---
title: "#llamacpp articles | Hashnode"
description: "Explore 29 articles tagged #llamacpp on Hashnode. Tutorials, guides, and how-tos from the community."
url: "https://hashnode.com/tag/llamacpp"
---

Tag

# #llamacpp

29 posts · 3 followers

[Gayatri Panse](https://hashnode.com/@gradientnotes) · [1d ago](https://hashnode.com/posts/backprop-01-where-did-the-memory-go/6afd92054f515efa31a3329a)

### [Backprop 01: Where Did the Memory Go?](https://gradientdecent.hashnode.dev/backprop-01-where-did-the-memory-go?utm_source=hashnode&utm_medium=feed)

Why the same Gemma 4 12B fit 1.83 users on vLLM and about 13 on llama.cpp

24 min read · [gradientdecent.hashnode.dev](https://gradientdecent.hashnode.dev/backprop-01-where-did-the-memory-go?utm_source=hashnode&utm_medium=feed)

[Internals Decoded](https://hashnode.com/@internalsdecoded) · [3d ago](https://hashnode.com/posts/local-inference/6afb0c6e57170ddd3bb9be08)

### [Local Inference: Running Models on Your Own Machine](https://internals-decoded.hashnode.dev/local-inference?utm_source=hashnode&utm_medium=feed)

Ollama, llama.cpp, and what hardware actually gets you usable speed.

16 min read · [internals-decoded.hashnode.dev](https://internals-decoded.hashnode.dev/local-inference?utm_source=hashnode&utm_medium=feed)

[Gayatri Panse](https://hashnode.com/@gradientnotes) · [Oct 13](https://hashnode.com/posts/epoch-1-try-it-time-it-then-buy-it/6ace36b10c270ecc322ef88d)

### [Epoch 1: Try It, Time It, Then Buy It](https://gradientdecent.hashnode.dev/epoch-1-try-it-time-it-then-buy-it?utm_source=hashnode&utm_medium=feed)

How three rounds of Gemma 4 experiments shaped our choice of GPU, inference engine and model

23 min read · [gradientdecent.hashnode.dev](https://gradientdecent.hashnode.dev/epoch-1-try-it-time-it-then-buy-it?utm_source=hashnode&utm_medium=feed)

[donald991234](https://hashnode.com/@donald991234) · [Oct 12](https://hashnode.com/posts/qwen3-8-27b-vram-measured-q4-k-m-uses-19-95-gib-at-32k-context/6accd6017fa5210e386410e5)

### [Qwen3.8 27B VRAM, measured: Q4\_K\_M uses 19.95 GiB at 32K context](https://modelvram.hashnode.dev/qwen3-8-27b-vram-measured-q4-k-m-uses-19-95-gib-at-32k-context?utm_source=hashnode&utm_medium=feed)

Real llama.cpp numbers, why the KV cache is only 64 KiB per token, and which GGUF fits a 16, 24 or 32 GB GPU

5 min read · [modelvram.hashnode.dev](https://modelvram.hashnode.dev/qwen3-8-27b-vram-measured-q4-k-m-uses-19-95-gib-at-32k-context?utm_source=hashnode&utm_medium=feed)

[ZyVOP](https://hashnode.com/@zyvop) · [Sep 25](https://hashnode.com/posts/can-a-2000-mini-pc-replace-your-ai-cloud-bill/6ab6a6c8488375a31a5b53f0)

### [Can a $2,000 Mini PC Replace Your AI Cloud Bill?](https://blog.zyvop.com/can-a-2000-mini-pc-replace-your-ai-cloud-bill?utm_source=hashnode&utm_medium=feed)

AMD's Strix Halo chips can run 120B-class models locally. Here's what the benchmarks, power costs, and real-world agent workflows actually look like.

11 min read · [blog.zyvop.com](https://blog.zyvop.com/can-a-2000-mini-pc-replace-your-ai-cloud-bill?utm_source=hashnode&utm_medium=feed)

[Mohammad Hatif Osmani](https://hashnode.com/@mhosm) · [Sep 7](https://hashnode.com/posts/building-llamini-cpp-a-from-scratch-llama-cpp-core-in-c99-zero-dependencies/6a9f17866f526dd1c8ea4868)

### [Building llamini.cpp: a from-scratch llama.cpp core in C99, zero dependencies](https://mhosm.hashnode.dev/building-llamini-cpp-a-from-scratch-llama-cpp-core-in-c99-zero-dependencies?utm_source=hashnode&utm_medium=feed)

At a glance Dialect Pure C99 (POSIX + libc) — 0 third-party dependencies Build One command: make → llamini Models TinyLlama 1.1B, Qwen2.5 0.5B, GPT-2 124M, Gemma 2B Quants Q4\_K, Q6\_K, Q5

26 min read · [mhosm.hashnode.dev](https://mhosm.hashnode.dev/building-llamini-cpp-a-from-scratch-llama-cpp-core-in-c99-zero-dependencies?utm_source=hashnode&utm_medium=feed)

[Tanisha Soni](https://hashnode.com/@tanisha-soni) · [Jul 5](https://hashnode.com/posts/custom-langchain-chat-model-local-llm/6a4a28818251985fc678d4c8)

### [I Taught My Local Model to Speak LangChain](https://tanisha-soni.hashnode.dev/custom-langchain-chat-model-local-llm?utm_source=hashnode&utm_medium=feed)

How to extend BaseChatModel, implement tool calling, parse JSON responses, and run self hosted GGUF model with LangChain no other API required.

9 min read · [tanisha-soni.hashnode.dev](https://tanisha-soni.hashnode.dev/custom-langchain-chat-model-local-llm?utm_source=hashnode&utm_medium=feed)

[Alexander Kerchum](https://hashnode.com/@ajkerchum) · [Jun 5](https://hashnode.com/posts/lessons-from-the-bottom-of-the-stack-shipping-a-quant/6a2328f6857478689f6b067a)

### [Lessons From the Bottom of the Stack: Shipping a Quant](https://blog.kerchum.dev/lessons-from-the-bottom-of-the-stack-shipping-a-quant?utm_source=hashnode&utm_medium=feed)

Shipping a 4-Bit LLM Quant into llama.cpp

12 min read · [blog.kerchum.dev](https://blog.kerchum.dev/lessons-from-the-bottom-of-the-stack-shipping-a-quant?utm_source=hashnode&utm_medium=feed)

[Phaneesh Yalavarthi](https://hashnode.com/@Phaneesh) · [May 11](https://hashnode.com/posts/local-ai-running-gemma-4-with-llama-cpp-and-docker/6a01a188fca21b0d4bf5aed6)

### [Local AI: Running Gemma 4 with llama.cpp and Docker](https://phaneesh.hashnode.dev/local-ai-running-gemma-4-with-llama-cpp-and-docker?utm_source=hashnode&utm_medium=feed)

Local AI: Running Gemma 4 with llama.cpp and Docker

1 min read · [phaneesh.hashnode.dev](https://phaneesh.hashnode.dev/local-ai-running-gemma-4-with-llama-cpp-and-docker?utm_source=hashnode&utm_medium=feed)

[강문규](https://hashnode.com/@devsnack) · [May 6](https://hashnode.com/posts/qwen3-6-on-dgx-spark-vllm-nvfp4-dflash-vs-llama-cpp-2x-faster-at-88-104-tok-s/69fb2aa950ecad45331b9b3a)

### [Qwen3.6 on DGX Spark: vLLM + NVFP4 + DFlash vs llama.cpp — 2x Faster at 88–104 tok/s](https://devsnack.hashnode.dev/qwen3-6-on-dgx-spark-vllm-nvfp4-dflash-vs-llama-cpp-2x-faster-at-88-104-tok-s?utm_source=hashnode&utm_medium=feed)

I benchmarked Qwen3.6-35B-A3B on DGX Spark GB10 using llama.cpp (45–50 tok/s) vs vLLM + NVFP4 + DFlash (88–104 tok/s). Full setup guide, docker-compose, troubleshooting, and real-world results.

10 min read · [devsnack.hashnode.dev](https://devsnack.hashnode.dev/qwen3-6-on-dgx-spark-vllm-nvfp4-dflash-vs-llama-cpp-2x-faster-at-88-104-tok-s?utm_source=hashnode&utm_medium=feed)
