Exascale Inference

Deep Learning at Scale @ SC25

Agent2Agent (A2A) is an open protocol enabling communication and interoperability between opaque agentic applications.

Code for SC24 Demonstartion

A parameterizable Weight-Stationary 2D Systolic Array implemented in Verilog for hardware-accelerated matrix multiplication

The first open-source agentic AI physicist, by Physical Superintelligence PBC (PSI).

TorchCFM: a Conditional Flow Matching library

Train transformer language models with reinforcement learning.

A C++ low Mach number stellar hydrodynamics code

Library for specialized dense and sparse matrix operations, and deep learning primitives.

Intel® Tensor Processing Primitives extension for Pytorch*

Official inference framework for 1-bit LLMs

Linux hwmon driver for the NVIDIA DGX Spark (GB10 SoC) that exposes full system power telemetry via standard sensors / sysfs interfaces.

🎯 aimrocks 🎸 — python & cython bindings for RocksDB. Batteries included! 🔋

Repo for issue tracking, documents and release. Learn more about MEDYAN at medyan.org.

Heterogeneous programming in Julia

Easy to maintain open source documentation websites.

TT-Metalium fork of the wolfpld/tracy repo on github

Frame profiler

🤘 TT-NN operator library, and TT-Metalium low level kernel programming model.

A simple, lightweight PowerShell script that allows you to remove pre-installed apps, disable telemetry, as well as perform various other changes to declutter and customize your Windows experience.…

Unified Policy as Code

A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.