Nvidia decided to unveil its new flagship Quadro GPU for visual computing at the Siggraph 2013 conference and exhibition that is all about computer graphics and interactive techniques.
Based on the Kepler architecture and featuring 2880 streaming multiprocessor cores (SMX), 240 texture units and 48 ROPs as well as a total of 12GB of GDDR5 memory paired up with a 384-bit memory interface, the new Quadro K6000 GPU delivers up to five times more compute performance when compared to its predecessor, the Quadro 6000. Although it already had the Quadro K5000 GPU, also based on the Kepler architecture and also announced at last year's Siggraph, the shere power of the new Quadro K6000 puts it at the computing GPU throne.
The 28nm Kepler GPU of the new Quadro K6000 is clocked at around 900MHz while 12GB of GDDR5 memory is at 6GHz. It has a 225W TDP that can be easily justified by up to 1.7 times more graphics performance and up to five times more compute performance.
As it was the case with the Quadro K5000, the new Nvidia Quadro K6000 also supports second-generation Maximus technology that allows the K6000 to be dedicated to graphics work while compute is switched to a more powerful Tesla based card that should benefit scenarios where multiple workloads are used.
According to Nvidia, the new Quadro K6000 should start to ship sometimes this fall and will be sold individually via AIB partners like ELSA, PNY and Leadtek as well as via Nvidia's OEM partner network that includes HP, Dell, Lenovo, Supermicro and others. Although the price was not announced you can expect them to be priced at anywhere between US $5000 and $7000 a piece.