CPP

SonicBoom – llama.cpp's own kernels, 14% faster decode via a graph compiler
4 points | 0 comments

Jeyzma is TypeSafeAI minus $850M plus Go/TinyGo/llama.cpp in the browser
3 points | 0 comments

PS5LM: Llama.cpp on a jailbroken PS5, run any GGUF model natively on the console
2 points | 0 comments

If llama.cpp enables MoE to activate multiple experts
1 points | 0 comments

TypeScript Compiler ported to Haxe(node, cpp, JVM, C# & Python)
1 points | 0 comments

Show HN: LoRA over GGUF – Train Qwen3.8-Flash-Next in 40G VRAM
1 points | 0 comments

C++26 Contracts: Introduction to Preconditions and Postconditions
140 points | 156 comments