It wasn’t too long ago that NVIDIA revealed its 10-series line of graphics cards, the GeForce GTX 1080, 1070, and budget-level 1060. While the big surprise with those cards was that they outperformed the 900-series, and were even more affordable to boot, NVDIA’s latest announcement goes in the opposite direction with the reveal of the new US$1,200 Titan X graphics card.
The Titan X is the latest in NVIDIA’s line of so-ridiculous-it-hurts GPUs, the kind of technology you’d only get when money is no option. NVIDIA claims the card is 60% faster than the previous Titan X, though how it compares to the GTX 1080 will likely be the more interesting benchmark. Both the Titan X and the rest of the 10-series are based off NVIDIA’s new Pascal architecture, which support a number of nifty VRWorks features.
In a press release, NVIDIA detailed the card’s performance stats:
- 12-billion transistors
- 11 TFLOPs FP32 (32-bit floating point)
- 44 TOPS INT8 (new deep learning inferencing instruction)
- 3,584 CUDA cores at 1.53GHz (versus 3,072 CUDA cores at 1.08GHz in previous TITAN X)
- 12GB of GDDR5X memory (480GB/s)
The Titan X will be available for – let’s say it again – US$1,200, across both the United States and Europe. The card will be made available in the Asia-Pacific region at a later date. NVIDIA is also looking into using the card for deep learning, as detailed in a follow-up blog post.