It’s von Neumann’s fault | GPU · TPU · LPU

Explains the memory bandwidth and scheduling tradeoffs behind GPU, TPU and LPU inference. That helps you reason about latency, cost and availability of the models you build on.

• inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.

• videoThis question inspired LoRA fine-tuning
• videoWhat makes a voice expressive?
• videoHow Continua AI Uses HyDE for RAG in Group Conversations
• videoWhat Are World Models?

• video"An endless demand for compute" | Jonathan Ross, founder of GroqJulia Turc
• videoGoogle's AI Infrastructure Chief, Amin Vahdat, on the Physics & Economics of Frontier AISequoia Capital
• videoReiner Pope of MatX on accelerating AI with transformer-optimized chipsCheeky Pint

Checking sign-in…

Loading comments…