It has been 3 weeks to the day since Nvidia posted an obscure tweet from its GeForce account stating, “Something super is coming…” Since at that point, AMD has turned out and declared its first run of Navi illustrations cards—Radeon RX 5700 XT and 5700—yet whatever “super” surprise lies in hold up from Nvidia has still to reveal itself. Well, officially at any rate.
There have been a couple of leaks pointing a few updated Super variants of Nvidia’s existing RTX cards. To be specific, it’s rumored Nvidia is preparing to report three new RTX cards, including Super versions of the RTX 2080, 2070, and 2060.
Nothing has been affirmed, so a portion of incredulity is in order, or a grain of salt if you prefer. However, both Videocardz and Igor Wallosek from our friends at Tom’s Hardware Germany are revealing the equivalent SKUs and specs, with the previous saying the latter “was the first to report the right variants.” In other words, Videocardz is claiming to have independently confirmed Wallossek’s information.
Let’s break this down into two sections. First, here is a look at the GPUs that the new cards will use, alongside the GPUs employed in at present shipping RTX cards, and the accompanying number of CUDA cores:
- GeForce RTX 2080 Super—TU104-450 (3,072 CUDA cores)
- GeForce RTX 2080—TU104-400 (2,944 CUDA cores)
- GeForce RTX 2070 Super—TU104-410 (2,560 CUDA cores)
- GeForce RTX 2070—TU106-400 (2,304 CUDA cores)
- GeForce RTX 2060 Super—TU106-410 (2,176 CUDA cores)
- GeForce RTX 2060—TU106-200 (1,920 CUDA cores)
As indicated by the leaked specs, the RTX 2070 and 2060 Supermodels will have 256 more CUDA cores as compared to the regular variants, while the RTX 2080 Super gains 128 CUDA cores over the standard version. There is no notice of clock speeds, yet the extra CUDA cores would give the Super cards increasingly crude muscle to throw at games and graphics.
Presently let’s have a look at the memory configuration of each and every card, including the sum, speed, and bandwidth, alongside their non-Super original versions:
- GeForce RTX 2080 Super—8GB GDDR6 @ 16Gbps, 512GB/s memory bandwidth
- GeForce RTX 2080—8GB GDDR6 @ 14Gbps, 448GB/s memory bandwidth
- GeForce RTX 2070 Super—8GB GDDR6 @ 14Gbps, 448GB/s memory bandwidth
- GeForce RTX 2070—8GB GDDR6 @ 14Gbps, 448GB/s memory bandwidth
- GeForce RTX 2060 Super—8GB GDDR6 @ 14Gbps, 448GB/s memory bandwidth
- GeForce RTX 2060—6GB GDDR6 @ 14Gbps, 336GB/s memory bandwidth
The regular RTX 2060 is the simple card of the bunch with 6GB of GDDR6 memory mated to a 192-bit bus. Every other SKU has 8GB of GDDR6 memory and a 256-bit bus, incorporating the RTX 2060 Super, consequently the higher memory bandwidth.
If the leaked specs are more precise, the RTX 2070 Super will not see an adjustment in the memory configuration, while the RTX 2080 Super will use faster-clocked memory chips, bringing about more bandwidth.
Both Videocardz and WCCFTech have pegged one week from now as a possible time frame for a declaration from Nvidia, while taking note of what an actual launch may be at some point in July. We’ll need to keep a watch out if that’s really the case.
- Analysis of Blazing Beaks, dungeons and looting in the form of rogue lite
- Analysis of Warhammer Chaosbane. The Warhammer universe adapts to the Devil formula
- Shenmue 3 is a 1-Year Epic Exclusive, Developer ‘Looking into’ Community Feedback
The same goes for value—nothing considerable has been leaked, however, Videocardz assumes that the Super variants will slip into the non-Super MSRPs, which in turn would drive the evaluating down on those current models.
If all this happens as expected, things will get intriguing in a rush. During its disclose, AMD indicated slides claiming that its Radeon RX 5700 XT is commonly quicker than the RTX 2070 when playing games at 1440p. The same goes for the non-XT model, though against the RTX 2060.
We clearly won’t draw any conclusions until we’ve completely tested AMD’s upcoming hardware, alongside whatever Nvidia might have in store.