NVIDIA is aiming to launch, and the right announced the Titan V, a prosumer graphics card related (based on) to the company’s brand new Volta architecture adopting a 12nm process that supersedes the previous Pascal architecture seen on the Titan XP. It’s a $3000 graphics card that is designed solely for computing performance in concern for people working on deep learning, AI, 3D rendering, and other computationally intense applications. On the top of that Nvidea Titan V might overtake its predecessors efficiently.
At the core of the Titan V is the GV100 GPU that was first recognized on the $10000 or 644805 INR (yes, that’s one with five zeroes) Tesla V100 following in May. For the most parts, the GPU is same, with 5120 CUDA, 21.1B transistors, and 640 Tensor Cores. The Tensor Cores facilitate specific compute operations and can handle them much fast and efficiently than the standard CUDA cores. This, copulated with the 1200MHz core clock and 1455MHz advance (boost) clock, produces a single precision of 13.8 T FLOPS & double precision performance of 6.9 TFLOPS, both just a shade under what the V100 can execute. The Titan V also credited 110 T FLOPS of Tensor performance, which is relevant for deep learning operations.
Where NVIDIA has lessened a little is in the memory department. The Titan V holds 12GB of HBM2 memory along with 850MHz, 1.7Gbps memory speed, 3072-bit full memory bus width and 653GB/s memory bandwidth, all awe-inspiring numbers but not as powerful as the 4096-bit V100 16 GB memory with 900GB/s bandwidth. The card has a TDP of 250W including one 6-pin and one 8-pin connectors.
From the outer side, the card looks almost to the previous Titan XP, save for the elegant cold gold color. It flaunts the same copper heatsink along with the vapor chamber design for cooling. Backwards, are three DisplayPorts and one HDMI.
It’s worth restating that this card is intended for compute applications. It’s expensive by consumer card standards but is genuinely affordable when compared to other workstation class cards on the market; such names are Teshla products and NVIDIA’s own Quadro. The Titan V is pointed at getting into AI or machine learning stuff and delivers most of the V100 performance at a surprising rate. Of course, it’s, still, possible to game on it and it will utilize NVIDIA’s standard GeForce suite of drivers, though the performance likely won’t be much more excellent than last year’s Titan XP. Including these new Tensor Cores, the Titan V will probably blow the Titan XP out of the water when it revolves around computer applications.
The Nvidea Titan V is opened for sale on NVIDIA’s website for $2999 or 193377 INR starting today.