Loading…
Tuesday February 4, 2025 11:45am - 12:25pm GMT
Modern AI and scientific computing come down to one core operation: matrix multiplication. Despite spending billions annually on data centers, we’re missing a massive opportunity to cut costs and boost performance by leveraging specialized assembly instructions and hardware-specific optimizations—rather than relying solely on compilers. In this talk, I’ll reveal powerful techniques that optimize matrix multiplication across a diverse range of platforms—including Intel, Apple, AMD, NVIDIA, and AWS chips. These strategies aren’t just about speed; they’re about democratizing access to high-performance computing, making AI and scientific research more efficient and accessible than ever before.
Speakers
avatar for Ash Vardanian

Ash Vardanian

Founder, Unum
Ash Vardanian is the founder of Unum, an advanced computing company developing primarily open-source infrastructure that is used by many modern database companies, as well as frontier AI labs and governments around the globe. With a background in Astrophysics, his current work focuses... Read More →
Tuesday February 4, 2025 11:45am - 12:25pm GMT
Doddington Forum 1
Feedback form is now closed.

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Share Modal

Share this link via

Or copy link