llama.cpp Slower Than Ollama? Local LLM Deployment Config Pitfalls
anshulsingh8326 · reddit · 2026-08-09
A developer testing on an RTX 4070 (12GB) found that running the same Q4K quantized model in llama.cpp yielded a lower generation speed (55 TPS) compared to Ollama's default configuration (61 TPS).
The poster shared detailed launch arguments (e.g., -ngl 99, -fa, -ctk q80) and hardware specs to ask for help. This highlights how default parameter optimizations and context processing mechanisms across different local tools significantly impact performance, and manual tuning can easily lead to bottlenecks.
More from Infra
- Matt Turck: Tech Industry Shouldn't Ignore Resistance to AI Data Centers — mattturck · 2026-08-09
- ComfyUI Quantization Toolkit Supports W4A8: 40% Drop in VRAM Usage — External_Quarter · 2026-08-09
- Running MiniMax H3 Video Generation Locally: 4070Ti Acceleration Test — harunandro · 2026-08-09
- DeepSeek-V4-Flash Local Inference Optimized: 27% Speedup on a Single DGX Spark — antirez · 2026-08-09
- Developer Releases LlamaPort: A macOS GUI Client for llama-server — Current-Quail-2503 · 2026-08-09
- Zero-Dependency C99 BitNet Inference Hits 36 tok/s on Intel Xeon CPU — shifu_legend · 2026-08-09