Radeon VII 16GB HBM 2 memory cost around $320
Published in Graphics
Saturday, 26 January 2019 08:55

Radeon VII 16GB HBM 2 memory cost around $320


Close to half of the MSRP

The marketing guys at AMD did a nice job naming the card Radeon VII (Roman 7) as it is a 7nm, launching on February 7 and costing $699 but our investigation shows that 16GB HBM 2 comes at a hefty price.

Xilinx helped AMD with HBM 2
Published in AI
Friday, 15 June 2018 12:34

Xilinx helped AMD with HBM 2


High bandwidth is the key

There is an interesting story that's very little known about the public cooperation that has happened between AMD and Xilinx about HBM memory.

Alibaba gets Xilinx FPGA for its F3 cloud
Published in Cloud


Faster deep learning and IO

Back in October 2017 Alibaba, a $502,179 billion company announced that it had chosen Xilinx for next generation FPGA acceleration in their public cloud.   

Intel to have one more Radeon adventure
Published in AI
Thursday, 30 November 2017 08:17

Intel to have one more Radeon adventure


10nm Intel part

Kaby Lake G might not be the only flirtation between Intel and AMD (RTG), there are some future parts in the pipeline,  too.

When Intel 8 Gen Core HBM 2 meets the GPU it's the P22
Published in AI


Vega based coming in Q1 18 ish

Intel has been working on what was internally called the Paolo Alto P22 project and that was the Vega HBM 2 based GPU that takes closely to Intel’s 8th Generation CPU.

Intel announces it will use Vega for notebooks
Published in Graphics


Fudzilla told you so in May 2017

Back in May, Fudzilla wrote about the fact Intel will work with AMD on graphics and guess what? We were right.

AMD Vega 64, water cooled and 56 revealed
Published in Graphics


Official specs, pricing but no bench

AMD has shared a few updates on the highly anticipated HBM 2 based Vega architecture and we finally know the official specs.

The next Geforce doesn’t use HBM 2
Published in Graphics
Tuesday, 13 June 2017 10:46

The next Geforce doesn’t use HBM 2


Too early for this memory

Volta is out for artificial intelligence, machine learning applications and it will be shipping in DGX 1 systems, mainly for deep learning and AI. The next Geforce will be a  completely separate chip. 

AMD's Vega GPU is late
Published in Graphics
Thursday, 01 June 2017 03:21

AMD's Vega GPU is late


July 30 is the first day

Vega is delayed and while AMD said it will arrive in Q2 2017, it won't. 

Vega HBM 2 8GB memory stack costs $160
Published in Graphics
Wednesday, 24 May 2017 11:35

Vega HBM 2 8GB memory stack costs $160


BOM  goes the bank balance 

It is hard to forget the fact that Vega is more than a year late. It was supposed to launch last year and AMD had to settle for Polaris 10 and 11 performance mainstream / mainstream parts.