Friday, November 19th 2010

Club 3D Announces Radeon HD 6870 Overclocked Custom-Design Graphics Card

Club 3D B.V. is pleased to announce the introduction of the next generation video card based on the Bart XT chipset, AMD’ second generation DirectX 11. The new Club 3D Radeon HD 6870 1GB GDDR5 Overclocked Edition video card features now the AMD HD3D technology which allows you to play your favorite games in full stereo 3D with a single video card (up to 3 monitors), and enjoy with the new EyeSpeed technology to watch the latest Blu-ray movies in 3D.

Do things faster with AMD Accelerated Parallel Processing (APP) technology on the most demanding applications and take advantage of the Dolby True HD support for a theatre-quality experience. Now is available to you with 940 Mhz GPU Clock, 4400 Mhz GDDR5 and 1120 Stream Processors.

Features
  • Second generation Microsoft DirectX 11 support
  • PCI Express 2.1 x16 bus interface
  • “Eye-Definition” graphics
  • AMD Eyefinity multi-display technology
  • AMD EyeSpeed visual acceleration
  • UVD 3 dedicated video playback accelerator
  • AMD Accelerated Parallel Processing technology
  • DirectCompute 11
  • OpenCL Support
  • AMD HD3D technology
  • Blu-ray 3D support
  • AMD CrossfireX multi-GPU technology
  • Integrated HDMI 1.4a with support for stereoscopic 3D
  • Integrated HD audio controller
Add your own comment

3 Comments on Club 3D Announces Radeon HD 6870 Overclocked Custom-Design Graphics Card

#1
arnold_al_qadr
does OpenCL is the same thing as Physx (on nvidia)?
sorry, newbie here..
:confused:
Posted on Reply
#2
chaotic_uk
same cooler as the powercolor pcs ?
Posted on Reply
#3
TheLaughingMan
by: arnold_al_qadr
does OpenCL is the same thing as Physx (on nvidia)?
sorry, newbie here..
:confused:
No. OpenCL is a parallel to DirectX more or less. PhysX's AMD counterpart would be the software based Havoc Physics Engine. It works well and is easy to access by game developers, but is very limited in it calculations due to them being done on the CPU which does not do that type (parallel processing) well.

Currently AMD, Nvidia, Intel, IBM, and several other major players are working out a deal to allow a free, open-sourced, middleware to run physics on GPU's no matter who makes them (AMD, Nvidia, and Intel holding the lion share of that domain). This middleware's status is on hold for now as they can't seem to agree on a variety of things. AMD also started a projected called AMD Stream which is suppose to be a direct counterpart to CUDA. While its finalization was announced years ago....nothing new has come up about it since as adoption seems to be either slow or not being properly marketed.

There is still talk about if GPU powered physics is still the way to go and AMD is holding out until it is confirmed, but there is no proof of that. Either way, OpenCL can handle CUDA, AMD Stream, and PhysX with some rather minor tweaking as it is designed to allow for plug-in style middleware. Think of OpenCL as your favorite programming language (such as C++, Java, Javascript, Perl, Python, etc.) but it runs programs like CUDA, PhysX, Stream, or whatever on your GPU instead of the CPU. While this doesn't help most programs and can hurt some, when massive parallel processing is needed, running on the GPU can increase performance ten fold.

Sorry I got off topic. There is a lot of detail left out and I may have generalized some info. a little too much, but that should get you started. By the way, the answer was no.

OpenCL Intro: http://developer.amd.com/gpu/ATIStreamSDK/pages/TutorialOpenCL.aspx

CUDA/GPU processing: http://www.nvidia.com/object/cuda_home_new.html

GPGPU: http://www.amd.com/us/products/technologies/stream-technology/opencl/pages/gpgpu-history.aspx
Posted on Reply