ai.canshu.com
中文
VRAM Calculator Advisor & TCO API Pricing Compare Models Leaderboard GPUs

Search models, GPUs, and tools

↑↓ Navigate · Enter Open · ★ Watchlist prioritized ai.canshu.com
Q4_K_M: Partial RAM Offload (Slow) 77.26 GB / 61.4 GB Usable

Qwen3.5-122B-A10B (MoE) on 海光深算 DCU K100-AI 64GB (HBM2e): VRAM Requirement & Speed Estimate

A single 海光 DCU K100-AI (61.4 GB usable VRAM) cannot fully fit Qwen3.5-122B-A10B (MoE) at Q4_K_M (77.26 GB required). Consider multi-GPU tensor parallelism, a smaller parameter variant, or cloud GPU / API deployment.

Concurrency (Batch):
MoE 122B / A10B·48L (d=3072)·32Q/2KV·Ctx 256K·Text+Image+Video
Runnable Models →
GPUs
64 GB VRAM·896 GB/s·Safe Usable 61.4 GB
Partial RAM Offload (Slow)77.26 / 61.4 GB~37.3 tok/s
3. Weight Quantization
4. Per-Request Context16K (16,384 tok)
Verified: 2026-11-20·Roofline bandwidth/compute estimate (Batch=1)
Partial RAM Offload (Slow)
77.26/ 61.4 GB usable

Physical 64 GB · Load 125.8%

Theoretical Decode Speed
~37.3tokens/s

MoE: reads only 6.06 GB active weights/step

Max Context at Batch=1
0K/ 256K

Enable Q8/Q4 KV quantization to scale batch concurrency

Weights: 73.96 GBKV Cache (1×16K): 0.52 GBOverhead: 2.78 GB
🏃 Smooth & Snappy~13.4s for a 500-token response
Reference Launch CLI:
OLLAMA_CONTEXT_LENGTH=16384 ollama run qwen3.5:122b-a10b

Insufficient Local VRAM? Ways to Run Qwen3.5-122B-A10B (MoE)

Rent an hourly high-VRAM cloud GPU pod, or check the 12-month Buy vs Cloud vs API break-even curve.

📊 All Quantization Levels for Qwen3.5-122B-A10B (MoE) on 海光 DCU K100-AI

Quantization Total VRAM (16K) Verdict Est. Speed Max Context
FP16 / BF16 250.7 GB Out of VRAM (OOM) — —
FP8 / Q8_0 134.04 GB Out of VRAM (OOM) — —
GGUF Q6_K 103.86 GB Out of VRAM (OOM) — —
GGUF Q5_K_M 90.33 GB Partial RAM Offload (Slow) ~23.6 tok/s —
Q4_K_M / AWQ 77.26 GB Partial RAM Offload (Slow) ~37.3 tok/s —
GGUF IQ4_XS 67.93 GB Partial RAM Offload (Slow) ~63.9 tok/s —
GGUF Q3_K_M 62.64 GB Partial RAM Offload (Slow) ~107.8 tok/s —
GGUF Q2_K 50.82 GB Smooth Fit ~153.7 tok/s 256K

💡 Recommended Alternative GPUs & Macs for Qwen3.5-122B-A10B (MoE)

Frequently Asked Questions

Can 海光深算 DCU K100-AI 64GB (HBM2e) run Qwen3.5-122B-A10B (MoE) locally?

A single 海光 DCU K100-AI (61.4 GB usable VRAM) cannot fully fit Qwen3.5-122B-A10B (MoE) at Q4_K_M (77.26 GB required). Consider multi-GPU tensor parallelism, a smaller parameter variant, or cloud GPU / API deployment.

What is the best quantization for Qwen3.5-122B-A10B (MoE) on 海光 DCU K100-AI?

The highest-quality quantization that fits cleanly on 海光 DCU K100-AI is GGUF Q2_K (Extreme Compression), using 50.82 GB VRAM at ~153.7 tokens/s with up to 256K context.

How much VRAM does Qwen3.5-122B-A10B (MoE) need in full FP16 precision?

In unquantized FP16/BF16 precision (16K context), Qwen3.5-122B-A10B (MoE) requires 250.7 GB VRAM (244 GB for weights alone). Because it uses a Mixture-of-Experts (MoE) architecture, all 122B parameters must sit in memory, while only 10B active parameters are read per token.