NVIDIA Announces Tesla P100 PCI Express Card For Deep Learning And HPC, Shipping Q4

Remember that scene when Al Pacino playing the part of Scarface famously said, “Say hello to my little friend!”? Well now NVIDIA is bringing out the big gun in PCI Express form in what it considers the “most advanced data center GPU ever built,” the Tesla P100, a crazy fast GPU accelerator based on NVIDIA’s powerful Pascal architecture built on a 16nm FinFET manufacturing process.This is the second iteration of the Tesla P100, the first of which NVIDIA unveiled during GTX 2016 a couple of months ago. Though it’s built on a fancy 16nm FinFET process, the GP100 GPU inside the Tesla P100 is quite large with a 600mm2 die, roughly the size of previous generation high-end Maxwell GPUs. However, it’s more advanced than Maxwell with a whopping 15.3 billion…


Link to Full Article: NVIDIA Announces Tesla P100 PCI Express Card For Deep Learning And HPC, Shipping Q4