2017 Google TPU Chips Update: Google's Second Generation Tensor Processing Units Is Crazily Fast
At the company's I/O 2017 developer's conference, Google unveiled the second-generation Tensor Processing Unit and it is capable of delivering a lot of computing power. Here are the details.
How Fast Is The Second Generation Google TPU?
Google claimed that the each second-generation TPU can deliver up to 180 teraflops of performance. Still, fans have to wait and see what the average benchmarks look like as Google are already a step forward in terms of speed.
The machine learning is normally carried out by GPUs made by NVIDIA. However, Google has decided to build some its own hardware and optimize it to work well with its software, according to WCCF Tech.
Tech Crunch reported when it comes to the machine learning purposes, the second generation Google TPU are supposed to be more capable than CPUs or even GPUs. Furthermore, it is faster than the upgrade from the first generation of chips the company released at last year's I/O.
2017 Google TPU Chips Update: Google's Second Generation TPU Details
The first generation TPU was only able to handle inference. According to The Verge, the latest TPU can also be used for training machine learning models, a significant part of the machine learning workflow all within this single, powerful chip.
That means that the users can build a machine learning model. For example, to correctly identify an object in a photo is a tree, a car or a cat.
Inference in machine learning refers to the statistical likelihood that the machine's conclusions are correct. For example, based on the model, the users may be 85 percent confident that this is actually a tree and not a broccoli stock.
Google is now packaging availability of this second generation Google TPU chips as a service on the Google Cloud Platform. They are also lowering the barrier to entry to this technology.
The second generation Google TPU allowed users to start building their models on competing chips like Intel's Skylake or GPUs like Nvidia's Volta. Hence, they will move the project to Google's TPU cloud for final processing.
© 2017 iSports Times All rights reserved. Do not reproduce without permission.