Ryan's coverage of the new Polaris architecture will be up momentarily but in the meantime you can take a peek at The Tech Report's coverage here. The new architecture will utilize FinFETs of an unspecified process node and is designed to power the new UHD displays and VR headsets due for release over this coming year. Raja Koduri discusses the two major goals of the new architecture, fast pixels and deep pixels. Fast pixels refers to the awe inspiring amount of bandwidth required to draw on UHD displays, twin 4K displays would require addressing 1.8 gigapixels per refresh which would certainly need some fast pixels. Deep pixels refers to improved support for variable refresh rates and likely encompasses support for the new HDR technology we will see appear on the market in the near future. If you can't hold off your curiosity for our coverage you can pop over here.
"AMD will release new Radeons built on its next-gen Polaris architecture in mid-2016. We got an early look at this new architecture and AMD's plans for building these chips with FinFETs last month at the company's Radeon Technologies Group tech summit."
Here is some more Tech News from around the web:
- The AMD Polaris GPU Architecture Preview @ Hardware Canucks
- Steam says a config error and DoS attack caused sensitive data leak @ The Inquirer
- Happy 2016, and here's the year's first ransomware story @ The Register
- Twitter To Revive Politwoops, Archive of Politicians' Deleted Tweets @ Slashdot
- Best Hardware of 2015 – The KitGuru Editorial Awards
- Windows 10 is now running on 10 percent of all PCs and laptops @ The Inquirer
- Netgear ProSAFE XS728T 24-Port 10GbE Ethernet Switch @ techPowerUp
Shouldn’t it be “GCAN”?
Shouldn’t it be “GCAN”? Graphics Core After Next? 😉
I’m for just calling it
I’m for just calling it Polaris, and forgetting the GCN to avoid any confusion! But at least AMD does its Asynchronous compute fully in hardware and Polaris will be coming with improved ACE units, and improved Asynchronous compute, as well as improved tessellation units. Nvidia still is not fully in the hardware with its GPU processor thread dispatch and management, and if Nvidia does not have hardware based thread dispatch and management fully in hardware by Pascal, Nvidia will be seriously behind AMD for the VR gaming market!
It’s no use for Nvidia to have their underutilized/idle GPU execution resources for lack of proper hardware based GPU processor thread dispatch and management is it if they can not be efficiently kept busy! That software based GPU processor thread dispatch and management is never going to be as efficient as the fully hardware based GPU processor thread dispatch and management of AMD’s GCN, and now Polaris based, GPU micro-architectures are.
GPU processor thread dispatch
GPU processor thread dispatch and management
hardware based thread dispatch and management
hardware based GPU processor thread dispatch and management
software based GPU processor thread dispatch and management
hardware based GPU processor thread dispatch and management
Okay listen. I’m as much of an AMD fan as anybody. But seriously, can I get you some pom-poms?