AMD has unveiled Radeon Pro Duo graphics card that is the currently fastest graphics card in the world. The graphic card is basically two Radeon R9 Nano cards (Fiji GPUs) that have been grafted into one. It features LiquidVR making it a perfect candidate for use in the company’s Virtual Reality platform. The graphics card boasts 16 teraflops in computing performance that is 1.5 faster than rival Nvidia’s GTX Titan X card. It is also about 4.4 Teraflops more than the recently launched Pascal-based Tesla P-100 GPU.
Radeon Pro Duo graphics card uses AMD’s HBM memory, allowing a far more compact package than a card with GDDR5 memory, commonly found in most GPUs. Each of the two Fiji GPUs used in making the graphics card features 4,097 stream processors, 256 texture units, 64 compute units and 64 ROPs. Each GPU offers 4 GB of HBM running at 500MHZ, which translates, into 1024 per second memory bandwidth. The graphics card has three 8-pin connectors, an HDMI port, and three Display Port connectors.
Power used by graphics card is a very reasonable 350 watts. AMD officials have in fact noted that they would have made the card with just 8-pin power plug but decided to add a third for overclocking headroom. This means that overclocking in the new graphics card is better than on previous based Fiji graphics core cards.
The period, which AMD will boast about having the fastest graphics card, is not known. One thing for sure is that Nvidia will be working to reclaim the throne. Two Titan X cards would compete closely with Radeon Pro Duo, though It is worth to note that the term fastest GPU usually refers to highest performance out of a single video card even if it has two GPUs.
AMD is not targeting only gamers with the new cards but also content creators in entertainment, cinema, education, medicine, and journalism.
AMD Radeon Pro Duo will be globally available on April 26th for $1,500.
Featured image credit: venturebeat