NVIDIA RTX TITAN is Official, Launches in December at $2500
We reported yesterday that first images of RTX TITAN have begun popping up through various sources. Today, NVIDIA is making it’s next behemoth official.
Codenamed “T-Rex” for all the right reasons, RTX TITAN will join the current RTX product stack: RTX 2070, RTX 2080, and the RTX 2080 Ti. Although the new TITAN shares the same architecture family, it’s distinct in a way that it costs way more than the most expensive RTX graphics card in the market right now. It costs a whopping $2500 USD, and will be available this December.
“Turing is NVIDIA’s biggest advance in a decade – fusing shaders, ray tracing, and deep learning to reinvent the GPU,” – Jen Hsun Huang, NVIDIA CEO
NVIDIA’s TITAN has always been a prosumer card and the new one is no different. Built on the same Turing architecture, it features the Tensor Cores found in TITAN V but with the addition of RT Cores. This makes the RTX TITAN a more affordable choice for Ray Tracing and other compute intensive workloads.
NVIDIA is equipping RTX TITAN with 24GB GDDR6 memory – that’s significantly more than TITAN V’s 16GB memory. NVIDIA is ensuring that this card can hold up to the demands of the most memory intensive workloads.
On the spec sheet, we’re looking at the full TU102 chip featuring 4608 CUDA Cores. It features 72 RT Cores and 576 Tensor Cores capable of achieving 11 GigaRays per second. The 24GB GDDR6 clocked at 14Gbps gives it 672GB/s memory bandwidth.
Make no mistake, RTX TITAN is not for gaming. Though we reckon there will be a few enthiasasts who would love to get their hands on it. Last year’s TITAN V based on Volta architecture was strictly aimed at AI/Deep learning workloads that includes training neural networks. The introduction of RT Cores on the silicon makes RTX TITAN a suitable choice for content creators as well, along with the full-suite of Turing’s AI/Deep Learning capabilities.
You must be logged in to post a comment.