With beefier TPUs, Google wants to lay claim to Machine Learning kingdom

Amos Gonzales
May 18, 2017

The first TPU, shown off a year ago as a special-purpose chip designed specifically for machine learning, is used by the AlphaGo artificial intelligence system as the foundation of its predictive and decision-making skills. That means machine learning customers that might have rented NVIDIA GPUs on Amazon Web Services, Microsoft, Azure, IBM's Bluemix, or elsewhere, now have the option of using Google's purpose-built solution. In the case of machine vision, the company announced its new Google Lens initiative, which aims to help machines understand the world in a way similar to how humans do.

Today at its I/O event, Google has announced that its second generation Tensor Processing Units are coming to the Google Cloud.

The company is introducing these TPUs to Google Compute engine as Cloud TPUs. But Google has opted over the last few years to build some of this hardware itself and optimize for its own software.

Simply put, this is bad news NVIDIA and other chipmakers that count on Google and others to buy their specialized processors. These devices are created to work in larger systems, for example a 64-TPU module pod can apply up to 11.5 petaflops of computation to a single ML (machine learning) training task. Google says each TPU module, as pictured above, can deliver up to 180 teraflops of floating-point performance.

Trump says Russia probe will reveal no collusion with his campaign
President Trump also responded, saying "There was no collusion . and I look forward to this matter concluding quickly". Kessler's latest book is " The First Family Detail: Secret Service Agents Reveal the Hidden Lives of the Presidents ".

In Dean's and Hölzle's blog, they describe the improvement in training times that the new TPUs have delivered: "One of our new large-scale translation models used to take a full day to train on 32 of the best commercially-available GPUs-now it trains to the same accuracy in an afternoon using just one eighth of a TPU pod", they write. One would assume additional frameworks like Caffe, Torch, Theano, and others will get ported over time, should the TPU attract a critical mass of users.

Beyond speed, the second-gen TPU is also going to allow Google's servers to do what is known as both inference and training simultaneously. The effect is easily visible across Google's wide portfolio of products including Google Translate, Photos and its Go Champion program AlphaGo. The newly hatched TPU is being used to accelerate Google's machine learning work and will also become the basis of a new cloud service. We'll be updating this article with more details.

Businesses will be able to use the new chips through Google's Cloud Platform, as part of its Compute Engine infrastructure-as-a-service offering. In order to qualify, researchers must publish their findings and potentially make their research's software available for free for others to access in an open-source model, he said.

Other reports by BadHub

Discuss This Article