NVIDIA finally announced the next generation of GeForce GTX 1080 video cards, the GeForce 1000 series. Controlled by the organization’s new Pascal engineering and based on TSMC’s 16nm FinFET process, NVIDIA is promising that the 1000 series will set another high bar for execution and effectiveness while likewise offering an accumulation of new elements to further separate it from its antecedents.
The initial two cards out of the entryway will be NVIDIA’s top of the line cards, the GTX 1080 and GTX 1070 series. 1080p jokes aside, these are clear successors to what has now been NVIDIA’s item stack naming plan following 2010, with the GTX 1080 speaking to their new top of the line card while the GTX 1070 is a less expensive card intended to hit the aficionado sweet spot. These cards will dispatch throughout the following month, with GTX 1080 hitting the business sector May 27th, and GTX 1070 two weeks after the fact, on June tenth. They will be estimated at $599 and $379 separately.
As of now we don’t have the foggiest idea about GP104’s die size, however at less than a large portion of the transistors than GP100, it’s an easy win this is a die size in accordance with past first-run 104 GPUs, for example, GK104 and GF104, which were 294mm2 and 332mm2 individually. GM204, by correlation, was a bigger GPU, coming later in the lifetime of the 28nm procedure. This means NVIDIA has spent some of their 16nm reserve funds on bringing the bucket size down when contrasted with GM204.
16nm is still generally at a very early stage in its life, and dissimilar to Tesla P100 where clients will pay high costs for early access to a lower-yielding chip, for consumer parts NVIDIA has should have been more preservationist. On that note, as of now it is not clear whether the GP104 in GTX 1080 is a completely empowered GP104 GPU which is what customers would love to buy instead of less bandwidth memory.