. For the purposes
44+
of this License, Derivative Works shall not include works that remain
45+
separable from, or merely link (or bind by name) to the interfaces of,
46+
the Work and Derivative Works thereof.
47+
48+
"Contribution" shall mean any work of authorship, including
49+
the original version of the Work and any modifications or additions
50+
to that Work or Derivative Works thereof, that is intentionally
51+
submitted to Licensor for inclusion in the Work by the copyright owner
52+
or by an individual or Legal Entity authorized to submit on behalf of
53+
the copyright owner. For the purposes of this definition, "submitted"
54+
means any form of electronic, verbal, or written communication sent
55+
to the Licensor or its representatives, including but not limited to
56+
communication on electronic mailing lists, source code control systems,
57+
and issue tracking systems that are managed by, or on behalf of, the
58+
Licensor for the purpose of discussing and improving the Work, but
59+
excluding communication that is conspicuously marked or otherwise
60+
designated in writing by the copyright owner as "Not a Contribution."
61+
62+
"Contributor" shall mean Licensor and any individual or Legal Entity
63+
on behalf of whom a Contribution has been received by Licensor and
64+
subsequently incorporated within the Work.
65+
66+
2. Grant of Copyright License. Subject to the terms and conditions of
67+
this License, each Contributor hereby grants to You a perpetual,
68+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
69+
copyright license to reproduce, prepare Derivative Works of,
70+
publicly display, publicly perform, sublicense, and distribute the
71+
Work and such Derivative Works in Source or Object form.
72+
73+
3. Grant of Patent License. Subject to the terms and conditions of
74+
this License, each Contributor hereby grants to You a perpetual,
75+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
76+
(except as stated in this section) patent license to make, have made,
77+
use, offer to sell, sell, import, and otherwise transfer the Work,
78+
where such license applies only to those patent claims licensable
79+
by such Contributor that are necessarily infringed by their
80+
Contribution(s) alone or by combination of their Contribution(s)
81+
with the Work to which such Contribution(s) was submitted. If You
82+
institute patent litigation against any entity (including a
83+
cross-claim or counterclaim in a lawsuit) alleging that the Work
84+
or a Contribution incorporated within the Work constitutes direct
85+
or contributory patent infringement, then any patent licenses
86+
granted to You under this License for that Work shall terminate
87+
as of the date such litigation is filed.
88+
89+
4. Redistribution. You may reproduce and distribute copies of the
90+
Work or Derivative Works thereof in any medium, with or without
91+
modifications, and in Source or Object form, provided that You
92+
meet the following conditions:
93+
94+
(a) You must give any other recipients of the Work or
95+
Derivative Works a copy of this License; and
96+
97+
(b) You must cause any modified files to carry prominent notices
98+
stating that You changed the files; and
99+
100+
(c) You must retain, in the Source form of any Derivative Works
101+
that You distribute, all copyright, patent, trademark, and
102+
attribution notices from the Source form of the Work,
103+
excluding those notices that do not pertain to any part of
104+
the Derivative Works; and
105+
106+
(d) If the Work includes a "NOTICE" text file as part of its
107+
distribution, then any Derivative Works that You distribute must
108+
include a readable copy of the attribution notices contained
109+
within such NOTICE file, excluding those notices that do not
110+
pertain to any part of the Derivative Works, in at least one
111+
of the following places: within a NOTICE text file distributed
112+
as part of the Derivative Works; within the Source form or
113+
documentation, if provided along with the Derivative Works; or,
114+
within a display generated by the Derivative Works, if and
115+
wherever such third-party notices normally appear. The contents
116+
of the NOTICE file are for informational purposes only and
117+
do not modify the License. You may add Your own attribution
118+
notices within Derivative Works that You distribute, alongside
119+
or as an addendum to the NOTICE text from the Work, provided
120+
that such additional attribution notices cannot be construed
121+
as modifying the License.
122+
123+
You may add Your own copyright statement to Your modifications and
124+
may provide additional or different license terms and conditions
125+
for use, reproduction, or distribution of Your modifications, or
126+
for any such Derivative Works as a whole, provided Your use,
127+
reproduction, and distribution of the Work otherwise complies with
128+
the conditions stated in this License.
129+
130+
5. Submission of Contributions. Unless You explicitly state otherwise,
131+
any Contribution intentionally submitted for inclusion in the Work
132+
by You to the Licensor shall be under the terms and conditions of
133+
this License, without any additional terms or conditions.
134+
Notwithstanding the above, nothing herein shall supersede or modify
135+
the terms of any separate license agreement you may have executed
136+
with Licensor regarding such Contributions.
137+
138+
6. Trademarks. This License does not grant permission to use the trade
139+
names, trademarks, service marks, or product names of the Licensor,
140+
except as required for reasonable and customary use in describing the
141+
origin of the Work and reproducing the content of the NOTICE file.
142+
143+
7. Disclaimer of Warranty. Unless required by applicable law or
144+
agreed to in writing, Licensor provides the Work (and each
145+
Contributor provides its Contributions) on an "AS IS" BASIS,
146+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
147+
implied, including, without limitation, any warranties or conditions
148+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
149+
PARTICULAR PURPOSE. You are solely responsible for determining the
150+
appropriateness of using or redistributing the Work and assume any
151+
risks associated with Your exercise of permissions under this License.
152+
153+
8. Limitation of Liability. In no event and under no legal theory,
154+
whether in tort (including negligence), contract, or otherwise,
155+
unless required by applicable law (such as deliberate and grossly
156+
negligent acts) or agreed to in writing, shall any Contributor be
157+
liable to You for damages, including any direct, indirect, special,
158+
incidental, or consequential damages of any character arising as a
159+
result of this License or out of the use or inability to use the
160+
Work (including but not limited to damages for loss of goodwill,
161+
work stoppage, computer failure or malfunction, or any and all
162+
other commercial damages or losses), even if such Contributor
163+
has been advised of the possibility of such damages.
164+
165+
9. Accepting Warranty or Additional Liability. While redistributing
166+
the Work or Derivative Works thereof, You may choose to offer,
167+
and charge a fee for, acceptance of support, warranty, indemnity,
168+
or other liability obligations and/or rights consistent with this
169+
License. However, in accepting such obligations, You may act only
170+
on Your own behalf and on Your sole responsibility, not on behalf
171+
of any other Contributor, and only if You agree to indemnify,
172+
defend, and hold each Contributor harmless for any liability
173+
incurred by, or claims asserted against, such Contributor by reason
174+
of your accepting any such warranty or additional liability.
175+
176+
END OF TERMS AND CONDITIONS
177+
178+
APPENDIX: How to apply the Apache License to your work.
179+
180+
To apply the Apache License to your work, attach the following
181+
boilerplate notice, with the fields enclosed by brackets "[]"
182+
replaced with your own identifying information. (Don't include
183+
the brackets!) The text should be enclosed in the appropriate
184+
comment syntax for the file format. We also recommend that a
185+
file or class name and description of purpose be included on the
186+
same "printed page" as the copyright notice for easier
187+
identification within third-party archives.
188+
189+
Copyright [yyyy] [name of copyright owner]
190+
191+
Licensed under the Apache License, Version 2.0 (the "License");
192+
you may not use this file except in compliance with the License.
193+
You may obtain a copy of the License at
194+
195+
http://www.apache.org/licenses/LICENSE-2.0
196+
197+
Unless required by applicable law or agreed to in writing, software
198+
distributed under the License is distributed on an "AS IS" BASIS,
199+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
200+
See the License for the specific language governing permissions and
201+
limitations under the License.

‎python/aibrix_kvcache/README.md‎

Lines changed: 45 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,45 @@
1+
# AIBrix KV Cache Offloading Framework for Cross-Engine KV Reuse
2+
AIBrix KV cache offloading framework provides several common functionalities for cross-engine KV reuse use cases:
3+
4+
**Tensor Parallelism Aware Management**: When inference engine (e.g., vLLM) uses tensor parallelism, each participating engine instance fetches KV tensors independently from the cache backend. In case of cache misses, before proceeding with prefill computation, participants must align the potentially different number of KV tensors fetched from the external KV cache service to ensure a consistent view .
5+
6+
**Embedded Cache w/ CPU Memory**: To meet performance requirements, it's common to have a small CPU memory-based cache embedded in the engine to avoid frequently accessing remote cache backends.
7+
8+
**Selective KV Cache Offloading**: Enables fine-grained control over offloading strategies and thus is crucial in optimizing performance across diverse deployment environments:
9+
1. Many cloud providers and companies deploy lower-end GPU instances without high-speed interconnects like RDMA, suited for tasks related to 7B/8B models running on 24/32GiB GPU cards. In these setups, GPUs within the same instance (typically 8-16 GPUs) share a single VPC NIC, leading to significant network bandwidth contention. Selective KV cache offloading (e.g., only offloading KV tensors identified by the employed eviction policy as hot rather than offloading all KV tensors) helps mitigate this issue by reducing unnecessary data transfers and conserving limited network bandwidth.
10+
2. Even in high-performance environments with RDMA-equipped GPUs, selective KV cache offloading can enhance efficiency by limiting the PCIe bandwidth consumed by remote data movement. While RDMA enables low-latency, high-bandwidth communication, remote data access still incurs higher latency than local memory access. By leveraging selective KV offloading, the framework reduces the frequency of remote data transfers, preserving PCIe bandwidth and ensuring that local memory access remains the preferred data pathway.
11+
To achieve selective KV cache offloading, we introduce an eviction policy layer that can be extended and customized with advanced offloading strategies to determine which KV tensors should be offloaded. Within this layer, multiple callbacks are available to support different offloading modes, including offloading all KV tensors, only hot KV tensors, or only cold KV tensors, with the definition of "hot" and "cold" being determined by the specific eviction policy in use. In this initial PR, the framework will provide built-in support for LRU, FIFO, and S3FIFO eviction policies.
12+
13+
## Quick Start
14+
### Installation
15+
AIBrix KV cache offloading framework can be installed by `pip`.
16+
17+
```sh
18+
pip install aibrix-kvcache
19+
```
20+
21+
## Contributing
22+
We welcome contributions from the community! Check out our [contributing guidelines](https://github.com/vllm-project/aibrix/blob/main/CONTRIBUTING.md) to see how you can make a difference.
23+
24+
### Build from source
25+
26+
```bash
27+
# This may take several minutes
28+
pip install -e .
29+
```
30+
31+
### Lint, Format and Type Check
32+
33+
Before contribute your code, please run the following commands to ensure that your code passes the tests and linting checks.
34+
35+
```bash
36+
# install dependencies
37+
poetry install --no-root --with dev
38+
39+
# linting, formatting and type checking
40+
bash ./scripts/format.sh
41+
```
42+
43+
## License
44+
45+
AI Runtime is licensed under the [APACHE License](https://github.com/vllm-project/aibrix/LICENSE.md).

‎python/aibrix_kvcache/aibrix_kvcache/cache_handle.py‎

Lines changed: 69 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,69 @@
1+
# Copyright 2024 The Aibrix Team.
2+
#
3+
# Licensed under the Apache License, Version 2.0 (the "License");
4+
# you may not use this file except in compliance with the License.
5+
# You may obtain a copy of the License at
6+
#
7+
# http://www.apache.org/licenses/LICENSE-2.0
8+
#
9+
# Unless required by applicable law or agreed to in writing, software
10+
# distributed under the License is distributed on an "AS IS" BASIS,
11+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12+
# See the License for the specific language governing permissions and
13+
# limitations under the License.
14+
15+
from abc import ABC, abstractmethod
16+
from typing import Sequence, Tuple
17+
18+
import torch
19+
20+
from .memory import MemoryRegion
21+
22+
23+
class KVCacheHandle(ABC):
24+
"""Cache handle to support zero-copy APIs."""
25+
26+
@property
27+
@abstractmethod
28+
def memory_regions(self) -> Sequence[MemoryRegion]:
29+
raise NotImplementedError
30+
31+
@abstractmethod
32+
def to_tensors(self) -> Sequence[torch.Tensor]:
33+
raise NotImplementedError
34+
35+
@abstractmethod
36+
def release(self) -> None:
37+
raise NotImplementedError
38+
39+
@abstractmethod
40+
def __len__(self) -> int:
41+
raise NotImplementedError
42+
43+
44+
class MemoryRegionKVCacheHandle(KVCacheHandle):
45+
def __init__(
46+
self,
47+
block_dtype: torch.dtype,
48+
block_shape: Tuple[int, ...],
49+
mrs: Sequence[MemoryRegion],
50+
) -> None:
51+
self._block_dtype = block_dtype
52+
self._block_shape = block_shape
53+
self._mrs = mrs
54+
55+
@property
56+
def memory_regions(self) -> Sequence[MemoryRegion]:
57+
return self._mrs
58+
59+
def to_tensors(self) -> Sequence[torch.Tensor]:
60+
return MemoryRegion.to_tensors(
61+
self._mrs, self._block_dtype, self._block_shape
62+
)
63+
64+
def release(self) -> None:
65+
for mr in self._mrs:
66+
mr.ref_down()
67+
68+
def __len__(self) -> int:
69+
return len(self._mrs)

0 commit comments

Comments
 (0)