Some leaks about the performance of the next NVIDIA graphics cards have appeared in the last few days, but we have to be very careful about the sources that shared them.
Scores on TimeSpy and in games for Lovelace?
Let’s start with the well known @kopite7Kimi, who has, over the last few years, been very well informed about the new NVIDIA cards (especially the Ampere generation). He had shared that the RTX 4090 would score over 19000 points on the Time Spy Extreme benchmark, which is 66% higher than the current RTX 3090 Ti.
Despite this being a leak, the author remains a reliable source, compared to the new rumors that popped up a day later by someone named @XpeaGPU on Twitter, but who is also known on the Anandtech forum for spouting many false rumors about older NVIDIA cards.
Got first in game score. Full AD102 at "high power draw" (whatever does it mean, maybe OC?) in Control Ultra 4k RT + DLSS reaches 160+fps. Don't know which CPU, RAM, GPU clock, but it looks very high to me. What a 3090Ti does in this bench?
— AGF (@XpeaGPU) July 19, 2022
This person published that the entire AD102 chip would reach more than 160 FPS on Control in ultra in 4K resolution, with Ray Tracing and DLSS enabled. Performance that would be twice that of a 3090.
This “leak” taken by several sites has quickly made the rounds and the buzz. But as mentioned before, this person is used to coming out with nonsense claiming to have reliable sources, even more reliable than kopite7kimi.
This same person (xpea) claimed in his last leaks 2 years ago that the NVIDIA Ampere gaming would be based on 7nm from TSMC, but that they would use Samsung too. All this while claiming loud and clear that kopite7Kimi, who had an almost perfect tracking on the Ampere range, was wrong.
So we must always be wary of leaks out of the ordinary, take it with a pinch of salt and go and find the source of his information. We are not targeting this person in particular, but we recommend to be attentive to all these famous leakers coming out of nowhere, already having a dubious past.