Kimi-K2.5 通過 KTransformers+SGLang 在混合 GPU/CPU 記憶體卸載配置上:4x RTX Pro 6000 Blackwells + 640GB RAM 原始基準是在 8x GPU 上使用合成編碼代理樣式工作負載,目標為 2k-45k 輸入標記,80-3k 最大輸出標記,並且最多可同時處理 10 個請求。重新在新的混合設置上運行 我能得到的最佳結果: - 23.03 輸出標記/秒 @ 10 個同時請求 - 平均 TTFT:~60 秒 - 中位數 TTFT:~64 秒 基準結果: - 74.39 輸出標記/秒 @ 10 個同時請求 - 平均 TTFT:~9 秒 - 中位數 TTFT:~3.7 秒
Yannick Nick
Yannick Nick2026年2月26日
Initial tests for Kimi-K2.5 via KTransformers+SGLang, on a hybrid 4x RTX Pro 6000 Blackwell + 640GB/1.5TB CPU memory offload. Compute provided by Lium pods: - 19.97 output tok/s @ 10 concurrent requests - Mean TTFT: ~120s - Median TTFT: ~102s Need to play with the KT flags to further optimize this setup, which is heavily dependent on the overall system's CPU core count & available RAM. GPU <-> PCIe <-> RAM interconnectivity is the most obvious bottleneck Experts per MoE Layer on GPU: --kt-num-gpu-experts=128 CPU cores dedicated to MoE inference: --kt-cpuinfer=104 CPU experts work overlapping GPU work: --kt-max-deferred-experts-per-token=2 Max tokens per prefill chunk: --chunked-prefill-size=32658 CUDA graph capture disabled: --disable-cuda-graph
完整命令: export CUDA_VISIBLE_DEVICES=0,1,2,3 export OMP_NUM_THREADS=1 export MKL_NUM_THREADS=1 export OPENBLAS_NUM_THREADS=1 export NUMEXPR_NUM_THREADS=1 export VECLIB_MAXIMUM_THREADS=1 export PYTHONUNBUFFERED=1 exec python -m sglang.launch_server \ --model-path /workspace/models/huggingface/models--moonshotai--Kimi-K2.5/snapshots/54383e83fa343a1331754112fb9e3410c55efa2f \ --kt-weight-path /workspace/models/huggingface/models--moonshotai--Kimi-K2.5/snapshots/54383e83fa343a1331754112fb9e3410c55efa2f \ --kt-threadpool-count 1 \ --kt-method RAWINT4 \ --trust-remote-code \ --served-model-name kimi_k2 \ --tool-call-parser kimi_k2 \ --reasoning-parser kimi_k2 \ --disable-radix-cache \ --disable-chunked-prefix-cache \ --tensor-parallel-size 4 \ --enable-p2p-check \ --disable-shared-experts-fusion \ --disable-cuda-graph \ --host 0.0.0.0 \ --port 8000 \ --kt-cpuinfer 32 \ --kt-num-gpu-experts 128 \ --kt-max-deferred-experts-per-token 2 \ --kt-gpu-prefill-token-threshold 1024 \ --kt-expert-placement-strategy uniform \ --mem-fraction-static 0.92 \ --enable-mixed-chunk \ --chunked-prefill-size 32658 \ --max-total-tokens 200000 \ --attention-backend flashinfer
111