Blog

In this blog post, we’ll continue our journey to build a state-of-the-art (SOTA) matmul kernel on NVIDIA Blackwell by exploring the cluster launch control (CLC) optimization. At the end of the post we’ll improve our performance by another 15% and achieve 1772 TFLOPs, exceeding that of the current SOTA.

In this post, we continue on this journey and discuss how to leverage the 2SM technique along with pipelining to increase our performance about 5x and get within 85% of state-of-the-art (SOTA).

This series of blog posts will showcase how one can: 1. Write a high-performance GPU kernel on Blackwell that offers performance competitive to that of NVIDIA's cuBLAS implementation. 2. Shows how one can leverage Mojo's special features to make the kernel as simple as possible.

Agentic Building Blocks: Creating AI Agents with MAX Serve and OpenAI Function Calling

Learn how quickly MAX and Open WebUI get you up-and-running with RAG, web search, and Llama 3.1 on GPU

Mojo 24.6 introduces key improvements in argument conventions, memory management, and reference tracking, enhancing code clarity and safety with features like 'mut' for mutable arguments, 'origins' for references, and new collection types.

Imagine unlocking a world of open innovation while ensuring secure, reliable, and enterprise-ready Gen AI deployments—MAX 24.6 enables enterprise AI teams to seamlessly run a vast range of cutting-edge AI models from Hugging Face on NVIDIA GPUs.

Measuring state of the art GPU performance compared to vLLM on Modular's MAX 24.6

Build a Chat Application with Llama 3 and MAX Serve