A custom-built chip for machine learning from Google. Introduced in 2016 and found only in Google datacenters, the Tensor Processing Unit (TPU) is optimized for matrix multiplications, which are essential computations in neural networks. TPUs do not support the TF32 floating point format (see
TF32).
Running software from Google's TensorFlow AI library, TPU circuits are designed to process huge volumes of low-precision math very quickly for neural network processing.
TPU Versions and Edge
TPUv1 through TPUv4 are the first four versions of the chip from 2016 to 2021. In 2018, a lighter-weight Edge v1 chip designed for AI processing was used in Google's Pixel 4 smartphone (see
Tensor chip). See
Tensor core,
TensorFlow,
PaLM and
neural network.