Guest_Jim_* Posted January 4, 2016 Posted January 4, 2016 It is really nice that CES 2016 is finally here, because we are finally getting a look at some long-awaited technologies. Today AMD has unveiled its upcoming Polaris GPU architecture, which features a 4th generation GCN core and will be built with 14 nm FinFET. A lot of improvements have been made with it, including new architecture features and a significant jump in power efficiency. It is already being sampled to OEMs and is expected to ship in mid-2016. The new 4th gen GCN will bring with it several features, including primitive discard acceleration. This is a feature that DirectX 12 will be able to take advantage of. It is also going to support the DisplayPort 1.3 and HDMI 2.0a standards, HDR monitors, and have h.265 decoding at up to 4K and 4K h.265 encoding at 60 FPS. All of this along with a significant improvement to power efficiency in part thanks to the switch to 14 nm FinFET. This technology allows for considerably better performance-per-watt compared to current 28 nm devices. To demonstrate the improved efficiency, a Polaris powered system was put up against a Maxwell NVIDIA GTX 950 that was otherwise identical and measured the power usage. At the medium preset in Star Wars Battlefront the GTX 950 system was pulling around 140-150 W while the Polaris wielding system used just 86 Watts. WCCFtech did a little math and determined the Polaris GPU was likely drawing just about 50 W on its own, and as the GTX 960 typically draws around 100 W that puts it at double the efficiency for the same performance. With Polaris launching months from now though, we could see further improvements to its efficiency by various hardware and software optimizations. This efficiency will help Polaris enter small form-factor designs and bring without sacrificing too much performance. Source: WCCFtech and AMD Back to original news post Share this post Link to post Share on other sites More sharing options...
Recommended Posts