Nvidia has been smashing out their GPU’s this summer. The GTX 1080 and GTX 1070 debuted in late May and early June, GTX 1060 dropped earlier this week, and now the Titan X is set to ship by August 2. The original GTX Titan released nearly a year after the GTX 780, and the GTX Titan X about six months after the GTX 980ti.
Existing Titan GPUs have been classed as GTX cards, ie Titan, Titan Black or Titan X. In this case, Nvidia is doing away with the GTX branding on the GPU and calling it the Nvidia Titan X. This could prove confusing for some customers, but it may also show that this card is meant to bridge the gap between Nvidia’s gaming and workstation divisions.
The new GPU is based on the GP102 architecture and will boast 3,584 CUDA cores at a base clock of 1417MHz and a boost clock of 1531MHz as well as a whopping 12gb of VRAM. Although this is slower than Nvidia’s GTX 1080’s 1607MHz / 1733MHz, the 40% increase in cores will easily offset the 12% decrease in clock. Also the new 384-bit memory bus allows the GPU 480GB/s of bandwidth.
- 11 TFLOPS FP32
- 44 TOPS INT8 (new deep learning inferencing instruction)
- 12B transistors
- 3,584 CUDA cores at 1.53GHz (versus 3,072 cores at 1.08GHz in previous TITAN X)
- Up to 60% faster performance than previous TITAN X
- High performance engineering for maximum overclocking
- 12 GB of GDDR5X memory (480 GB/s)
The Titan X will be the single GPU gamers ultimate choice until we see what the GTX 1080ti can do, given that Nvidia’s GTX 1080 is supposed to retail for around $599 and 2 GTX 1080’s would easily surpass a single Nvidia Titan X. There are plenty of gamers who choose not to use multi GPU configurations, since a single high-end GPU typically provides better frame times and a smoother experience. Although this could change thanks to new methods of sharing workloads between cards. Until now we have only seen a very limited support for SLI in DX12 titles. Nvidia may have a hard time justifying this GPU’s price tag to the masses. Even if the card is 40% faster than the GTX 1080, you’d be stumping up twice the cash for 1.4x the performance. There are however people with money to burn who want the bragging rights will jump in regardless of the $1200 price tag.
As far as competition from AMD goes, their isn’t really any at the moment, as AMD’s Fury line has nothing to match the power of Nvidia’s TitanX, this will no doubt change later this year/early next year. Both async compute and slightly better performance in DX12 has already helped the Fury and Nano card’s, but if you want the high end performance in both APIs, Nvidia’s GTX 1070 & GTX 1080 have quickly taken back the title belt.
Brought to you by Greig Crichton (VR Gaming Evolved) of the VRSpies