What is LPDDR4?
Low-Power DDR4 — often soldered memory in mini PCs. Lower bandwidth than desktop DDR4 or DDR5. Limits tokens-per-second compared to high-end alternatives.
Full Explanation
LPDDR4 (Low-Power Double Data Rate 4) is a power-optimized RAM standard commonly soldered onto mini PC motherboards to reduce size and cost. It typically runs at 51–68 GB/s — about 10× less bandwidth than GDDR7. While sufficient for productivity workloads, this bandwidth ceiling means CPU-based LLM inference on LPDDR4 systems produces 6–10 tokens per second on 7B models, which feels sluggish for interactive chat.
Why It Matters for Local AI
Budget mini PCs using LPDDR4 are best suited as always-on background AI servers rather than interactive chat machines. For tasks that tolerate latency — document summarization running overnight, automation scripts — the low power draw and cost make them compelling despite the speed penalty.
Hardware Relevant to LPDDR4
mini-pc · Check Price on Amazon · 16 GB Unified · 34 GB/s
mini-pc · Check Price on Amazon · 16 GB Unified · 34 GB/s
Related Terms
Memory Bandwidth→
How fast data moves between memory and the processor, measured in GB/s. Tokens per second scales nearly linearly with bandwidth — this is the single most important GPU spec for LLM speed.
Unified Memory→
Apple Silicon uses a single pool of fast RAM shared between CPU and GPU. Larger unified memory = larger models run entirely at full bandwidth — no PCIe bottleneck.
Tokens/s→
Tokens per second — the standard speed metric for LLMs. One token ≈ 0.75 words. Above 10 t/s feels interactive; below 5 t/s feels like watching paint dry.
CPU Inference→
Running LLMs on the CPU rather than a GPU. Works on any hardware, no special drivers needed. Limited to ~8–12 t/s on 7B models — fine for background tasks, slow for interactive use.