An open-source parameterizable NPU generator with full-stack multi-target compilation stack for intelligent workloads.

LLMServingSim: A HW/SW Co-Simulation Infrastructure for LLM Inference Serving at Scale

Inference Llama 2 in one file of pure C