A blizzard of snippets of information regarding a purported Nvidia GeForce RTX Super series of graphics cards has swept across the internet in recent hours. Thankfully, VideoCardz has distilled all the latest info, backed up by Igor Wallossek from Tom's Hardware Germany, into a cogent piece that is worth a sip - though of course a pinch of salt is still required.
VideoCardz says that it has confirmed three new GeForce RTX Super series graphics cards that were outlined by the Igor's Lab TV YouTube channel earlier this week. In brief its sources confirm the names, extra CUDA cores, and memory buffs that are destined for the series. This is summed up in the screenshot directly below.
Staring at the entry-level for RTX graphics cards, it is reported that the GeForce RTX 2060 Super "will basically be a cut-down version of RTX 2070 non-Super, featuring TU106-410 GPU with full 256-bit memory bus". It is good to see this SKU get 8GB of GDDR6.
Moving upwards we see that there will be a faster Nvidia GeForce RTX 2070 variant, augmented by a further 256 CUDA cores thanks to its TU104-410 GPU.
Finally, the GeForce RTX 2080 Super, based upon the TU104-450 GPU will boast 3072 CUDA cores (up from 2944) and faster GDDR6 memory (16Gbps vs 14Gbps). There isn't any sign of any other Super RTX or GTX variants as yet.
On the topic of pricing, the new Super variants are supposed to launch at the MSRP of their predecessors, with the non-Super variants getting a price cut. Timing-wise, the rumour mill expects Nvidia to announce these new GeForce RTX Super graphics cards next Thursday, with availability scheduled for mid July.
It will be interesting to see, if the above pans out, how it will change the competitive landscape with AMD. Remember that AMD is introducing its Radeon RX 5700 and RX 5700 XT cards, which compete with the existing RTX 2060 and RTX 2070 respectively, on 7th July.