Nvidia’s “something super” is looking ever more like it is going to be new graphics hardware, as fresh rumours peg new GeForce RTX 2070 Super cards for a mid-July launch with UL dropping a fresh update for its ray tracing specific Port Royal benchmark “to improve compatibility with upcoming hardware.” Now we know that AMD’s Navi is coming soon, and will get an airing at E3 next week, but team Radeon is only just gearing up to talk about its ray tracing strategy, not to launch one.
And everything we’re hearing from inside and out of Nvidia is that the ‘super’ teaser will come to fruition at the LA gaming show next week. We’re going to be right there at the heart of E3, so you can bet we’ll have the goods for you as soon as we’re allowed to talk GeForce tech.
With AMD trying to steal the gaming show, in the same way it absolutely pwned Computex, Nvidia definitely wants to have a response ready for when it goes into detail about the new mainstream Navi graphics cards. And that’s got to be a new range of its own GeForce RTX cards, which look like they’ll be appearing very soon.
The 3DMark Port Royal update was dropped yesterday, and is pegged as a minor update that doesn’t actually affect any of the current or future benchmarking scores. But the most important part about it is that the patch has been made almost entirely just to improve its functioning with new hardware.
There are just two bullet points on the Steam patch post:
- Updated DirectX 12 resource handling in Port Royal to improve compatibility with upcoming hardware and drivers.
- Improved presentation of benchmark results when using custom run settings.
And it’s that “upcoming hardware” bit that’s key here. Port Royal is all about the ray tracing, so it’s got to be referencing new Nvidia RTX GPUs rather than the upcoming mainstream AMD cards.
Pixel pushers: These are the best graphics cards around today
The current expectation for the Nvidia ‘super’ is that team GeForce will look to leverage the speedier iterations of GDDR6 to allow it to create mildly updated versions of its current range of RTX cards using 16Gbps memory chips. There is also some speculation that Nvidia will create a specific RTX 2070 Ti card, using the same TU 104 GPU as the RTX 2080, but with fewer CUDA cores inside it.
A recent post on Chinese forum, Weibo, (via WCCFTech) purports to show the specifications for three new Super RTX cards launching in mid-July, however, which flies in the face of Nvidia launching an Ti version of the RTX 2070.
The post lists the CUDA core counts and new GPU designation for the three rumoured cards; the RTX 2080 Super, RTX 2070 Super, and RTX 2060 Super.
2080 super 3072 TU104-450 16Gbps 8G,
2070 super 2560 TU104-410 14Gbps 8G,
2060 super 2176 TU106-410 14Gbps 8G
Honestly, none of the information in the Weibo post strikes us as particularly prescient given that it could all be cobbled together from previous rumours and Nvidia GPU naming passim. We’ve known all along there was a full-fat TU104 GPU available, should yields of the chips be strong enough that team GeForce didn’t need to make sure it could put any slightly b0rked silicon to good use.
And we’ve spoken before about the likelihood of cutting the TU104 down even further to make an updated version of the RTX 2070 too, but it’s still interesting to note that this latest rumour only pegs the updated GDDR6 memory to the RTX 2080 Super.
However it turns out we think it could all be a smart play, especially if Nvidia chooses to launch these new Super RTX cards at the same price as the current-gen versions. That will allow it to drop the sticker prices of the already released RTX graphics cards without necessarily looking like it’s reacting to AMD’s Navi out of fear.
But whatever the truth of it, we’re only a few days away from finding out what the plan is for both AMD’s Navi and Nvidia’s Super GPUs.