Technology-news site Fudzilla recently published some information about graphics specialist NVIDIA's (NVDA 3.65%) next gaming-oriented graphics processors.

"The upcoming Geforce [sic] GPU architecture will be more incremental than revolutionary," Fudzilla writes. GeForce is NVIDIA's gaming-oriented graphics processor brand.

Fudzilla says those new GeForce GPUs won't be based on the company's recently announced Volta architecture, which is expected to ship in the form of a data-center-specific chip next quarter, but instead a "Pascal-influenced design-derived shrink down."

The metal shroud on NVIDIA's GeForce GTX 1080 Ti.

Image source: NVIDIA.

The implication here, then, is that the new gaming-oriented chips will be built on the same customized-for-NVIDIA 12-nanometer technology by Taiwan Semiconductor Manufacturing Company (TSM 2.84%), but all the design changes in Volta compared with the current Pascal-architecture products won't carry over.

Here's why I'm not buying this rumor.

That's not how NVIDIA does things

It's true that in recent years, NVIDIA's professional and data-center graphics processor designs have diverged somewhat from the designs that make their way into the hands of gamers.

For example, the NVIDIA Tesla P100 -- the first product announced based on the company's Pascal architecture last year -- uses a fast, efficient, but expensive type of memory known as HBM2. The P100 also includes a lot of cores dedicated to handling high-precision floating point math that's useful in some data-center applications but not for games.

The GeForce GTX 10-series processors, which are also based on the Pascal architecture, use different memory types -- either GDDR5 or GDDR5X, depending on the product -- and include far fewer of those high-precision processor cores. There are also some slight differences in the way the architecture itself is organized.

But at the end of the day, from NVIDIA's highest-end supercomputer chips to its lowest-cost gaming chips, the company's entire product line uses the same basic Pascal architecture.

I suspect we're going to see the same thing with NVIDIA's next gaming cards. Things like the high-precision cores and the dedicated tensor processors will either be entirely absent or there in much smaller proportions relative to the technology that's useful for gaming.

Remember: NVIDIA said in its white paper describing the Volta architecture that Volta "features a major new redesign of the [streaming multiprocessor] architecture that is at the center of the GPU" and that this new design "is 50% more energy efficient than the previous generation Pascal design."

I find it hard to believe that NVIDIA would go through the trouble to redesign the streaming multiprocessor for dramatically better energy efficiency and not deploy that technology in what is still easily NVIDIA's highest-revenue business -- gaming!

Expect Volta for gaming, and expect it soon

Considering the third-quarter release schedule of NVIDIA's Tesla V100 based on the Volta architecture for data centers, the consumer-oriented Volta-based products can't be too far behind. At the very least, I expect a Titan-class premium gaming/prosumer part to launch in the first quarter of 2018, based on a press release from graphics memory maker, SK Hynix, and I wouldn't be surprised to see the rumors of gaming-oriented Volta products launching this fall also coming true.