Gemma-4-31B-IT on 海光深算 DCU K100-AI 64GB (HBM2e): VRAM Requirement & Speed Estimate
Yes. At Q4_K_M 4-bit quantization with a 16K context window, Gemma-4-31B-IT requires 22.47 GB of VRAM (18.61 GB weights + 2.18 GB KV cache + 1.67 GB overhead), fitting within 海光 DCU K100-AI's 61.4 GB usable memory at ~32.7 tokens/s.
Physical 64 GB · Load 36.6%
Actual speed varies ~±25% by runtime
Enable Q8/Q4 KV quantization to scale batch concurrency
OLLAMA_CONTEXT_LENGTH=16384 ollama run gemma4:31b海光 DCU K100-AI Fits Gemma-4-31B-IT · Cost & TCO
Approx. $8,000 (~$125/GB VRAM) · 350W TDP
📊 All Quantization Levels for Gemma-4-31B-IT on 海光 DCU K100-AI
| Quantization | Total VRAM (16K) | Verdict | Est. Speed | Max Context |
|---|---|---|---|---|
| FP16 / BF16 | 66.11 GB | Partial RAM Offload (Slow) | ~6.5 tok/s | — |
| FP8 / Q8_0 | 36.75 GB | Smooth Fit | ~19.1 tok/s | 256K |
| GGUF Q6_K | 29.16 GB | Smooth Fit | ~24.6 tok/s | 256K |
| GGUF Q5_K_M | 25.75 GB | Smooth Fit | ~28.1 tok/s | 256K |
| Q4_K_M / AWQ | 22.47 GB | Smooth Fit | ~32.7 tok/s | 256K |
| GGUF IQ4_XS | 20.12 GB | Smooth Fit | ~37.1 tok/s | 256K |
| GGUF Q3_K_M | 18.79 GB | Smooth Fit | ~40.1 tok/s | 256K |
| GGUF Q2_K | 15.81 GB | Smooth Fit | ~49 tok/s | 256K |
💡 Recommended Alternative GPUs & Macs for Gemma-4-31B-IT
Frequently Asked Questions
Can 海光深算 DCU K100-AI 64GB (HBM2e) run Gemma-4-31B-IT locally?
Yes. At Q4_K_M 4-bit quantization with a 16K context window, Gemma-4-31B-IT requires 22.47 GB of VRAM (18.61 GB weights + 2.18 GB KV cache + 1.67 GB overhead), fitting within 海光 DCU K100-AI's 61.4 GB usable memory at ~32.7 tokens/s.
What is the best quantization for Gemma-4-31B-IT on 海光 DCU K100-AI?
The highest-quality quantization that fits cleanly on 海光 DCU K100-AI is FP8 / Q8_0 (99.9% Near-Lossless), using 36.75 GB VRAM at ~19.1 tokens/s with up to 256K context.
How much VRAM does Gemma-4-31B-IT need in full FP16 precision?
In unquantized FP16/BF16 precision (16K context), Gemma-4-31B-IT requires 66.11 GB VRAM (61.4 GB for weights alone).