Lockless and ordered, parallel chain reaction processing
-
Updated
Dec 28, 2022 - C++
Lockless and ordered, parallel chain reaction processing
Ryzentosh - Hackintosh based on TRX40-E and Threadripper 3960x
MojoLlama is a high-throughput inference engine for CPU, built on Modular MAX. GGUF native, MoE-optimized, with support for 50+ architectures — from Llama to Gemma 4 to hybrid SSM models. GPU acceleration via MAX engine for supported models.
Not a product. Not a framework. Nothing here is packaged for you to deploy. This is my house, my desk, my power bill, and the machine that thinks in it. It exists so I can point at something on my wall and say that box runs my world.
AMD rcraid Linux 6.14 DKMS port, NVMe RAID0 setup, mdadm/XFS tuning and fio benchmarks for high-speed workstation storage.
Add a description, image, and links to the threadripper topic page so that developers can more easily learn about it.
To associate your repository with the threadripper topic, visit your repo's landing page and select "manage topics."