Hardware
32 items on noze.it
Articles (32)
16 July 2026Codex Micro and the AI hardware that works: augment, don't replace
OpenAI launches the Codex Micro, a $230 macropad (with Work Louder) to drive fleets of Codex agents: its first branded h...
14 July 2026colibri: GLM 5.2 (744B) on 25 GB of RAM, streaming the experts from disk
colibri, a C inference engine released by JustVugg (Apache 2.0), runs GLM 5.2, a 744-billion-parameter MoE model, on a c...
20 January 2026NVIDIA GB10: local AI arrives on the desktop with 128 GB unified memory
NVIDIA GB10 Grace Blackwell: desktop workstations (DGX Spark, ThinkStation PGX) with 128 GB unified memory and 1 PFLOP F...
18 March 2025NVIDIA DGX: the integrated platform for AI training and inference
Evolution of NVIDIA DGX platforms: from DGX-1 (2016, P100) to DGX-2 (2018, Volta + NVSwitch), DGX A100 (2020), DGX H100 ...
14 June 2024Nemotron: NVIDIA's open models
NVIDIA releases Nemotron-4 340B under the NVIDIA Open Model License: a Base + Instruct + Reward triplet trained on 9 tri...
12 May 2024Edge AI on Raspberry Pi: local LLMs with Ollama and llama.cpp
In 2024 quantised LLM execution becomes practical on Raspberry Pi 5 and Jetson Orin Nano thanks to llama.cpp and Ollama....
18 March 2024NVIDIA: evolution of GPU architectures from Tesla to Blackwell
Technical history of NVIDIA GPU architectures: GeForce 256 (1999), Tesla and CUDA (2006), Fermi, Kepler, Maxwell, Pascal...
20 November 2023Arduino and Python for IoT: healthcare and industry deploy
Integration pattern Arduino/ESP32 + MQTT + Python + cloud or on-premise backend for IoT projects in healthcare (environm...
15 September 2023Ollama: Open Source language models on your own hardware
Ollama simplifies local execution of Open Source LLMs: Modelfile for configuration, REST API, GGUF quantization for cons...
8 December 2022CUDA: the programming model for NVIDIA GPUs
CUDA as SIMT programming model for NVIDIA GPUs: grid, block, thread, warp, memory hierarchy, streaming multiprocessors. ...
6 June 2022Apple Metal: graphics and compute API for iOS, macOS and Apple Silicon
Apple Metal as a low-overhead graphics and compute API: WWDC 2014 announcement (iOS 8), Metal 2 (2017), Metal 3 (WWDC 20...
1 May 2022YK-Robotics: multi-brand robotic programming via web
The collaboration with YK-Robotics begins for the development of a multi-brand robotic programming software with a cloud...