Google Unveils Most Powerful AI AcceleratorTPU v5p
The new TPU, coupled with the concept of an 'AI hypercomputer', aims to revolutionize AI training and large language model handling.
- Google has unveiled its most powerful Tensor Processing Unit (TPU) yet, the v5p, designed to accelerate AI training and reduce the time commitment associated with training large language models.
- The TPU v5p can push 459 teraFLOPS of bfloat16 performance or 918 teraOPS of Int8, backed by 95GB of high bandwidth memory capable of transferring data at a speed of 2.76 TB/s.
- As many as 8,960 v5p accelerators can be coupled together in a single pod using Google's 600 GB/s inter-chip interconnect to train models faster or at greater precision.
- Google also introduced the concept of an 'AI hypercomputer', a supercomputing architecture that employs a close knit system of hardware, software, ML frameworks, and consumption models.
- Google's new hardware and AI supercomputing architecture debuted alongside Gemini, a multi-modal large language model capable of handling text, images, video, audio, and code.
The Register 
Engadget 
TechCrunch 
