xc-llm-ascend/docs/source/user_guide/feature_guide/context_parallel.md

Context Parallel Guide

Overview

This guide shows how to use Context Parallel, a long sequence inference optimization technique. Context Parallel includes PCP (Prefill Context Parallel) and DCP (Decode Context Parallel), which reduces NPU memory usage and improves inference speed in long sequence LLM inference.

Benefits of Context Parallel

Context parallel mainly solves the problem of serving long context requests. As prefill and decode present quite different characteristics and have quite different SLO (service level objectives), we need to implement context parallel separately for them. The major considerations are:

• For long context prefill, we can use context parallel to reduce TTFT (time to first token) by amortizing the computation time of the prefill across query tokens.
• For long context decode, we can use context parallel to reduce KV cache duplication and offer more space for KV cache to increase the batch size (and hence the throughput).

To learn more about the theory and implementation details of context parallel, please refer to the context parallel developer guide.

Supported Scenarios

Currently context parallel can be used together with most other features, supported features are as follows:

	Eager	Graph	Prefix Cache	Chunked Prefill	SpecDecode (MTP)	PD disaggregation	MLAPO
PCP	✅	✅	✅	✅	✅	✅	✅
DCP	✅	✅	✅	✅	✅	✅	✅

How to use Context Parallel

You can enable PCP and DCP by prefill_context_parallel_size and decode_context_parallel_size, refer to the following example:

• Offline example:

  from vllm import LLM, SamplingParams
  
  prompts = [
      "The future of AI is",
  ]
  sampling_params = SamplingParams(temperature=0.8, top_p=0.95)
  
  llm = LLM(
      model="deepseek-ai/DeepSeek-V2-Lite",
      tensor_parallel_size=2,
      decode_context_parallel_size=2,
      prefill_context_parallel_size=2,
  )
  outputs = llm.generate(prompts, sampling_params)
  

• Online example:

  vllm serve deepseek-ai/DeepSeek-V2-Lite \
      --tensor-parallel-size 2 \
      --decode-context-parallel-size 2 \
      --prefill-context-parallel-size 2 \
  

The total world size is tensor_parallel_size * prefill_context_parallel_size, so the examples above need 4 NPUs for each.

Constraints

• While using DCP, the following constraints must be met:

  • For MLA-based model, such as DeepSeek-R1:
    • tensor_parallel_size >= decode_context_parallel_size
    • tensor_parallel_size % decode_context_parallel_size == 0
  • For GQA-based model, such as Qwen3-235B:
    • (tensor_parallel_size // num_key_value_heads) >= decode_context_parallel_size
    • (tensor_parallel_size // num_key_value_heads) % decode_context_parallel_size == 0

• While using Context Parallel in KV cache transfer-needed scenario (e.g. KV pooling, PD disaggregation), to simplify KV cache transmission, cp_kv_cache_interleave_size must be set to the same value of KV cache block_size(default: 128), which specifies CP to split KV cache in a block-interleave style. For example:

  vllm serve deepseek-ai/DeepSeek-V2-Lite \
      --tensor-parallel-size 2 \
      --decode-context-parallel-size 2 \
      --prefill-context-parallel-size 2 \
      --cp-kv-cache-interleave-size 128 \
      --kv-transfer-config {...} \
  

Experimental Results

To evaluate the effectiveness of Context Parallel in long sequence LLM inference scenarios, we use DeepSeek-R1-W8A8 and Qwen3-235B, deploy PD disaggregate instances in the environment of 64 cards Ascend 910C*64G (A3), the configuration and performance data are as follows.

• DeepSeek-R1-W8A8:

  Configuration	Input length 32k	Input length 64k	Input length 128k
  P node: (DP2 TP8 EP16) *2 D node: (DP32 EP32)*1	TTFT: 9.3s TPOT: 72ms	TTFT: 22.8s TPOT: 74ms	TTFT: 73.2s TPOT: 82ms
  P node: (PCP2 TP8 DCP8 EP16) *2 D node: (DP32 EP32)*1	TTFT: 7.9s TPOT: 74ms	TTFT: 15.9s TPOT: 78ms	TTFT: 46.0s TPOT: 83ms

• Qwen3-235B:

  Configuration	Input length 32k	Input length 64k	Input length 120k
  P node: (DP2 TP8 EP16) *2 D node: (DP32 EP32)*1	TTFT: 5.1s TPOT: 65ms	TTFT: 13.1s TPOT: 85ms	TTFT: 33.9s TPOT: 120ms
  P node: (PCP2 TP8 DCP2 EP16) *2 D node: (DP32 EP32)*1	TTFT: 3.0s TPOT: 66ms	TTFT: 8.9s TPOT: 86ms	TTFT: 22.7s TPOT: 121ms