In this talk from AI Infra @ Scale 2024, Joel Colburn, a software engineer at Meta, technical lead Junqiang Lan, and software engineer Jack Montgomery discuss the second generation of MTIA, Meta’s in-house training and inference accelerator.
They cover the co-design process behind building the second generation of Meta’s first-ever custom silicon for AI workloads, including the PyTorch software ecosystem, and the model architectures for Meta’s key applications. They demonstrate how MTIA achieves the performance, efficiency, and developer experience to successfully launch models into production. They also highlight several co-design examples where special silicon features are utilized to accelerate Meta’s models.
