Site icon

Nvidia RTX 4090 GPU could blow the RTX 3090 out of the water

We’ve received a new update on Lovelace’s high-end graphics cards from a regular source of Nvidia leaks on Twitter, which provided some additional details, including some enticing information about the RTX 4090.

This comes from Kopite7kimi, who gave us a whole bunch of details about the GPUs coming in just a week and a half ago.

see more

As you might imagine, in such a short period of time, a lot of the alleged leaked specs remained the same as in the previous leak from Kopite7kimi. But there are some key differences here, notably some of the slick clock speeds of the flagship Lovelace graphics card (based on what a recent leaker hinted at).

The clock speed of the RTX 4090 is set to be a base clock of 2235MHz with a boost of up to 2520MHz, although the actual maximum that overclocking can reach exceeds 2750MHz, the leaker believes (which is just a belief in what has been said them by the sources – so let’s not read too much into any of this yet).

Kopite7kimi also pins a power usage of 450W on the flagship Lovelace, which matches what we heard earlier. The RTX 4090 would be a shortened AD102 chip, of course, and power usage could be pushed much further (like 600W) with a full-fat version (it could be the 4090 Ti, or maybe even a new RTX Titan).

The info on the RTX 4080 is pretty much the same story, except now Kopite7kimi thinks 16GB of VRAM will be GDDR6X – the leaker wasn’t sure about this last time – and the prediction for 18Gbps of RAM has been revised upwards to 21Gbps In the second, that’s some good news (the theory).

Power usage is tied to 420W, which again echoes what the leaker said a month ago, although they added a question mark to that wattage at the time – so this seems to indicate that Kopite7kimi is getting more confident that this is the RTX 4080 can be installed in.

Finally, regarding the RTX 4070, there is nothing new in terms of specifications since the last major update of Kopite7kimi. 300W power usage is mentioned again, which has been a leaker’s best guess for some time now. In fact, Kopite7kimi first brought up this number in April, before he had much clues about the RTX 4080’s power requirements.

Analysis: Is Lovelace’s flagship twice as powerful as the RTX 3090?

What it tells us here is that some of the energy use numbers presented seem to have been held as forecasts for some time now, suggesting that these numbers may be set in stone, as they were. Or at least they can now represent a more likely correct estimate of where Nvidia will end up with TGPs for its RTX 4000 graphics cards.

None of the information presented here is much of a surprise, then, and it’s more about consistency than the new finds — except for the (alleged) specific clock speeds of the RTX 4090. With 2750MHz mentioned in addition, Nvidia appears to be looking toward the 3GHz mark for the flagship, unlike The most conservative clocks seen on the current-generation Ampere (the 3090 has a 1.7GHz boost speed, as it’s officially rated by Nvidia – although of course it can be pushed faster – and even the Ti version only hits 1.86GHz).

And with the RTX 4090 loaded with CUDA cores as is – 16,384 of them – along with that theoretical 50% increase in boost speed, this signals a very powerful next generation. In fact, it can be twice as powerful as the RTX 3090, which is a jump in performance we’ve seen before, as you might remember.

Although nothing is mentioned about other clock speeds here, given the relative counts of the CUDA Core, and the RTX 4080 very close to the RTX 4090 for TGP (420W vs 450W), it’s easy to imagine that the 4080 will push harder with clock speeds.

It makes sense that Nvidia would want to drive as hard as possible to get the best possible gains from the premium design to 5nm with the Lovelace, and in fact with the flagship, there would be plenty of room to build something better than the initial 4090 (remember we’re looking more at 600 watts for the 4090 Ti, According to the rumor maker).

Finally, it is worth noting that while it is not new information, there is still disappointment circulating online about the alleged memory load of the RTX 4070, and it may be only 10 GB of VRAM. Previous rumors suggested 12GB would be more palatable, but it looks like this configuration might be reserved for the 4070 Ti – take all of that with the help of proper salt, naturally.

Across Wccftech (Opens in a new tab)

Source link

Exit mobile version