hi@aiweekly.co.in

MoonMath AI Open-Sources a HIP Attention Kernel for AMD MI300X That Beats AITER v3 on Every Shape and Rounding Mode

MoonMath AI team has released a bf16 forward attention kernel for AMD’s MI300X GPU. It is written in HIP, not hand-written assembly. The code is open-source under the MIT license. The MoonMath.ai team reports it beats AITER v3, AMD’s own optimized kernel, on every tested shape. Bare-metal access came from HotAisle, an AMD cloud provider.

MoonMath AI Open-Sources a HIP Attention Kernel for AMD MI300X That Beats AITER v3 on Every Shape and Rounding Mode Read More »

ET AI Hackathon 2.0 enters phase 2, challenging builders to turn ideas into AI-powered solutions

ET AI Hackathon 2.0 has entered Phase 2, where shortlisted participants will transform ideas into AI-powered prototypes by tackling real-world problem statements. The stage focuses on innovation, technical execution, and practical impact, challenging teams to build solutions that address meaningful challenges.

ET AI Hackathon 2.0 enters phase 2, challenging builders to turn ideas into AI-powered solutions Read More »