The leaks ahead of the NVIDIA GeForce RTX 20 SUPER launch were pretty insane, but the embargo has been lifted and we can finally talk about the new GeForce RTX 2060 SUPER and GeForce RTX 2070 SUPER models. For starters these two cards are priced at $399 and $499, respectively and both will be available for purchase on July 9th. There will also be an NVIDIA GeForce RTX 2080 SUPER model coming on July 23rd, 2019 for $699 that we’ll be covering on that date. Today we will be focusing on the 2060 SUPER and 2070 SUPER as these are the cards that we’ve been able to test over the past week.
The ‘refresh’ of the RTX 20 series shouldn’t come as too big of a surprise as AMD is launching their first 7nm GPU next week with the release of the Radeon RX 5700 XT 50th Anniversary Edition – $499 , Radeon 5700 XT ($449) and Radeon 5700 ($379). That means the main competition for the NVIDIA GeForce RTX 2060 SUPER will eventually be the AMD Radeon RX 5700 and then the NVIDIA GeForce RTX 2070 SUPER is set to face off against the AMD Radeon RX 5700 XT cards. Since the Radeon 5700 ‘Navi’ series cards are under embargo until July 7th, we won’t be able to give you the full picture until next week.
Thankfully our review of the GeForce RTX 2070 SUPER and GeForce RTX 2060 SUPER should be pretty simple and straight forward. NVIDIA didn’t do any architecture changes on these cards and kept the same GPU coolers. Other than some minor cosmetic changes on the front and rear plates the outside physical appearance is the same.
The new GeForce RTX 2060 SUPER is based on the same ‘Turing’ TU106 GPU that the original RTX 2060 used, but it comes with an increased core count, wider memory bus, and 8GB of memory. These improvements increase the price by $50 to $399, but the GeForce RTX 2060 SUPER is up to 22% faster (15% average) than the GeForce RTX 2060 GPU. In fact, NVIDIA says the GeForce RTX 2060 SUPER is just 1% slower than the stock GeForce RTX 2070 on average overall at 1440p and delivers higher performance than the GeForce GTX 1080. This is one of the primary reasons that the NVIDIA GeForce RTX 2070 will no longer be produced and will be gone once remaining inventory is sold. The NVIDIA GeForce RTX 2080 will also be sharing the same fate.
The GeForce RTX 2060 SUPER is said to be using the full Turing TU106-410 GPU die with 2176 CUDA Cores, 136 TMUs, and 64 ROPs. It also features 272 Tensor Cores and 32 RT cores. The clock speeds on the GeForce RTX 2060 SUPER Founders Edition model are 1470 MHz base and 1650 MHz boost with the TGP (Total Graphics Power) being 175W. The GDDR6 on this 2060 SUPER runs on a wider 256-bit memory interface and there is now a total of 8GB of memory running at 14Gbps. This gives the card an impressive memory bandwidth of 448 GB/s, which happens to be a 33% increase over the RTX 2060 model that was just introduced last year.
Next up we have the NVIDIA GeForce RTX 2070 SUPER that uses a new derivative of NVIDIA’s ‘Turing’ TU104-410 GPU, This new model has 2560 CUDA Cores, 184 TMUs, and 64 ROPs. It also features 320 Tensor Cores and 40 RT cores. The price tag on this graphics card is $499 and NVIDIA says it is not only faster than the GeForce GTX 1080 Ti GPU, but is also up to 24% faster (16% average) overall than the standard GeForce RTX 2070 at 2560×1440 resolution. The RTX 2070 SUPER brings no changes to the memory bus, clock speed or amount of GDDR6 memory. The big enhancements for this GPU come from 4 extra SMs being unlocked and higher clock speeds of 1605 MHz base and 1770 MHz boost. The higher core clock speeds and extra cores increase the TGP (Total Graphics Power) of this card 215W, which is a 40 Watt increase over the RTX 2070.
Let’s move along to the benchmarks to see how these two new GPUs perform!