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

[CPU][Bugfix] Using custom allreduce for CPU backend (#15934)

Browse Source 
Signed-off-by: jiang1.li <jiang1.li@intel.com>
This commit is contained in:
Li, Jiang
2025-04-02 22:46:47 +08:00
committed by GitHub
parent cefb9e5a28
commit 550b2801ad
8 changed files with 1013 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
docs/source/getting_started/installation/cpu.md
View File

@@ -272,12 +272,14 @@ $ python examples/offline_inference/basic/basic.py
- Decouple the HTTP serving components from the inference components. In a GPU backend configuration, the HTTP serving and tokenization tasks operate on the CPU, while inference runs on the GPU, which typically does not pose a problem. However, in a CPU-based setup, the HTTP serving and tokenization can cause significant context switching and reduced cache efficiency. Therefore, it is strongly recommended to segregate these two components for improved performance.
- On CPU based setup with NUMA enabled, the memory access performance may be largely impacted by the [topology](https://github.com/intel/intel-extension-for-pytorch/blob/main/docs/tutorials/performance_tuning/tuning_guide.inc.md#non-uniform-memory-access-numa). For NUMA architecture, two optimizations are to recommended: Tensor Parallel or Data Parallel.
- On CPU based setup with NUMA enabled, the memory access performance may be largely impacted by the [topology](https://github.com/intel/intel-extension-for-pytorch/blob/main/docs/tutorials/performance_tuning/tuning_guide.inc.md#non-uniform-memory-access-numa). For NUMA architecture, Tensor Parallel is a option for better performance.
- Using Tensor Parallel for a latency constraints deployment: following GPU backend design, a Megatron-LM's parallel algorithm will be used to shard the model, based on the number of NUMA nodes (e.g. TP = 2 for a two NUMA node system). With [TP feature on CPU](gh-pr:6125) merged, Tensor Parallel is supported for serving and offline inferencing. In general each NUMA node is treated as one GPU card. Below is the example script to enable Tensor Parallel = 2 for serving:
- Tensor Parallel is supported for serving and offline inferencing. In general each NUMA node is treated as one GPU card. Below is the example script to enable Tensor Parallel = 2 for serving:
```console
VLLM_CPU_KVCACHE_SPACE=40 VLLM_CPU_OMP_THREADS_BIND="0-31|32-63" vllm serve meta-llama/Llama-2-7b-chat-hf -tp=2 --distributed-executor-backend mp
```
- Using Data Parallel for maximum throughput: to launch an LLM serving endpoint on each NUMA node along with one additional load balancer to dispatch the requests to those endpoints. Common solutions like [Nginx](#nginxloadbalancer) or HAProxy are recommended. Anyscale Ray project provides the feature on LLM [serving](https://docs.ray.io/en/latest/serve/index.html). Here is the example to setup a scalable LLM serving with [Ray Serve](https://github.com/intel/llm-on-ray/blob/main/docs/setup.inc.md).
- For each thread id list in `VLLM_CPU_OMP_THREADS_BIND`, users should guarantee threads in the list belong to a same NUMA node.
- Meanwhile, users should also take care of memory capacity of each NUMA node. The memory usage of each TP rank is the sum of `weight shard size` and `VLLM_CPU_KVCACHE_SPACE`, if it exceeds the capacity of a single NUMA node, TP worker will be killed due to out-of-memory.