Nvidia’s Next-Gen GPU ‘Ada Lovelace Is No Longer a Simple Ampere Upgrade’, Rumors Suggest

The latest rumors about Nvidia’s upcoming GPU architecture – codenamed Ada, Lovelace, or Ada Lovelace, depending on who you talk to – are far more positive than those ugly rumors of 900W boards. Where once Lovelace was hitched like little more than a slight Ampere updateit is now being suggested that it is much more than that and closer to an entirely new GPU architecture.

Is this further evidence that AMD’s upcoming RDNA 3 GPUs are worrying Jen-Hsun and company? That’s certainly the suggestion of tweaker @kopite7kimi, who seems to suggest that while it was once an update, it’s not anymore. If “the current AD102 is NOT the original AD102”, something has caused Nvidia to change course.

View more

Yes, you can tell that we’re getting closer and closer to a GPU battle royale as rumor has been aggressively backing away from leak after leak from some major Twitter accounts. We’re betting on new high-end graphics cards from both AMD and Nvidia to land on our platforms later in the year, and the closer we get to launch, that’s only going to get bigger.

But it’s providing us with some interesting potential insights and, if nothing else, some interesting things to consider as we head into what could be one of the closest battles for GPU dominance we’ve ever seen. And maybe even one where AMD can pull a Ryzen and come out on top.

And the potential for the RDNA 3 rumors to translate into a dominant GPU architecture for the first time could be the reason we’re hearing about Nvidia testing an AD102-based graphics card that draws around 900W of juice. Likewise, why might he have chosen to appeal to Hopper a little more than he might have been planning for Lovelace.

Either way, we’re looking at a rubber-meets-the-road moment when RDNA 3 and Lovelace come face to face, or chip to chip, where two very different architectural design options will come into close contact.

Nvidia is accumulating a monolithic chip approach, packing as much N4 silicon as possible into a GPU you could call a consumer chip. While AMD is dishonest, aiming to make the leap to GPU chiplets is worth it for its graphics card division in the same way it did when Ryzen took the chiplet approach.

One is brute force, the other is advanced, lean and efficient. But the other way of looking at it is that one is tried and true and the other is almost entirely unknown. How games respond to GPU chiplets will be fascinating and potentially one of the pain points when it comes to testing them later this year.

Nvidia, however, expects AMD to pull a Bulldozer instead of a Ryzen.

Leave a Comment