Way back in the ancient mists of February, we reported that RTX 4060 rumours were shaping up well in terms of performance. The latest news(opens in new tab) is that it and its Ti sibling are go for launch in May.
Unfortunately, in the intervening period the specs of the GPUs, at least according to the rumours, have had a pretty serious haircut. Back in Feb, an early RTX 4060 was supposedly benchmarked, revealing performance pretty much on par with an RTX 3070 Ti.
That «RTX 4060» GPU was said to have 4,608 CUDA cores and a 2.5GHz boost clock courtesy of the AD106 GPU. The latest rumblings(opens in new tab), however, put the RTX 4060 on the smaller AD107 chip with just 3,072 cores.
As for the RTX 4060 Ti, that's said to be the one to use the AD106 chip, but slightly cut down to 4,352 cores running at 2.5GHz. Oh and both AD106 and AD107 have 128-bit memory buses.
Obviously all this information remains nothing more than rumour. But we really hope that it's somehow wrong. Here's why. Take the rumoured 4060 Ti. It supposedly has its memory bus cut in half versus the old 256-bit RTX 3060 Ti(opens in new tab) and also fewer shaders at 4,352 versus 4,864.
OK, the 4060 Ti is clocked a lot higher and will have more cache. So, the raw compute performance will be a decent step over the 3060 Ti. But that memory bus is a killer. Not only will bandwidth be way down. But it also limits options in terms of memory size.
Unless a new video 32Gb memory chip spec appears, both the 4060 and the 4060 Ti will have the technical option of either 4GB or 8GB of total video memory using existing 16Gb GDDR6 memory.
Best CPU for gaming(opens in new tab): The top chips from Intel and AMDBest gaming motherboard(opens in new tab): The right boardsBest
Read more on pcgamer.com