Nvidia RTX 2060 Super and RTX 2070 Super Review

Nvidia RTX 2060 Super and RTX 2070 Super Review

RTX 2070 Super Specifications

Demonstrating that there is more than one way to skin a cat, the RTX 2070 Super takes a slightly different approach than the upgrades applied to the RTX 2060 Super. Both start off the same, with the RTX 2070 Super having four extra Shader Modules which give you another 256 CUDA Cores. However, whilst the RTX 2060 Super was 105 MHz faster at base clock and 30 MHz slower when utilising GPU Boost, the RTX 2070 is just faster, gaining 195 MHz base clock and 150 MHz when boosted.

This has a significant effect upon the calculative abilities of the card. The RTX 2060 Super was 0.7 TFLOPS faster than its predecessor but the RTX 2070 Super is a scarcely believable 1.6 TFLOPS faster than the already speedy RTX 2070. Similarly the Tensor Cores saw a 11% boost on the RTX 2060 between vanilla and Super, but the RTX 2070 Super gains 20%. Lastly the RTX 2060 Super gained 16 Texture Units over the regular model for a 11% increase in Giga Texels Per Second fill-rate, whilst the RTX 2070 Super gains 40 new TUs for a 39.6% increase in Texel fill-rate.

The only area that the RTX 2070 Super doesn’t have a massive lead on its illustrious predecessor when compared to the improvements on the RTX 2060 Super is in Memory performance, but that’s largely because the RTX 2070 was already well endowed.

  GeForce RTX 2070 GeForce RTX 2070 Super
SMs  36 40
CUDA Cores  2304 2560
Base Clock  1410 MHz 1605 MHz
GPU Boost Clock  1620 MHz 1770 MHz
FLOPS  7.5 TFLOPS (INT32/FP32) 9.1 TFLOPS (INT32/FP32)
Tensor FLOPS  60 TFLOPS 72 TFLOPS
Rays Cast  6 Giga Rays 7 Giga Rays
Texture Units  144 184
Texel Fill-Rate  233.3 GT/s  325.7 GT/s
Memory Interface  256-bit 256-bit
Memory Data Rate  14 Gbps 14 Gbps
Memory Bandwidth  448 GB/s 448 GB/s
Memory Size  8 GB 8 GB
Max L1 Cache Size  2304 KB 2560 KB
TGP  175 W 215 W

Let’s look at them in the flesh.