NVIDIA is set to announce GeForce RTX SUPER series next week, here is everything that you need to know about the new series.
NVIDIA RTX 2070 SUPER & RTX 2060 SUPER
On July 2nd, NVIDIA will announce the new series featuring three SKUs: RTX 2080 SUPER, RTX 2070 SUPER and RTX 2060 SUPER. Only the last two will launch simultaneously. Currently, no date is set for the RTX 2080 SUPER launch.
The GeForce RTX 2070 SUPER is using TU104 GPU with 2560 CUDA cores. This model comes with the same memory configuration as RTX 2070 Non-Super: 8GB GDDR6 256-bit with 14 Gbps bandwidth. This SKU will require 30W more than its predecessor (215W)
The RTX 2060 SUPER is now featuring TU106 GPU with 2176 CUDA cores. The Wattage has increased by 15W to 175W. The biggest change came to memory configuration. Unlike RTX 2060 Non-Super the new card has 8GB GDDR6 memory across a 256-bit interface.
As for the RTX 2080 SUPER, the card is only mentioned in a press deck with a claim of “15 Gbps Mem”.
The embargo on NVIDIA RTX SUPER lifts on:
GeForce RTX 2060 SUPER / RTX 2070 SUPER Under Embargo Until Tuesday, 7/2 @ 6AM PST (California Time).
Both SKUs (2070S and 2060S) will be available on July 9th. On the same day, the embargo on AIC cards will lift. Only Founders Edition models reviews will be posted on July 2nd. Pricing information is not available yet.
| NVIDIA RTX SUPER Official Performance Claims | |||
|---|---|---|---|
| VideoCardz.com | Claim #1 | Claim #2 | Claim #3 |
| RTX 2080 SUPER | 11+11 TOPs* / 89 Tensor TFLOPs | “Faster than Titan Xp” | – |
| RTX 2070 SUPER | 9+9 TOPs* / 73 Tensor TFLOPs | “Faster than GTX 1080 Ti” | 16% AVG faster than RTX 2070 |
| RTX 2060 SUPER | 7+7 TOPs* / 57 Tensor TFLOPs | “Faster than GTX 1080” | 15% AVG faster than RTX 2060 |
| * (FP32+INT32) | |||
Here is how the full stack of RTX cards would look like, bear in mind, RTX 2080S specs have not yet been confirmed (they are not listed in Reviewer’s Guide):
| NVIDIA GeForce RTX SUPER Specifications | ||||
|---|---|---|---|---|
| VideoCardz.com | GPU | CUDA Cores | Memory | Memory Speed |
| TITAN RTX | TU102-400 | |||
| RTX 2080 Ti | TU102-300 | |||
| RTX 2080 SUPER (TBC) | TU104-450 | |||
| RTX 2080 | TU104-400 | |||
| RTX 2070 SUPER | TU104-410 | |||
| RTX 2070 | TU106-400 | |||
| RTX 2060 SUPER | TU106-410 | |||
| RTX 2060 | TU106-300 | |||
The full specs as listed in reviewer’s guide:
| NVIDIA GeForce RTX SUPER Full Specifications | ||
|---|---|---|
| VideoCardz.com | RTX 2070 SUPER | RTX 2060 SUPER |
| Graphics Processing Clusters | 5 or 6 | 3 |
| Texture Processing Clusters | 20 | 17 |
| Streaming Multiprocessors | 40 | 34 |
| CUDA Cores (single precision) | 2560 | 2176 |
| Tensor Cores | 320 | 272 |
| RT Cores | 40 | 32 |
| Texture Units | 184 | 136 |
| ROP Units | 64 | 64 |
| Rays Cast | 7 Giga Rays/second | 6 Giga Rays/second |
| Base Clock | 1605 MHz | 1470 MHz |
| Boost Clock | 1770 MHz | 1650 MHz |
| Memory Clock | 7000 MHz | 7000 MHz |
| Memory Data Rate | 14 Gbps | 14 Gbps |
| L2 Cache Size | 4096 K | 4096 K |
| Total Video Memory | 8192 MB GDDR6 | 8192 MB GDDR6 |
| Memory Interface | 256-bit | 256-bit |
| Total Memory Bandwidth | 448 Gb/s | 448 Gb/s |
| Texture Rate (Bilinear) | 326 GigaTexels/s | 246 GigaTexels/s |
| Fabriaction Process | 12nm FFN | 12 nm FFN |
| Transistor Count | 13.6 Billion | 10.8 Billion |
| Connectors | 3x DisplayPort 1x HDMI 1x USB Type-C | 2x DisplayPort 1x Dual-Link DVI 1x HDMI 1x USB Type-C |
| Form Factor | Dual Slot | Dual Slot |
| Power Connectors | One 6-pin, One 8-pin | One 8-pin |
| Recommended Power Supply | 650 Watts | 550 Watts |
| Total Graphics Power (TGP) | 215 Watts | 175 Watts |
