Technical Deep Dive
The Next IPU: Supercharge your AI Performance with Bow Systems
Join our deep dive to explore how Bow, the world's first 3D wafer-on-wafer processor, delivers huge performance gains for AI at scale.
Whether you’re tackling scientific problems at supercomputing scale, exploring graph neural networks or implementing GPT-based language applications, new Graphcore Bow systems, based on the world’s first 3D wafer-on-wafer processor, deliver a huge power and efficiency boost to enable AI acceleration at scale.
Bow’s new wafer-stacking technology allows enterprises and research organisations to push the boundaries of machine learning innovation, with 40% performance gains and 16% better power efficiency compared with its predecessors.
Join Graphcore SVP Software Matt Fyles to discover how new Bow systems’ substantial performance boost will help innovators accelerate and explore next generation models and, ultimately, build the future of AI.
Key takeaways:
Meet the speaker
Matt Fyles
Matt Fyles is a computer scientist with over 20 years of proven experience in the design, delivery and the support of software and hardware within the microprocessor market. As SVP Software at Graphcore, Matt has built the company’s Poplar software stack from scratch, co-designed with the IPU for machine intelligence. He currently oversees the Software team’s work on the Poplar SDK, helping to support Graphcore’s growing community of developers.