549be94397
### What this PR does / why we need it?
Fix the bug in the PCP overlay feature
1、Fix the bug related to PCP and EPLB overlap by including PCP size in
the word_size calculation.
2、In the PCP pooling scenario, a prompt has been added for setting the
cp_kv_cache_interleave_size.
- vLLM version: v0.13.0
- vLLM main:
7157596103
Signed-off-by: weiguihua2 <weiguihua2@huawei.com>
99 lines
3.0 KiB
YAML
99 lines
3.0 KiB
YAML
test_name: "test Qwen3-235B-A22B-W8A8-longseq disaggregated_prefill"
|
|
model: "vllm-ascend/Qwen3-235B-A22B-W8A8"
|
|
num_nodes: 2
|
|
npu_per_node: 16
|
|
env_common:
|
|
VLLM_USE_MODELSCOPE: true
|
|
OMP_PROC_BIND: false
|
|
OMP_NUM_THREADS: 100
|
|
HCCL_BUFFSIZE: 1024
|
|
SERVER_PORT: 8080
|
|
NUMEXPR_MAX_THREADS: 128
|
|
DYNAMIC_EPLB: true
|
|
disaggregated_prefill:
|
|
enabled: true
|
|
prefiller_host_index: [0]
|
|
decoder_host_index: [1]
|
|
|
|
deployment:
|
|
-
|
|
server_cmd: >
|
|
vllm serve "vllm-ascend/Qwen3-235B-A22B-W8A8"
|
|
--host 0.0.0.0
|
|
--port $SERVER_PORT
|
|
--data-parallel-size 1
|
|
--decode-context-parallel-size 2
|
|
--prefill-context-parallel-size 2
|
|
--tensor-parallel-size 8
|
|
--cp-kv-cache-interleave-size 128
|
|
--seed 1024
|
|
--enforce-eager
|
|
--enable-expert-parallel
|
|
--max-num-seqs 16
|
|
--max-model-len 8192
|
|
--max-num-batched-tokens 8192
|
|
--quantization ascend
|
|
--trust-remote-code
|
|
--no-enable-prefix-caching
|
|
--gpu-memory-utilization 0.9
|
|
--kv-transfer-config
|
|
'{"kv_connector": "MooncakeConnectorV1",
|
|
"kv_role": "kv_producer",
|
|
"kv_port": "30000",
|
|
"engine_id": "0",
|
|
"kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector",
|
|
"kv_connector_extra_config": {
|
|
"prefill": {
|
|
"dp_size": 1,
|
|
"tp_size": 8
|
|
},
|
|
"decode": {
|
|
"dp_size": 2,
|
|
"tp_size": 8
|
|
}
|
|
}
|
|
}'
|
|
--additional-config
|
|
'{"dynamic_eplb":true}'
|
|
|
|
-
|
|
server_cmd: >
|
|
vllm serve "vllm-ascend/Qwen3-235B-A22B-W8A8"
|
|
--host 0.0.0.0
|
|
--port $SERVER_PORT
|
|
--data-parallel-size 2
|
|
--decode-context-parallel-size 2
|
|
--prefill-context-parallel-size 1
|
|
--tensor-parallel-size 8
|
|
--cp-kv-cache-interleave-size 128
|
|
--seed 1024
|
|
--quantization ascend
|
|
--max-num-seqs 16
|
|
--max-model-len 8192
|
|
--max-num-batched-tokens 8192
|
|
--enable-expert-parallel
|
|
--trust-remote-code
|
|
--no-enable-prefix-caching
|
|
--gpu-memory-utilization 0.9
|
|
--compilation_config '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
|
|
--kv-transfer-config
|
|
'{"kv_connector": "MooncakeConnectorV1",
|
|
"kv_role": "kv_consumer",
|
|
"kv_port": "30100",
|
|
"engine_id": "1",
|
|
"kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector",
|
|
"kv_connector_extra_config": {
|
|
"prefill": {
|
|
"dp_size": 1,
|
|
"tp_size": 8
|
|
},
|
|
"decode": {
|
|
"dp_size": 2,
|
|
"tp_size": 8
|
|
}
|
|
}
|
|
}'
|
|
--additional-config
|
|
'{"dynamic_eplb":true}'
|
|
benchmarks:
|