First RTX 2070 Super / 2060 Super laptop GPU scores spotted on 3DMark and Geekbench
Even though Nvidia might not release its next gen Ampere desktop gaming GPUs this spring, the green team is surely ready to unleash a refreshed RTX Super lineup aimed at gaming laptops in mid-March. Last month we were estimating that the performance gains of the new Super laptop GPUs would hover around 10% over the older non-Super chips, and, since we are quite close to launch, some early samples have already been spotted on 3DMark and Geekbench thanks to the investigative work of twitter user _rogame.
First up are the Geekbench specs and scores for the RTX 2060 Super. The refreshed GPU will integrate 2,176 Cuda cores amounting to 34 compute units running at 1.49 GHz and coupled with 8 GB GDDR6. This model managed to score 231312 points in the Geekbench OpenCL test. For reference, the older RTX 2070 and 2070 Max-Q laptop GPUs score 245K and 220K points, respectively.
Now, _rogame suggests that non-Max-Q Super GPUs use 14 Gbps GDDR6 VRAM, while all Max-Q models (even the new Super versions) come with 11 Gbps GDDR6 VRAM, so this could explain how the new RTX 2060 Super is faster than the older RTX 2070 Max-Q model.
_rogame also found some interesting 3Dmark results for the upcoming RTX 2070 Super laptop GPU. This one was tested together with an i9-10880H CPU and managed to score 9,531 points in Time Spy, which is almost on par with the non-Super RTX 2080 (usually 10-11K points). A pretty good gain considering that the RTX 2080 is a 150W chip and the RTX 2070 Super has a 115 W TDP. For reference, the older RTX 2070 is a 115 W chip, as well, and only scores ~8K points, while the RTX 2080 Max-Q scores 8.85K at 90 W.
All things considered, it looks like the Super versions are actually more than 15% faster than their non-Super counterparts. and we just need to find out the exact pricing scheme for the Super models.