Ok, so first of all, I want to acknowledge that this is my second video in 24 hours, Go check out the other here, I'm sorry, but I saw this info at 2 AM and had to make a video about it. So, Let’s get started with Nvidia, With the release of the RTX 3050 back at CES it looks like Nvidia has completed their GPU lineup for the RTX 3000 series. And with that, it's time to turn our attention to the RTX 4000 series.
So let’s recap what we have real quick. This week, we learned that Ada Lovelace, that’s the architecture for the next-gen graphics, Well it would be based on TSMC’s 4nm process. A process more advanced than what AMD will use for their RDNA 3 graphics. RDNA 3 will be based on TSMC’s 5 and 6nm process nodes.
Something else that has been floating around is the specs of RTX 4000. Specifically, the specs for the highest-end RTX 4090. It would sport the AD102 chip with up to 18 432 Cuda cores. I say up to because it's likely to not use 100% of the chip. Also, since its board design is compatible with the 3090ti, the RTX 4090 could sport up to 24GB of GDDR6X memory. The speed of the memory is faster but we’ll get back to that in a moment.
And with all of those specs comes a HUGE total graphics power draw or TGP of 600W. Oh, and the card would apparently still use the PCIe Gen 4 interface instead of the newer PCIe Gen 5, but that’s not an issue, Gen 4 has enough bandwidth as it is.
So what’s new? Well, legendary leaker, Kopite7Kimi got us some of the specs for the more reasonable RTX 4080 and 4070. As well as a monster that is apparently roaming Nvidia’s test chambers, but we’ll get to that in a moment.