Open-Source Software & Data

We release open artifacts so that others can reproduce, extend, and build on our work. Most of them are led and maintained by students in the lab.

Open-Source Libraries & Tools

libCacheSim High-performance cache library and simulator

A high-performance, open-source C++ library designed to simulate and analyze cache systems efficiently. It provides a flexible framework for exploring cache design choices and evaluating their impact on performance. Used by more than 100 research institutes and companies worldwide.

libCacheSim-python Python bindings for libCacheSim

Python bindings for libCacheSim, designed for rapid cache-simulation experiments. Install with pip install libcachesim.

RouteWise Latency-cost optimization library for LLM routers

An open-source library that LLM routers can use to optimize cost and latency across multiple providers, from our EuroSys'27 paper. Available on PyPI as llm-routewise.

Datasets

Cache trace dataset Open collection of production cache traces

A large open collection of key-value, object/CDN, and block cache traces (2007–2025) from Meta, Twitter, Alibaba, Tencent, and others, including the Twitter production traces from our OSDI'20 measurement study. Widely used in cache research and evaluation.

LLM serving trace dataset One year of production LLM serving traces (to be released)

A year-long production LLM serving trace collected from Chutes, spanning many models and users, supporting research on workload evolution, caching, and load balancing in LLM serving.

Services

FreeInference Free inference API for open-source models

A free, OpenAI- and Anthropic-compatible inference API for frontier open-source models, aimed at research, teaching, and prototyping.

traceViz Workload visualization tool

A web tool for visualizing and exploring storage and cache workloads.

Open Infrastructure

FreeInference dashboard Public GPU and serving dashboard

A public dashboard for the GPUs and serving engine behind FreeInference.