NVIDIA PAX East finally released the first video chipsets based on its architecture Fermi. The high-end GeForce GTX 480 in the head group and is billed as the fastest GPU in the world: "It has 480 processing cores visual, 16 units of geometry and four units that combined matrix must fight card ATI Radeon HD 5800 Series Real-time testing world. It also has major enhancements in multi-card SLI which produce a 90 percent rate with a second card, which makes the case for several series GeForce 400 cards in the high-end systems.
This island and the smallest are the GTX 470 cards from NVIDIA to support DirectX 11 (and optionally OpenGL 4.0) and the accompanying benefits, including geometry tessellation hardware. Fermi design much more widespread should also contribute significantly to the general-purpose computer in CUDA, and DirectCompute OpenCL. Either new GeForce card is powerful enough to handle raytracing lighting and realistic hair in real time under certain conditions.
3D Vision has received a boost as two or three cards can drive up to three screens with 3D active.
The GTX 480-499 $ makes a main clock speed of 700MHz core, 1.5 GB of GDDR5 memory clocked at 1.85 GHz and a maximum of 250W of power draw. Scales of the NVIDIA GTX 470 back considerably to a heart of 607MHz and 1.2 GB of GDDR5 memory 1.67GHz, but it costs much less $ 349 and has a slightly less demanding energy needs 215W. Advice based on the technology should be available from third parties as of today in limited quantities. Third-party manufacturers such as CyberPower PC have already signed on board to bring the fastest video equipment.
(Via electronista)
No comments:
Post a Comment