Google and Amazon launch tools to reduce reliance on Nvidia
In brief
- Google releases TorchTPU, an open-source tool to run PyTorch on its TPUs.
- Amazon launches TorchNeuron to migrate PyTorch apps to its Trainium chips on AWS.
- Both tools aim to reduce Nvidia’s grip on the PyTorch ecosystem.
On Tuesday, Google announced the open-source release of TorchTPU, a PyTorch version optimized for its Tensor Processing Units (TPUs). According to the announcement, the tool lets developers easily move PyTorch apps to Google’s TPUs without rewriting code.
PyTorch, used by 85% of AI apps, was originally developed by Meta in 2016 and is now managed by the Linux Foundation. Google is collaborating with Meta, Hugging Face, and others to ensure TorchTPU’s compatibility with their libraries. Amazon made a similar move the day before with TorchNeuron, its tool for its Trainium chips on AWS. Both companies also offer dedicated toolkits to simplify migration and debugging.
These efforts aim to challenge Nvidia’s dominance in the PyTorch ecosystem, where most apps are designed for its GPUs. However, the article notes that developers won’t immediately ditch Nvidia en masse, though these tools could gradually diversify hardware options for AI workloads.
This summary gives you the essentials. The full article, its figures and quotes are on AI News & Artificial Intelligence | TechCrunch.
Read the original articleAI-generated summary of the source, checked by rules. How we write · Report an error