At least five are deployed in Europe
We've had confirmation that at least five super expensive, 170 TeraFlops servers with eight P100 cards, each having a 16GB HBM 2, two Xeon processors, 512 GB RAM memory have been reaching its customers.
Rent deep learning
Scan in the UK plans to offer the Nvidia DGX-1 deep learning super computer to buy or even to rent. This is a unique approach as many deep learning startups will get the chance to try the system before they buy it. And there are many universities which would love to get their hands on 170 teraFlops of deep learning performance.
For professional market
Our well-informed sources have confirmed that Vega 10, AMD’s first HBM 2 card will by announced before the end of this year, at least for the professional market.
Ai Supercomputer with SoC 7 B transistors
Nvidia CEO Jen-Hsun Huang has just announced a 2017 processor s codenamed Xavier to succeed Parker.
Four cores with eight threads
A few of you might remember that we exclusively posted the news that AMD is working on a 7nm CPU codenamed Starship. The 7nm APU is codenamed Gray Hawk and it aims to attain lower TDPs.
Next, next generation
Vega 10 will evolve to Vega 20 in 2017 and 2018 and this will be AMD’s next big thing. It is codenamed Navi and will be a 7nm FinFET GlobalFoundries product that should surface in late 2018 or early 2019.
Replaces Vega 10, 32GB HBM 2
Vega 10 is Radeon Technology Group’s first HBM 2 part and should launch in early 2017. AMD – RTG is also planning a successor called Vega 20 that might launch later in the year.
Xiaomi Note 2 Pro with Snapdragon 821
Xiaomi will launch two new phones on 27 September and both will have the snappiest Snapdragon 821 SoC.
Slightly faster than P100
Vega 10 is coming next year and it seems that it will be faster than the P100 card. P100 (21.2 TFlops) is a supercomputer part and has much higher theoretical performance than GP102 (10.97 TFlops). According to the raw numbers, AMD Vega 10 with 24Tflops, Vega 10 will significantly outperform Geforce Titan X.