biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Hardware] [Intel] Enable Multiprocessing and tensor parallel in CPU backend and update documentation (#6125)

Browse Source
This commit is contained in:
Li, Jiang
2024-07-27 04:50:10 +08:00
committed by GitHub
parent aa4867791e
commit 3bbb4936dc
14 changed files with 403 additions and 89 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
docs/source/getting_started/cpu-installation.rst
View File

@@ -10,6 +10,7 @@ Table of contents:
#. :ref:`Requirements <cpu_backend_requirements>`
#. :ref:`Quick start using Dockerfile <cpu_backend_quick_start_dockerfile>`
#. :ref:`Build from source <build_cpu_backend_from_source>`
#. :ref:`Related runtime environment variables <env_intro>`
#. :ref:`Intel Extension for PyTorch <ipex_guidance>`
#. :ref:`Performance tips <cpu_backend_performance_tips>`
@@ -47,7 +48,7 @@ Build from source
.. code-block:: console
$ sudo apt-get update -y
$ sudo apt-get install -y gcc-12 g++-12
$ sudo apt-get install -y gcc-12 g++-12 libnuma-dev
$ sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-12 10 --slave /usr/bin/g++ g++ /usr/bin/g++-12
- Second, install Python packages for vLLM CPU backend building:
@@ -71,6 +72,15 @@ Build from source
- If you want to force enable AVX512_BF16 for the cross-compilation, please set environment variable VLLM_CPU_AVX512BF16=1 before the building.
.. _env_intro:
Related runtime environment variables
-------------------------------------
- ``VLLM_CPU_KVCACHE_SPACE``: specify the KV Cache size (e.g, ``VLLM_CPU_KVCACHE_SPACE=40`` means 40 GB space for KV cache), larger setting will allow vLLM running more requests in parallel. This parameter should be set based on the hardware configuration and memory management pattern of users.
- ``VLLM_CPU_OMP_THREADS_BIND``: specify the CPU cores dedicated to the OpenMP threads. For example, ``VLLM_CPU_OMP_THREADS_BIND=0-31`` means there will be 32 OpenMP threads bound on 0-31 CPU cores. ``VLLM_CPU_OMP_THREADS_BIND=0-31|32-63`` means there will be 2 tensor parallel processes, 32 OpenMP threads of rank0 are bound on 0-31 CPU cores, and the OpenMP threads of rank1 are bound on 32-63 CPU cores.
.. _ipex_guidance:
Intel Extension for PyTorch
@@ -78,15 +88,11 @@ Intel Extension for PyTorch
- `Intel Extension for PyTorch (IPEX) <https://github.com/intel/intel-extension-for-pytorch>`_ extends PyTorch with up-to-date features optimizations for an extra performance boost on Intel hardware.
- IPEX after the ``2.3.0`` can be enabled in the CPU backend by default if it is installed.
.. _cpu_backend_performance_tips:
Performance tips
-----------------
- vLLM CPU backend uses environment variable ``VLLM_CPU_KVCACHE_SPACE`` to specify the KV Cache size (e.g, ``VLLM_CPU_KVCACHE_SPACE=40`` means 40 GB space for KV cache), larger setting will allow vLLM running more requests in parallel. This parameter should be set based on the hardware configuration and memory management pattern of users.
- We highly recommend to use TCMalloc for high performance memory allocation and better cache locality. For example, on Ubuntu 22.4, you can run:
.. code-block:: console
@@ -96,11 +102,44 @@ Performance tips
$ export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libtcmalloc_minimal.so.4:$LD_PRELOAD # prepend the library to LD_PRELOAD
$ python examples/offline_inference.py # run vLLM
- vLLM CPU backend uses OpenMP for thread-parallel computation. If you want the best performance on CPU, it will be very critical to isolate CPU cores for OpenMP threads with other thread pools (like web-service event-loop), to avoid CPU oversubscription.
- When using the online serving, it is recommended to reserve 1-2 CPU cores for the serving framework to avoid CPU oversubscription. For example, on a platform with 32 physical CPU cores, reserving CPU 30 and 31 for the framework and using CPU 0-29 for OpenMP:
- If using vLLM CPU backend on a bare-metal machine, it is recommended to disable the hyper-threading.
.. code-block:: console
- If using vLLM CPU backend on a multi-socket machine with NUMA, be aware to set CPU cores and memory nodes, to avoid the remote memory node access. ``numactl`` is an useful tool for CPU core and memory binding on NUMA platform. Besides, ``--cpuset-cpus`` and ``--cpuset-mems`` arguments of ``docker run`` are also useful.
$ export VLLM_CPU_KVCACHE_SPACE=40
$ export VLLM_CPU_OMP_THREADS_BIND=0-29
$ vllm serve facebook/opt-125m
- If using vLLM CPU backend on a machine with hyper-threading, it is recommended to bind only one OpenMP thread on each physical CPU core using ``VLLM_CPU_OMP_THREADS_BIND``. On a hyper-threading enabled platform with 16 logical CPU cores / 8 physical CPU cores:
.. code-block:: console
$ lscpu -e # check the mapping between logical CPU cores and physical CPU cores
# The "CPU" column means the logical CPU core IDs, and the "CORE" column means the physical core IDs. On this platform, two logical cores are sharing one physical core.
CPU NODE SOCKET CORE L1d:L1i:L2:L3 ONLINE MAXMHZ MINMHZ MHZ
0 0 0 0 0:0:0:0 yes 2401.0000 800.0000 800.000
1 0 0 1 1:1:1:0 yes 2401.0000 800.0000 800.000
2 0 0 2 2:2:2:0 yes 2401.0000 800.0000 800.000
3 0 0 3 3:3:3:0 yes 2401.0000 800.0000 800.000
4 0 0 4 4:4:4:0 yes 2401.0000 800.0000 800.000
5 0 0 5 5:5:5:0 yes 2401.0000 800.0000 800.000
6 0 0 6 6:6:6:0 yes 2401.0000 800.0000 800.000
7 0 0 7 7:7:7:0 yes 2401.0000 800.0000 800.000
8 0 0 0 0:0:0:0 yes 2401.0000 800.0000 800.000
9 0 0 1 1:1:1:0 yes 2401.0000 800.0000 800.000
10 0 0 2 2:2:2:0 yes 2401.0000 800.0000 800.000
11 0 0 3 3:3:3:0 yes 2401.0000 800.0000 800.000
12 0 0 4 4:4:4:0 yes 2401.0000 800.0000 800.000
13 0 0 5 5:5:5:0 yes 2401.0000 800.0000 800.000
14 0 0 6 6:6:6:0 yes 2401.0000 800.0000 800.000
15 0 0 7 7:7:7:0 yes 2401.0000 800.0000 800.000
# On this platform, it is recommend to only bind openMP threads on logical CPU cores 0-7 or 8-15
$ export VLLM_CPU_OMP_THREADS_BIND=0-7
$ python examples/offline_inference.py
- If using vLLM CPU backend on a multi-socket machine with NUMA, be aware to set CPU cores using ``VLLM_CPU_OMP_THREADS_BIND`` to avoid cross NUMA node memory access.